run bfloat16 tensor core in metal benchmark (#7808)

* run bfloat16 tensor core in metal benchmark

* separate task
This commit is contained in:
chenyu
2024-11-20 15:34:07 -05:00
committed by GitHub
parent 33a496279b
commit c815d7b56e
+7 -4
View File
@@ -60,10 +60,12 @@ jobs:
run: BIG=2 MPS=1 python3.11 test/test_speed_v_torch.py | tee torch_speed.txt
- name: Test tensor cores
run: METAL=1 python3.11 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded
- name: Run Tensor Core GEMM
run: |
DEBUG=2 python3.11 extra/gemm/simple_matmul.py | tee matmul.txt
DEBUG=2 HALF=1 python3.11 extra/gemm/simple_matmul.py | tee matmul_half.txt
- name: Run Tensor Core GEMM (float)
run: DEBUG=2 python3.11 extra/gemm/simple_matmul.py | tee matmul.txt
- name: Run Tensor Core GEMM (half)
run: DEBUG=2 HALF=1 python3.11 extra/gemm/simple_matmul.py | tee matmul_half.txt
- name: Run Tensor Core GEMM (bfloat16)
run: DEBUG=2 BFLOAT16=1 python3.11 extra/gemm/simple_matmul.py | tee matmul_bfloat16.txt
- name: Fuzz Padded Tensor Core GEMM
run: METAL=1 M_START=6 M_STOP=10 M_STEP=1 N_START=6 N_STOP=10 N_STEP=1 K_START=6 K_STOP=24 K_STEP=1 TC_OPT=2 DEBUG=2 python3.11 ./extra/gemm/fuzz_matmul.py
- name: Run LLaMA
@@ -114,6 +116,7 @@ jobs:
gpt2_half_beam.txt
matmul.txt
matmul_half.txt
matmul_bfloat16.txt
sd.txt
sd_fp16.txt
sdxl.txt