llama.cpp/examples/sycl
2024-05-24 10:06:56 +08:00
..
build.sh fix memcpy() crash, add missed cmd in guide, fix softmax (#6622) 2024-04-14 10:42:29 +08:00
CMakeLists.txt ggml : add unified SYCL backend for Intel GPUs (#2690) 2024-01-28 17:56:23 +02:00
ls-sycl-device.cpp Support multiple GPUs (split mode) on SYCL backend (#5806) 2024-03-02 19:49:30 +08:00
README.md docs: fix typos (#7124) 2024-05-07 18:20:33 +03:00
run-llama2.sh fix memcpy() crash, add missed cmd in guide, fix softmax (#6622) 2024-04-14 10:42:29 +08:00
win-build-sycl.bat add build shared lib in win release package (#7438) 2024-05-24 10:06:56 +08:00
win-run-llama2.bat update readme sycl for new update (#6151) 2024-03-20 11:21:41 +08:00

llama.cpp/example/sycl

This example program provides the tools for llama.cpp for SYCL on Intel GPU.

Tool

Tool Name Function Status
ls-sycl-device List all SYCL devices with ID, compute capability, max work group size, ect. Support

ls-sycl-device

List all SYCL devices with ID, compute capability, max work group size, ect.

  1. Build the llama.cpp for SYCL for all targets.

  2. Enable oneAPI running environment

source /opt/intel/oneapi/setvars.sh
  1. Execute
./build/bin/ls-sycl-device

Check the ID in startup log, like:

found 4 SYCL devices:
  Device 0: Intel(R) Arc(TM) A770 Graphics,	compute capability 1.3,
    max compute_units 512,	max work group size 1024,	max sub group size 32,	global mem size 16225243136
  Device 1: Intel(R) FPGA Emulation Device,	compute capability 1.2,
    max compute_units 24,	max work group size 67108864,	max sub group size 64,	global mem size 67065057280
  Device 2: 13th Gen Intel(R) Core(TM) i7-13700K,	compute capability 3.0,
    max compute_units 24,	max work group size 8192,	max sub group size 64,	global mem size 67065057280
  Device 3: Intel(R) Arc(TM) A770 Graphics,	compute capability 3.0,
    max compute_units 512,	max work group size 1024,	max sub group size 32,	global mem size 16225243136

Attribute Note
compute capability 1.3 Level-zero running time, recommended
compute capability 3.0 OpenCL running time, slower than level-zero in most cases