]> git.djapps.eu Git - pkg/ggml/sources/ggml/shortlog
pkg/ggml/sources/ggml
2025-12-11 Jiacheng (Jason... HIP: Patch failed testcase in WMMA-MMQ kernels for...
2025-12-11 hipuddingCANN: Add MROPE and IMROPE support (llama/17401)
2025-12-11 Jeff Bolzvulkan: Implement GGML_OP_CUMSUM (llama/17479)
2025-12-11 Georgi Gerganovggml : add ggml_top_k (llama/17365)
2025-12-11 TianHao324CANN: supports out_prod operator for F32 and F16 (llama...
2025-12-11 Jeff Bolzvulkan: Use fewer rows for scalar FA when HS is not...
2025-12-11 Jeff Bolzvulkan: more FA details in vk_perf_logger (llama/17443)
2025-12-11 Jiacheng (Jason... HIP: WMMA-MMQ kernels for RDNA 4 (llama/17156)
2025-12-11 Alberto Cabrera... ggml-cpu: arm64: q4_K repack gemm and gemv implementati...
2025-12-11 ixgbeggml: add RISC-V cpu-feats (llama/17461)
2025-12-11 Max Krasnyanskyhexagon: add support for ROPE_NEOX (llama/17458)
2025-12-11 Raul TorresCANN: Define `cann_graph_update_required` before macro...
2025-12-11 M. Mediouniggml-hexagon: Initial Hexagon v68/v69 support (llama...
2025-12-11 nullnameggml-hexagon: add `hex_supported_buffer` for better...
2025-12-11 Sigbjørn Skjæretcuda : support non-contiguous i32 to i32 copy (llama...
2025-12-11 Jeff Bolzvulkan: remove a couple unnecessary switches (llama...
2025-12-11 Masato NakasakaRevive MUL_MAT_ID to perf testing (llama/17397)
2025-12-11 yuloHIP: RDNA4 tensor core support for MMF (llama/17077)
2025-12-11 lhezopencl: refine condition for kqv mm (llama/17392)
2025-12-11 Jeff Bolzvulkan: disable async for older Intel devices (llama...
2025-12-11 Raul TorresCANN: Refactor `evaluate_and_capture_cann_graph` (llama...
2025-12-11 nullnameggml-hexagon: fix swiglu failure at `test-backend-ops...
2025-12-11 Piotr Wilkin... ggml : Fix transposed SOLVE_TRI result (llama/17323)
2025-12-11 Scott FudallyDGX Spark: UMA support (llama/17368)
2025-12-11 Adrien Gallouëtggml : remove useless and error-prone variadic macros...
2025-12-11 sudhiarmkleidiai: fix zero-size array declaration (llama/17240)
2025-12-11 ixgbeggml-cpu:add RISC-V RVV (Zvfh) optimization for FP16...
2025-12-11 Giuseppe Scrivanovulkan: implement ADD1, ARANGE, FILL, SOFTPLUS, STEP...
2025-12-11 Jeff Bolzvulkan: support larger argsort (llama/17313)
2025-12-11 Jeff Bolzvulkan: Add copy_transpose shader (llama/17371)
2025-12-11 Aman Guptacuda: fix rope fusion for gemma3 (llama/17378)
2025-12-11 Piotr Wilkin... Fix too relaxed check on CUDA "fast copy" (can_be_trans...
2025-12-11 Ruben Ortlamvulkan: force full subgroups for flash attention to...
2025-12-11 Jeremy Randggml-cpu: Don't pass -mpowerpc64 when -mcpu already...
2025-12-11 Chenguang LiCANN: fix acl_tensor_ptr usage in ASCEND_310P ROPE...
2025-12-11 Jeff Bolzvulkan: support noncontig i32 copy (llama/17328)
2025-12-11 Ruben Ortlamvulkan: add log RTE support to fix Nvidia CI (llama...
2025-12-11 Adrien Gallouëtcmake : fix ARM feature verification (llama/17170)
2025-12-11 Adrien Gallouëtggml : add missing AVX512 feature checks (llama/17270)
2025-11-24 Daniel Beveniusggml : remove dirty flag from version string (#1391)
2025-11-20 Georgi Gerganovsync : whisper.cpp
2025-11-20 YangLemetal : fix compile on macos 11 (whisper/3533)
2025-11-17 Georgi Gerganovsync : llama.cpp
2025-11-17 Georgi Gerganovmetal : support I32 -> I32 copy (llama/17317)
2025-11-17 Georgi Gerganovmetal : faster argsort (llama/17315)
2025-11-17 Georgi Gerganovmetal : add cumsum (llama/17305)
2025-11-17 hipuddingCANN: Use smart pointers to manage ACL objects (llama...
2025-11-17 Pavels Zaicenkovsvulkan: add LOG operation support for F32 and F16 ...
2025-11-17 Ruben Ortlamvulkan: fix MMQ quantize_y condition (llama/17301)
2025-11-17 Georgi Gerganovmetal : remove obosolete asserts (llama/17295)
2025-11-17 lhezopencl: fix rms_norm_mul (llama/17250)
2025-11-17 shaofeiqiopencl: add kernel to handle mat mul in attention to...
2025-11-17 shani-fsycl : unify unary kernels with a generic implementatio...
2025-11-17 Jeff Bolzvulkan: Fuse mul_mat_id+add_id+mul and mul_mat+add...
2025-11-17 Ruben Ortlamvulkan: Replace 16-bit unpack8 calls to work around...
2025-11-17 Giuseppe Scrivanovulkan: implement ABS and NEG (llama/17245)
2025-11-17 Jeff Bolzvulkan: Use ggml_vk_tensor_subbuffer in mul_mat_vec...
2025-11-17 Jeff Bolzvulkan: skip all-negative-inf blocks in FA (llama/17186)
2025-11-17 Jeff Bolzvulkan: change graph_compute to be async and enable...
2025-11-17 Georgi Gerganovmetal : support argsort for ne00 > 1024 (llama/17247)
2025-11-17 Georgi Gerganovmetal : make the FA extra sizes consistent (llama/17143)
2025-11-17 Alberto Cabrera... ggml-cpu: handle 3d tensors in repack mat_mul (llama...
2025-11-17 Piotr Wilkin... ggml : add ops SOFTPLUS, EXPM1, TRI, SOLVE_TRI, CUMSUM...
2025-11-17 Ruben Ortlamvulkan: remove shell call from vulkan-shaders-gen tool...
2025-11-17 Diego Devesasched : fix reserve ignoring user tensor assignments...
2025-11-17 ixgbeggml-cpu : add RISC-V vector intrinsic support for...
2025-11-17 bagheerametal: accelerated conv2d (llama/17175)
2025-11-17 Georgi GerganovRevert "ggml-cpu: handle 3d tensors in repack mat_mul...
2025-11-17 Diego Devesaggml-cpu : use template for argsort (llama/17222)
2025-11-17 TecJeshCANN: Add cross_entropy_loss op support (llama/16886)
2025-11-17 Aman GuptaCUDA: fuse rope + set_rows (llama/16884)
2025-11-17 Johannes GäßlerCUDA: static assert to prevent misuse of memcpy_1 ...
2025-11-17 Georgi Gerganovggml : use std::sort in ggml_argsort CPU implementation...
2025-11-17 Alberto Cabrera... ggml-cpu: handle 3d tensors in repack mat_mul (llama...
2025-11-17 TecJeshCANN: Add L2_NORM op support (llama/16856)
2025-11-17 Neo Zhang Jianyufix ci crash about SSM_CONV (llama/17169)
2025-11-17 Max Krasnyanskyhexagon: various Op fixes (llama/17135)
2025-11-17 Evedisable rms norm mul rope for chips with no fp16 rte...
2025-11-17 ixgbeggml-cpu : add RISC-V RVV (Zvfh) optimization for FP16...
2025-11-17 dudutaggml-cpu: templateify ggml_compute_forward_rope_f32...
2025-11-17 Charles Xukleidiai: add optimized per-channel kernels for Q8_0...
2025-11-17 Mike Abbottcmake : add version to all shared object files (llama...
2025-11-17 lhezopencl: add fastdiv and use it in set_rows, ported...
2025-11-17 Max Krasnyanskycpu: skip NOPs to avoid barriers (llama/17133)
2025-11-17 Georgi Gerganovmetal : cap threadgroups size of set_rows (llama/17146)
2025-11-17 Adrien Gallouëtggml-cpu : inspect -march and -mcpu to found the CPU...
2025-11-17 Ruben Ortlamvulkan: check glslc executable string (llama/17144)
2025-11-17 Ruben Ortlamvulkan: fix validation issue introduced by #16868 ...
2025-11-17 Georgi Gerganovmetal : enable tensor API for A19 (llama/17087)
2025-11-17 fj-y-saitoarm64: add i8mm route with SVE ggml_vec_dot_q4_K_q8_K...
2025-11-17 Aclycuda/vulkan : bicubic interpolation (llama/17022)
2025-11-17 Ruben Ortlamvulkan: fix memory allocations (llama/17122)
2025-11-11 Mathieu BaudierFix arm64 build debian/0.9.4.185-1
2025-11-11 Mathieu BaudierImprove Ubuntu ARM64 build
2025-11-11 Mathieu BaudierUpdate upstream and improve packaging
2025-11-11 Mathieu BaudierMerge tag 'upstream/0.9.4.185' into debian/latest
2025-11-09 Georgi Gerganovsync : whisper.cpp upstream/0.9.4.185
2025-11-09 Georgi Gerganovsync : llama.cpp
2025-11-09 Ruben Ortlamvulkan: iGPU memory reporting fix (llama/17110)
2025-11-09 Ruben Ortlamvulkan: fix mmq out of bounds reads (llama/17108)
next