sirhcm/tinygrad
1
0
Fork 0
forked from tinygrad/tinygrad
Code Pull Requests 2 Actions Activity

run llvmspeed Speed Test (BEAM=2) without cache (#17869)

Browse Source 
* run llvmspeed Speed Test (BEAM=2) without cache

* HEVC too
This commit is contained in:
chenyu
2026-08-31 20:16:32 -04:00
committed by GitHub
parent 79250d76de
commit c12eefe447
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/benchmark.yml
+4 -4
View File
@@ -403,7 +403,7 @@ jobs:
run: python test/external/external_benchmark_multitensor_allreduce.py
- name: HEVC Decode Benchmark
if: ${{ matrix.dev == 'NV' }}
run: VALIDATE=1 MAX_FRAMES=100 ASSERT_FPS=1400 JITBEAM=1 PYTHONPATH=. python3 extra/hevc/decode.py
run: IGNORE_BEAM_CACHE=1 VALIDATE=1 MAX_FRAMES=100 ASSERT_FPS=1400 JITBEAM=1 PYTHONPATH=. python3 extra/hevc/decode.py
- uses: actions/upload-artifact@v7
if: ${{ matrix.dev != 'AMD' }}
with:
@@ -619,7 +619,7 @@ jobs:
GRAPH_ONE_KERNEL=1 NSZ=8192 python3 test/speed/external_test_copy_speed.py TestCopySpeed.testCopyCPUtoDefaultJit
- name: HEVC Decode Benchmark
if: ${{ matrix.dev == 'NV' }}
run: VALIDATE=1 MAX_FRAMES=100 ASSERT_FPS=1400 JITBEAM=1 PYTHONPATH=. python3 extra/hevc/decode.py
run: IGNORE_BEAM_CACHE=1 VALIDATE=1 MAX_FRAMES=100 ASSERT_FPS=1400 JITBEAM=1 PYTHONPATH=. python3 extra/hevc/decode.py
- name: Run 10 MLPerf ResNet50 training steps (1 gpu)
if: ${{ matrix.dev == 'NV' }}
run: BENCHMARK_LOG=resnet_10steps MNISTMOCK=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=256 GPUS=1 MODEL=resnet python3 examples/mlperf/model_train.py
@@ -642,7 +642,7 @@ jobs:
llvmspeed:
name: LLVM Speed
runs-on: [self-hosted, Linux, tinyboxrandom]
timeout-minutes: 5
timeout-minutes: 10
if: github.repository_owner == 'tinygrad'
steps:
- name: Checkout Code
@@ -650,4 +650,4 @@ jobs:
- name: Speed Test
run: DEV=CPU:LLVM python3 test/speed/external_test_speed_v_torch.py
- name: Speed Test (BEAM=2)
run: BEAM=2 DEV=CPU:LLVM python3 test/speed/external_test_speed_v_torch.py
run: IGNORE_BEAM_CACHE=1 BEAM=2 DEV=CPU:LLVM python3 test/speed/external_test_speed_v_torch.py