]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commitdiff
overview / pkg / ggml / sources / llama.cpp / commitdiff
summary | shortlog | log | commit | commitdiff | tree
raw | inline | side by side (parent: 73d284a)
ggml webgpu: add CEIL operation support (#18605)
authornwyin <redacted>
Mon, 5 Jan 2026 19:38:57 +0000 (13:38 -0600)
committerGitHub <redacted>
Mon, 5 Jan 2026 19:38:57 +0000 (11:38 -0800)
* ggml-webgpu: add CEIL operation support

      Add support for the CEIL unary operation in the WebGPU backend:
      - Add CEIL_FUNC shader template in unary_op.wgsl
      - Add 4 shader variants (f32, f16, inplace versions)
      - Initialize CEIL pipelines in ggml-webgpu.cpp
      - Register CEIL in supports_op function

* docs: update WebGPU ops support for CEIL

docs/ops.md patch | blob | history
docs/ops/WebGPU.csv patch | blob | history
ggml/src/ggml-webgpu/ggml-webgpu.cpp patch | blob | history
ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl patch | blob | history

diff --git a/docs/ops.md b/docs/ops.md
index 2b2770cb765456c4f7fb95b704ccd47f3854ba2e..142f401d03a2ad8f676b8fc815ae4e724fd0fd2c 100644 (file)
--- a/docs/ops.md
+++ b/docs/ops.md
@@ -22,7 +22,7 @@ Legend:
 |                           ARANGE | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                           ARGMAX | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                          ARGSORT | ❌ | ✅ | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | ❌ | ❌ | ❌ |
-|                             CEIL | â\9d\8c | â\9d\8c | â\9c\85 | ð\9f\9f¡ | â\9d\8c | â\9d\8c | ð\9f\9f¡ | ð\9f\9f¡ | â\9d\8c | ❌ | ❌ |
+|                             CEIL | â\9d\8c | â\9d\8c | â\9c\85 | ð\9f\9f¡ | â\9d\8c | â\9d\8c | ð\9f\9f¡ | ð\9f\9f¡ | â\9c\85 | ❌ | ❌ |
 |                            CLAMP | ❌ | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | 🟡 | ❌ | ❌ | ❌ |
 |                           CONCAT | ❌ | ✅ | ✅ | 🟡 | ✅ | 🟡 | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                             CONT | ❌ | 🟡 | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | 🟡 | ❌ | ❌ |
diff --git a/docs/ops/WebGPU.csv b/docs/ops/WebGPU.csv
index bfff75e66f0c9e2322c86fbf21ef409d4f8be80e..8cd7e12001e6e32cb06d1cd7479dcb7d4a7b76ee 100644 (file)
--- a/docs/ops/WebGPU.csv
+++ b/docs/ops/WebGPU.csv
@@ -35,8 +35,8 @@
 "WebGPU: WebGPU","GELU_ERF","type=f16,ne_a=[5,7,11,13],v=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne_a=[5,7,11,13],v=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne_a=[5,7,11,13],v=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne_a=[128,2,2,2],v=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne_a=[5,7,11,13],v=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne_a=[5,7,11,13],v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
@@ -77,8 +77,8 @@
 "WebGPU: WebGPU","GELU_ERF","type=f16,ne_a=[5,7,11,13],v=1","support","1","yes","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne_a=[5,7,11,13],v=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne_a=[5,7,11,13],v=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne_a=[128,2,2,2],v=1","support","1","yes","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne_a=[5,7,11,13],v=1","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne_a=[5,7,11,13],v=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
@@ -119,8 +119,8 @@
 "WebGPU: WebGPU","GELU_ERF","type=f32,ne_a=[5,7,11,13],v=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne_a=[5,7,11,13],v=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne_a=[5,7,11,13],v=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne_a=[128,2,2,2],v=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne_a=[5,7,11,13],v=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne_a=[5,7,11,13],v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f32,ne_a=[128,2,2,2],v=0","support","0","no","WebGPU"
@@ -161,8 +161,8 @@
 "WebGPU: WebGPU","GELU_ERF","type=f32,ne_a=[5,7,11,13],v=1","support","1","yes","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne_a=[5,7,11,13],v=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne_a=[5,7,11,13],v=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne_a=[128,2,2,2],v=1","support","1","yes","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne_a=[5,7,11,13],v=1","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne_a=[5,7,11,13],v=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f32,ne_a=[128,2,2,2],v=1","support","0","no","WebGPU"
@@ -965,6 +965,7 @@
 "WebGPU: WebGPU","IM2COL","type_input=f32,type_kernel=f16,dst_type=f16,ne_input=[12,12,1,2560],ne_kernel=[3,3,1,2560],s0=1,s1=1,p0=1,p1=1,d0=1,d1=1,is_2D=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","IM2COL","type_input=f32,type_kernel=f16,dst_type=f16,ne_input=[12,12,2,2560],ne_kernel=[3,3,2,2560],s0=1,s1=1,p0=1,p1=1,d0=1,d1=1,is_2D=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","IM2COL","type_input=f32,type_kernel=f16,dst_type=f16,ne_input=[5,5,1,32],ne_kernel=[3,4,1,32],s0=1,s1=1,p0=0,p1=0,d0=1,d1=1,is_2D=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","IM2COL","type_input=f32,type_kernel=f32,dst_type=f32,ne_input=[2,2,1536,729],ne_kernel=[2,2,1536,4096],s0=1,s1=1,p0=0,p1=0,d0=1,d1=1,is_2D=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","IM2COL_3D","type_input=f32,type_kernel=f32,dst_type=f32,ne_input=[10,10,10,9],ne_kernel=[3,3,3,1],IC=3,s0=1,s1=1,s2=1,p0=1,p1=1,p2=1,d0=1,d1=1,d2=1,v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","IM2COL_3D","type_input=f32,type_kernel=f16,dst_type=f32,ne_input=[10,10,10,9],ne_kernel=[3,3,3,1],IC=3,s0=1,s1=1,s2=1,p0=1,p1=1,p2=1,d0=1,d1=1,d2=1,v=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","IM2COL_3D","type_input=f32,type_kernel=f16,dst_type=f16,ne_input=[10,10,10,9],ne_kernel=[3,3,3,1],IC=3,s0=1,s1=1,s2=1,p0=1,p1=1,p2=1,d0=1,d1=1,d2=1,v=0","support","0","no","WebGPU"
@@ -4964,6 +4965,7 @@
 "WebGPU: WebGPU","CONV_TRANSPOSE_1D","ne_input=[2,1,1,1],ne_kernel=[3,1,1,1],s0=1,p0=0,d0=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","CONV_TRANSPOSE_2D","ne_input=[3,2,3,1],ne_kernel=[2,2,1,3],stride=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","CONV_TRANSPOSE_2D","ne_input=[10,10,9,1],ne_kernel=[3,3,1,9],stride=2","support","0","no","WebGPU"
+"WebGPU: WebGPU","CONV_TRANSPOSE_2D","ne_input=[129,63,35,1],ne_kernel=[3,3,48,35],stride=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","COUNT_EQUAL","type=f32,ne=[4,500,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","COUNT_EQUAL","type=f32,ne=[4,5000,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","ARGMAX","type=f32,ne=[32,1,1,1]","support","0","no","WebGPU"
@@ -5715,15 +5717,15 @@
 "WebGPU: WebGPU","L2_NORM","type=f32,ne=[64,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","RMS_NORM","type=f32,ne=[64,5,4,3],v=0,eps=0.000001,inplace=1","support","1","yes","WebGPU"
 "WebGPU: WebGPU","L2_NORM","type=f32,ne=[64,5,4,3]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,1024,1,1],ne_b=[3,1024,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[8,1024,1,1],ne_b=[3,1024,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,1024,4,1],ne_b=[3,1024,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,1536,1,1],ne_b=[3,1536,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[8,1536,1,1],ne_b=[3,1536,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,1536,4,1],ne_b=[3,1536,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,2048,1,1],ne_b=[3,2048,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[8,2048,1,1],ne_b=[3,2048,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,2048,4,1],ne_b=[3,2048,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[3,1024,1,1],ne_b=[3,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[6,1024,1,1],ne_b=[3,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[3,1024,4,1],ne_b=[3,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[3,1536,1,1],ne_b=[3,1536,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[6,1536,1,1],ne_b=[3,1536,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[3,1536,4,1],ne_b=[3,1536,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[3,2048,1,1],ne_b=[3,2048,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[6,2048,1,1],ne_b=[3,2048,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[3,2048,4,1],ne_b=[3,2048,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,1024,1,1],ne_b=[4,1024,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[8,1024,1,1],ne_b=[4,1024,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,1024,4,1],ne_b=[4,1024,1,1]","support","0","no","WebGPU"
@@ -5733,6 +5735,15 @@
 "WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,2048,1,1],ne_b=[4,2048,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[8,2048,1,1],ne_b=[4,2048,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[4,2048,4,1],ne_b=[4,2048,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[9,1024,1,1],ne_b=[9,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[18,1024,1,1],ne_b=[9,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[9,1024,4,1],ne_b=[9,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[9,1536,1,1],ne_b=[9,1536,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[18,1536,1,1],ne_b=[9,1536,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[9,1536,4,1],ne_b=[9,1536,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[9,2048,1,1],ne_b=[9,2048,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[18,2048,1,1],ne_b=[9,2048,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SSM_CONV","type=f32,ne_a=[9,2048,4,1],ne_b=[9,2048,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_SCAN","type=f32,d_state=16,head_dim=1,n_head=1024,n_group=1,n_seq_tokens=32,n_seqs=4","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_SCAN","type=f32,d_state=128,head_dim=64,n_head=16,n_group=2,n_seq_tokens=32,n_seqs=4","support","0","no","WebGPU"
 "WebGPU: WebGPU","SSM_SCAN","type=f32,d_state=256,head_dim=64,n_head=8,n_group=2,n_seq_tokens=32,n_seqs=4","support","0","no","WebGPU"
@@ -8662,7 +8673,7 @@
 "WebGPU: WebGPU","CLAMP","type=f16,ne=[10,5,4,3],min=-0.500000,max=0.500000","support","0","no","WebGPU"
 "WebGPU: WebGPU","LEAKY_RELU","type=f16,ne_a=[10,5,4,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne=[10,2,2,2]","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne=[10,2,2,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne=[10,2,2,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne=[10,2,2,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SQR","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
@@ -8674,8 +8685,8 @@
 "WebGPU: WebGPU","LEAKY_RELU","type=f16,ne_a=[7,1,5,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne=[1024,1024,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f16,ne=[1024,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne=[1024,1024,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
@@ -8688,7 +8699,7 @@
 "WebGPU: WebGPU","CLAMP","type=f32,ne=[10,5,4,3],min=-0.500000,max=0.500000","support","0","no","WebGPU"
 "WebGPU: WebGPU","LEAKY_RELU","type=f32,ne_a=[10,5,4,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne=[10,2,2,2]","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne=[10,2,2,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne=[10,2,2,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f32,ne=[10,2,2,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SQR","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
@@ -8700,8 +8711,8 @@
 "WebGPU: WebGPU","LEAKY_RELU","type=f32,ne_a=[7,1,5,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne=[1024,1024,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
-"WebGPU: WebGPU","CEIL","type=f32,ne=[1024,1024,1,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","CEIL","type=f32,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne=[1024,1024,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
@@ -8916,6 +8927,8 @@
 "WebGPU: WebGPU","SOFT_MAX","type=f32,ne=[32,2,32,1],mask=1,sinks=0,m_prec=f16,nr23=[1,1],scale=0.100000,max_bias=0.000000,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","SOFT_MAX","type=f32,ne=[32,2,32,1],mask=1,sinks=1,m_prec=f32,nr23=[1,1],scale=0.100000,max_bias=8.000000,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","SOFT_MAX","type=f32,ne=[32,2,32,1],mask=1,sinks=1,m_prec=f16,nr23=[1,1],scale=0.100000,max_bias=8.000000,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SOFT_MAX","type=f32,ne=[200001,2,3,1],mask=1,sinks=1,m_prec=f32,nr23=[1,1],scale=0.100000,max_bias=8.000000,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SOFT_MAX","type=f32,ne=[200001,2,3,1],mask=1,sinks=1,m_prec=f16,nr23=[1,1],scale=0.100000,max_bias=8.000000,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","SOFT_MAX_BACK","type=f32,ne=[16,16,1,1],scale=1.000000,max_bias=0.000000","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOFT_MAX_BACK","type=f32,ne=[15,15,1,1],scale=1.000000,max_bias=0.000000","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOFT_MAX_BACK","type=f32,ne=[16,16,2,3],scale=1.000000,max_bias=0.000000","support","0","no","WebGPU"
@@ -8968,6 +8981,7 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
@@ -8977,6 +8991,7 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
@@ -8987,11 +9002,13 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
@@ -9001,6 +9018,7 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
@@ -9011,11 +9029,13 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
@@ -9025,6 +9045,7 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
@@ -9035,11 +9056,13 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
@@ -9049,6 +9072,7 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
@@ -9059,6 +9083,7 @@
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
+"WebGPU: WebGPU","ROPE","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f16,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROPE","type=f16,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","1","yes","WebGPU"
@@ -9184,6 +9209,7 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
@@ -9193,6 +9219,7 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
@@ -9203,11 +9230,13 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
@@ -9217,6 +9246,7 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
@@ -9227,11 +9257,13 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
@@ -9241,6 +9273,7 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
@@ -9251,11 +9284,13 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,40,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,52,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,64,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,1,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,71,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,8,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
@@ -9265,6 +9300,7 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=20,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,2,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,32,4,1],n_dims=32,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=128,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,12,2,1],n_dims=20,mode=8,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
@@ -9275,6 +9311,7 @@
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,28,2,1],n_dims=32,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[80,16,2,1],n_dims=80,mode=24,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[128,16,2,1],n_dims=128,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[16,16,8192,1],n_dims=16,mode=40,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f32,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=1,v=1,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f16,ne_a=[128,32,2,1],n_dims=128,mode=0,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROPE_BACK","type=f16,ne_a=[64,128,2,1],n_dims=64,mode=2,n_ctx=512,fs=1.000000,ef=0.000000,af=1.000000,ff=0,v=0,inplace=0","support","0","no","WebGPU"
@@ -9542,333 +9579,333 @@
 "WebGPU: WebGPU","ARGSORT","type=f32,ne=[2048,2,1,3],order=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","ARGSORT","type=f32,ne=[2049,2,1,3],order=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","ARGSORT","type=f32,ne=[2,8,8192,1],order=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[12,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[13,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[13,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[15,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[15,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[15,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=100","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=500","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=1023","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=9999","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=2","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=3","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=7","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=15","support","0","no","WebGPU"
-"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=15","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[12,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[13,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[13,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[15,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[15,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[15,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[19,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[27,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[43,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[64,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[75,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[128,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[139,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[256,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[267,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[512,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[523,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1035,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2059,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4096,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[4107,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8192,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[8203,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16395,1,2,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32768,1,1,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[32779,1,2,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65536,1,1,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[65547,1,2,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131072,1,1,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[131083,1,2,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262144,1,1,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[262155,1,2,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=100,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=500,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=1023,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524288,1,1,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[524299,1,2,1],k=9999,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=1,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=2,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=3,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=7,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16,10,10,10],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[60,10,10,10],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1023,2,1,3],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1024,2,1,3],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[1025,2,1,3],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[16384,1,1,1],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2047,2,1,3],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2048,2,1,3],k=15,ties=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","TOP_K","type=f32,ne=[2049,2,1,3],k=15,ties=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=nearest,transpose=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=nearest,transpose=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=nearest,flags=none","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=nearest,flags=none","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=nearest","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=nearest","support","0","no","WebGPU"
 "WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=bilinear,transpose=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=bilinear,transpose=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bilinear,flags=none","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=bilinear,flags=none","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bilinear","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=bilinear","support","0","no","WebGPU"
 "WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=bicubic,transpose=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=bicubic,transpose=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bicubic,flags=none","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=bicubic,flags=none","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=513,transpose=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=513,transpose=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bilinear,flags=none","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=bilinear,flags=none","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bilinear,flags=align_corners","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[1,4,3,2],ne_tgt=[2,8,3,2],mode=bilinear,flags=align_corners","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[4,1,3,2],ne_tgt=[1,1,3,2],mode=bilinear,flags=align_corners","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bicubic,flags=align_corners","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[1,4,3,2],ne_tgt=[2,8,3,2],mode=bicubic,flags=align_corners","support","0","no","WebGPU"
-"WebGPU: WebGPU","UPSCALE","type=f32,ne=[4,1,3,2],ne_tgt=[1,1,3,2],mode=bicubic,flags=align_corners","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bicubic","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=bicubic","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=bilinear|antialias,transpose=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[512,512,3,2],scale_factor=2,mode=bilinear|antialias,transpose=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bilinear|antialias","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[5,7,11,13],ne_tgt=[2,5,7,11],mode=bilinear|antialias","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bilinear|align_corners","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[1,4,3,2],ne_tgt=[2,8,3,2],mode=bilinear|align_corners","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[4,1,3,2],ne_tgt=[1,1,3,2],mode=bilinear|align_corners","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[2,5,7,11],ne_tgt=[5,7,11,13],mode=bicubic|align_corners","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[1,4,3,2],ne_tgt=[2,8,3,2],mode=bicubic|align_corners","support","0","no","WebGPU"
+"WebGPU: WebGPU","UPSCALE","type=f32,ne=[4,1,3,2],ne_tgt=[1,1,3,2],mode=bicubic|align_corners","support","0","no","WebGPU"
 "WebGPU: WebGPU","SUM","type=f32,ne=[10,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SUM_ROWS","type=f32,ne=[10,5,4,3],permute=0,slice=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","SUM","type=f32,ne=[11,5,6,3],permute=[0,2,1,3]","support","0","no","WebGPU"
@@ -9891,8 +9928,9 @@
 "WebGPU: WebGPU","GROUP_NORM","type=f32,ne=[64,64,320,1],num_groups=32,eps=0.000001","support","0","no","WebGPU"
 "WebGPU: WebGPU","GROUP_NORM","type=f32,ne=[9,9,1280,1],num_groups=32,eps=0.000001","support","0","no","WebGPU"
 "WebGPU: WebGPU","ACC","type=f32,ne_a=[256,17,1,1],ne_b=[256,16,1,1]","support","0","no","WebGPU"
-"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],pad_0=1,pad_1=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,3,1],lp0=1,rp0=1,lp1=1,rp1=1,lp2=1,rp2=1,lp3=1,rp3=1,v=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],pad_0=1,pad_1=1,circular=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[33,17,2,1],pad_0=4,pad_1=3,circular=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,3,1],lp0=1,rp0=1,lp1=1,rp1=1,lp2=1,rp2=1,lp3=1,rp3=1,v=0,circular=0","support","0","no","WebGPU"
 "WebGPU: WebGPU","PAD_REFLECT_1D","type=f32,ne_a=[512,34,2,1],pad_0=10,pad_1=9","support","0","no","WebGPU"
 "WebGPU: WebGPU","PAD_REFLECT_1D","type=f32,ne_a=[3000,384,4,1],pad_0=10,pad_1=9","support","0","no","WebGPU"
 "WebGPU: WebGPU","ROLL","shift0=3,shift1=-2,shift3=1,shift4=-1","support","0","no","WebGPU"
@@ -9903,6 +9941,7 @@
 "WebGPU: WebGPU","CUMSUM","type=f32,ne=[10,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CUMSUM","type=f32,ne=[127,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CUMSUM","type=f32,ne=[128,5,4,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","CUMSUM","type=f32,ne=[128,128,4,4]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CUMSUM","type=f32,ne=[255,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CUMSUM","type=f32,ne=[256,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CUMSUM","type=f32,ne=[511,5,4,3]","support","0","no","WebGPU"
@@ -9922,17 +9961,41 @@
 "WebGPU: WebGPU","FILL","type=f32,ne=[303,207,11,3],c=2.000000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FILL","type=f32,ne=[800,600,4,4],c=-152.000000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FILL","type=f32,ne=[2048,512,2,2],c=3.500000","support","0","no","WebGPU"
+"WebGPU: WebGPU","DIAG","type=f32,ne=[10,1,4,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","DIAG","type=f32,ne=[79,1,19,13]","support","0","no","WebGPU"
+"WebGPU: WebGPU","DIAG","type=f32,ne=[256,1,8,16]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[10,10,4,3],ne_rhs=[3,10,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[11,11,1,1],ne_rhs=[5,11,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[17,17,2,4],ne_rhs=[9,17,2,4]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[30,30,7,1],ne_rhs=[8,30,7,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[42,42,5,2],ne_rhs=[10,42,5,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[64,64,2,2],ne_rhs=[10,64,2,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[64,64,2,2],ne_rhs=[64,64,2,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[79,79,5,3],ne_rhs=[417,79,5,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[128,128,4,2],ne_rhs=[32,128,4,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[80,80,2,8],ne_rhs=[80,80,2,8]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[80,80,2,8],ne_rhs=[79,80,2,8]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[80,80,2,8],ne_rhs=[81,80,2,8]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[80,80,8,8],ne_rhs=[80,80,8,8]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[80,80,8,8],ne_rhs=[79,80,8,8]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[80,80,8,8],ne_rhs=[81,80,8,8]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[84,84,4,4],ne_rhs=[32,84,4,4]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[95,95,8,8],ne_rhs=[40,95,8,8]","support","0","no","WebGPU"
 "WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[100,100,4,4],ne_rhs=[41,100,4,4]","support","0","no","WebGPU"
-"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],lp0=0,rp0=1,lp1=0,rp1=1,lp2=0,rp2=0,lp3=0,rp3=0,v=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","PAD","type=f32,ne_a=[11,22,33,44],lp0=1,rp0=2,lp1=3,rp1=4,lp2=5,rp2=6,lp3=7,rp3=8,v=0","support","0","no","WebGPU"
-"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],lp0=0,rp0=1,lp1=0,rp1=1,lp2=0,rp2=0,lp3=0,rp3=0,v=1","support","0","no","WebGPU"
-"WebGPU: WebGPU","PAD","type=f32,ne_a=[11,22,33,44],lp0=1,rp0=2,lp1=3,rp1=4,lp2=5,rp2=6,lp3=7,rp3=8,v=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[128,128,4,4],ne_rhs=[31,128,4,4]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[128,128,4,4],ne_rhs=[32,128,4,4]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[128,128,3,4],ne_rhs=[32,128,3,4]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[128,128,4,1],ne_rhs=[32,128,4,1]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[64,64,4,4],ne_rhs=[200,64,4,4]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SOLVE_TRI","type=f32,ne_lhs=[64,64,4,4],ne_rhs=[384,64,4,4]","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],lp0=0,rp0=1,lp1=0,rp1=1,lp2=0,rp2=0,lp3=0,rp3=0,v=0,circular=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[11,22,33,44],lp0=1,rp0=2,lp1=3,rp1=4,lp2=5,rp2=6,lp3=7,rp3=8,v=0,circular=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],lp0=0,rp0=1,lp1=0,rp1=1,lp2=0,rp2=0,lp3=0,rp3=0,v=0,circular=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[11,22,33,44],lp0=1,rp0=2,lp1=3,rp1=4,lp2=5,rp2=6,lp3=7,rp3=8,v=0,circular=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],lp0=0,rp0=1,lp1=0,rp1=1,lp2=0,rp2=0,lp3=0,rp3=0,v=1,circular=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[11,22,33,44],lp0=1,rp0=2,lp1=3,rp1=4,lp2=5,rp2=6,lp3=7,rp3=8,v=1,circular=0","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[512,512,1,1],lp0=0,rp0=1,lp1=0,rp1=1,lp2=0,rp2=0,lp3=0,rp3=0,v=1,circular=1","support","0","no","WebGPU"
+"WebGPU: WebGPU","PAD","type=f32,ne_a=[11,22,33,44],lp0=1,rp0=2,lp1=3,rp1=4,lp2=5,rp2=6,lp3=7,rp3=8,v=1,circular=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLASH_ATTN_EXT","hsk=40,hsv=40,nh=4,nr23=[1,1],kv=113,nb=1,mask=1,sinks=1,max_bias=0.000000,logit_softcap=0.000000,prec=f32,type_KV=f32,permute=[0,1,2,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLASH_ATTN_EXT","hsk=40,hsv=40,nh=4,nr23=[1,1],kv=113,nb=1,mask=1,sinks=1,max_bias=0.000000,logit_softcap=0.000000,prec=f32,type_KV=f16,permute=[0,1,2,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLASH_ATTN_EXT","hsk=40,hsv=40,nh=4,nr23=[1,1],kv=113,nb=1,mask=1,sinks=1,max_bias=0.000000,logit_softcap=0.000000,prec=f32,type_KV=bf16,permute=[0,1,2,3]","support","0","no","WebGPU"
diff --git a/ggml/src/ggml-webgpu/ggml-webgpu.cpp b/ggml/src/ggml-webgpu/ggml-webgpu.cpp
index d0e99b6fe29eb16c03e0927fd85b31bdc291a97e..c7afdfb8e92f3700d51052f1591fdfbf90bdac94 100644 (file)
--- a/ggml/src/ggml-webgpu/ggml-webgpu.cpp
+++ b/ggml/src/ggml-webgpu/ggml-webgpu.cpp
@@ -2273,6 +2273,16 @@ static void ggml_webgpu_init_unary_pipeline(webgpu_context & webgpu_ctx) {
         ggml_webgpu_create_pipeline(webgpu_ctx->device, wgsl_xielu_inplace_f32, "xielu_inplace_f32", constants);
     webgpu_ctx->unary_pipelines[GGML_UNARY_OP_XIELU][GGML_TYPE_F16][1] =
         ggml_webgpu_create_pipeline(webgpu_ctx->device, wgsl_xielu_inplace_f16, "xielu_inplace_f16", constants);
+
+    // CEIL
+    webgpu_ctx->unary_pipelines[GGML_UNARY_OP_CEIL][GGML_TYPE_F32][0] =
+        ggml_webgpu_create_pipeline(webgpu_ctx->device, wgsl_ceil_f32, "ceil_f32", constants);
+    webgpu_ctx->unary_pipelines[GGML_UNARY_OP_CEIL][GGML_TYPE_F16][0] =
+        ggml_webgpu_create_pipeline(webgpu_ctx->device, wgsl_ceil_f16, "ceil_f16", constants);
+    webgpu_ctx->unary_pipelines[GGML_UNARY_OP_CEIL][GGML_TYPE_F32][1] =
+        ggml_webgpu_create_pipeline(webgpu_ctx->device, wgsl_ceil_inplace_f32, "ceil_inplace_f32", constants);
+    webgpu_ctx->unary_pipelines[GGML_UNARY_OP_CEIL][GGML_TYPE_F16][1] =
+        ggml_webgpu_create_pipeline(webgpu_ctx->device, wgsl_ceil_inplace_f16, "ceil_inplace_f16", constants);
 }
 
 static void ggml_webgpu_init_scale_pipeline(webgpu_context & webgpu_ctx) {
@@ -2528,6 +2538,7 @@ static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const
                     case GGML_UNARY_OP_EXP:
                     case GGML_UNARY_OP_GELU_ERF:
                     case GGML_UNARY_OP_XIELU:
+                    case GGML_UNARY_OP_CEIL:
                         supports_op = supports_op =
                             (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
                         break;
diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl
index d474ab107b42f4d40671e2238dfc2de2c0c55234..25fe2854518005e8bcef2d12d6fe4da4b0a9d531 100644 (file)
--- a/ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl
+++ b/ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl
@@ -16,7 +16,8 @@
     "HARDSWISH_FUNC": "{{MUTATE}}[dst_i] = src[src_i] * min(1.0, max(0.0, (src[src_i] + 3.0) / 6.0));",
     "GELU_FUNC": "{{MUTATE}}[dst_i] = 0.5 * src[src_i] * (1.0 + tanh(clamp(sqrt(2.0 / 3.14159265) * (src[src_i] + 0.044715 * pow(src[src_i], 3.0)), -9.010913, 9.010913))); // Regarding tanh() domain restrictions in wgsl https://github.com/gpuweb/gpuweb/issues/4458",
     "GELU_QUICK_FUNC": "{{MUTATE}}[dst_i] = src[src_i] * 0.5 * (1.0 + tanh(clamp(0.79788456 * (src[src_i] + 0.044715 * src[src_i] * src[src_i] * src[src_i]), -9.010913, 9.010913))); // Regarding tanh() domain restrictions in wgsl https://github.com/gpuweb/gpuweb/issues/4458",
-    "GELU_ERF_FUNC": "{{MUTATE}}[dst_i] = 0.5 * src[src_i] * (1.0 + tanh(clamp(0.79788456 * (src[src_i] + 0.044715 * src[src_i] * src[src_i] * src[src_i]), -9.010913, 9.010913))); // Regarding tanh() domain restrictions in wgsl https://github.com/gpuweb/gpuweb/issues/4458"
+    "GELU_ERF_FUNC": "{{MUTATE}}[dst_i] = 0.5 * src[src_i] * (1.0 + tanh(clamp(0.79788456 * (src[src_i] + 0.044715 * src[src_i] * src[src_i] * src[src_i]), -9.010913, 9.010913))); // Regarding tanh() domain restrictions in wgsl https://github.com/gpuweb/gpuweb/issues/4458",
+    "CEIL_FUNC": "{{MUTATE}}[dst_i] = ceil(src[src_i]);"
 }
 
 #end(REPL_TEMPLATES)
@@ -357,6 +358,27 @@
         "SHADER_NAME": "gelu_erf_inplace_f16",
         "REPLS": { "TYPE": "f16", "FUNC": "GELU_ERF_FUNC", "EXT_PARAMS": "", "MUTATE": "src" },
         "DECLS": ["INPLACE"]
+    },
+
+    {
+        "SHADER_NAME": "ceil_f32",
+        "REPLS": { "TYPE": "f32", "FUNC": "CEIL_FUNC", "EXT_PARAMS": "", "MUTATE": "dst" },
+        "DECLS": ["NOT_INPLACE"]
+    },
+    {
+        "SHADER_NAME": "ceil_f16",
+        "REPLS": { "TYPE": "f16", "FUNC": "CEIL_FUNC", "EXT_PARAMS": "", "MUTATE": "dst" },
+        "DECLS": ["NOT_INPLACE"]
+    },
+    {
+        "SHADER_NAME": "ceil_inplace_f32",
+        "REPLS": { "TYPE": "f32", "FUNC": "CEIL_FUNC", "EXT_PARAMS": "", "MUTATE": "src" },
+        "DECLS": ["INPLACE"]
+    },
+    {
+        "SHADER_NAME": "ceil_inplace_f16",
+        "REPLS": { "TYPE": "f16", "FUNC": "CEIL_FUNC", "EXT_PARAMS": "", "MUTATE": "src" },
+        "DECLS": ["INPLACE"]
     }
 ]
 
Packaging of ggml-org/llama.cpp