From: Peter Sugihara <redacted>
Date: Fri, 3 Nov 2023 19:18:18 +0000 (-0700)
Subject: metal : round up to 16 to fix MTLDebugComputeCommandEncoder assertion (#3938)
X-Git-Tag: upstream/0.0.4488~3004
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=d9b33fe95bd257b36c84ee5769cc048230067d6f;p=pkg%2Fggml%2Fsources%2Fllama.cpp

metal : round up to 16 to fix MTLDebugComputeCommandEncoder assertion (#3938)
---

diff --git a/ggml-metal.m b/ggml-metal.m
index acdb8384..78ae4485 100644
--- a/ggml-metal.m
+++ b/ggml-metal.m
@@ -1017,7 +1017,7 @@ void ggml_metal_graph_compute(
                             [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:2];
                             [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:3];
                             [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:4];
-                            [encoder setThreadgroupMemoryLength:nth/32*sizeof(float) atIndex:0];
+                            [encoder setThreadgroupMemoryLength:MAX(16, nth/32*sizeof(float)) atIndex:0];
 
                             [encoder dispatchThreadgroups:MTLSizeMake(ne01*ne02*ne03, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)];
                         } break;
@@ -1348,7 +1348,7 @@ void ggml_metal_graph_compute(
                             [encoder setBytes:&ne00    length:sizeof( int64_t) atIndex:2];
                             [encoder setBytes:&nb01    length:sizeof(uint64_t) atIndex:3];
                             [encoder setBytes:&eps     length:sizeof(   float) atIndex:4];
-                            [encoder setThreadgroupMemoryLength:nth*sizeof(float) atIndex:0];
+                            [encoder setThreadgroupMemoryLength:MAX(16, nth*sizeof(float)) atIndex:0];
 
                             const int64_t nrows = ggml_nrows(src0);