github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-02-17 10:02:00 -05:00
Code Issues Packages Projects Releases Wiki Activity

tf32 tc for nv and ptx (#8635)

Browse Source 
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
This commit is contained in:
ignaciosica
2025-01-17 22:43:57 -03:00
committed by GitHub
parent 5afb0a4a81
commit d2234e308a
7 changed files with 23 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
.github/workflows/benchmark.yml vendored
View File

@@ -172,12 +172,13 @@ jobs:
run: NV=1 python test/external/external_benchmark_multitensor_allreduce.py
- name: Test tensor cores
run: |
NV=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded
PTX=1 NV=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded
NV=1 ALLOW_TF32=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded
PTX=1 ALLOW_TF32=1 NV=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded
- name: Run Tensor Core GEMM (CUDA)
run: |
CUDA=1 HALF=1 DEBUG=2 python3 extra/gemm/simple_matmul.py | tee matmul.txt
CUDA=1 BFLOAT16=1 DEBUG=2 python3 extra/gemm/simple_matmul.py | tee matmul_bfloat16.txt
CUDA=1 ALLOW_TF32=1 DEBUG=2 python3 extra/gemm/simple_matmul.py | tee matmul_tf32.txt
- name: Run Tensor Core GEMM (PTX)
run: NV=1 PTX=1 HALF=1 DEBUG=2 python3 extra/gemm/simple_matmul.py | tee matmul_ptx.txt
- name: Run Tensor Core GEMM (NV)
@@ -226,6 +227,7 @@ jobs:
torch_speed.txt
matmul.txt
matmul_bfloat16.txt
matmul_tf32.txt
matmul_ptx.txt
matmul_nv.txt
sd.txt