llama.cpp/ggml/src/ggml-opencl/kernels
lhez 34a846b584
opencl: fix for small models (#11950)
* opencl: fix small shape gemv, remove unused extensions

* opencl: fix `transpose_16`, `dump_tensor`, enforce subgroup size

* opencl: fix for token length < 4

* opencl: use wave size of 64 for all Adreno GPUs

---------

Co-authored-by: Shawn Gu <quic_shawngu@quicinc.com>
Co-authored-by: Skyler Szot <quic_sszot@quicinc.com>
2025-02-24 14:47:07 -07:00
..
embed_kernel.py Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (#10693) 2024-12-13 12:23:52 -08:00
ggml-opencl.cl opencl: fix for small models (#11950) 2025-02-24 14:47:07 -07:00
ggml-opencl_cvt.cl Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (#10693) 2024-12-13 12:23:52 -08:00
ggml-opencl_gemv_noshuffle.cl opencl: fix for small models (#11950) 2025-02-24 14:47:07 -07:00
ggml-opencl_gemv_noshuffle_general.cl opencl: fix for small models (#11950) 2025-02-24 14:47:07 -07:00
ggml-opencl_mm.cl Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (#10693) 2024-12-13 12:23:52 -08:00
ggml-opencl_mul_mat_Ab_Bi_8x4.cl opencl: fix for small models (#11950) 2025-02-24 14:47:07 -07:00
ggml-opencl_transpose_16.cl opencl: fix for small models (#11950) 2025-02-24 14:47:07 -07:00
ggml-opencl_transpose_32.cl Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (#10693) 2024-12-13 12:23:52 -08:00
ggml-opencl_transpose_32_16.cl Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (#10693) 2024-12-13 12:23:52 -08:00