From 6327b50e51ad4478517e795f9ffd8f86cf2c5a22 Mon Sep 17 00:00:00 2001 From: nimlgen <138685161+nimlgen@users.noreply.github.com> Date: Sat, 8 Jun 2024 23:24:46 +0300 Subject: [PATCH] amd in benchmarks (#4861) * amd in benchmarks * remove all hsa --- .github/workflows/benchmark.yml | 41 +++++++++++++++------------------ 1 file changed, 19 insertions(+), 22 deletions(-) diff --git a/.github/workflows/benchmark.yml b/.github/workflows/benchmark.yml index 62c28a7faa..2ff721565b 100644 --- a/.github/workflows/benchmark.yml +++ b/.github/workflows/benchmark.yml @@ -285,35 +285,32 @@ jobs: # LD_PRELOAD="/opt/rocm/lib/libhsa-runtime64.so" HSA=1 BIG=2 TORCHCUDA=1 python3 test/test_speed_v_torch.py | tee torch_speed.txt - name: Test tensor cores run: | - HSA=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded AMD=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores TestLinearizer.test_tensor_cores_padded - - name: Run Tensor Core GEMM (HSA) - run: HSA=1 HALF=1 DEBUG=2 python3 extra/gemm/simple_matmul.py | tee matmul.txt - name: Run Tensor Core GEMM (AMD) run: AMD=1 HALF=1 DEBUG=2 python3 extra/gemm/simple_matmul.py | tee matmul_amd.txt # TODO: AMD compiler bug causes this to fail #- name: Fuzz Padded Tensor Core GEMM # run: HSA=1 M_START=12 M_STOP=20 M_STEP=1 N_START=12 N_STOP=20 N_STEP=1 K_START=28 K_STOP=36 K_STEP=1 HALF=1 TC_OPT=2 DEBUG=2 python3 ./extra/gemm/fuzz_matmul.py - name: Run Stable Diffusion - run: HSA=1 python3 examples/stable_diffusion.py --seed 0 --noshow --timing | tee sd.txt + run: AMD=1 python3 examples/stable_diffusion.py --seed 0 --noshow --timing | tee sd.txt - name: Run LLaMA 7B run: | - HSA=1 JIT=0 python3 examples/llama.py --gen 1 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_unjitted.txt - HSA=1 JIT=1 python3 examples/llama.py --gen 1 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_jitted.txt + AMD=1 JIT=0 python3 examples/llama.py --gen 1 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_unjitted.txt + AMD=1 JIT=1 python3 examples/llama.py --gen 1 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_jitted.txt - name: Run LLaMA 7B with BEAM - run: HSA=1 JIT=1 BEAM=2 CACHELEVEL=0 python3 examples/llama.py --gen 1 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_beam.txt + run: AMD=1 JIT=1 BEAM=2 CACHELEVEL=0 python3 examples/llama.py --gen 1 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_beam.txt - name: Run LLaMA 7B on 4 GPUs - run: HSA=1 python3 examples/llama.py --gen 1 --size 7B --shard 4 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_four_gpu.txt + run: AMD=1 python3 examples/llama.py --gen 1 --size 7B --shard 4 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_four_gpu.txt - name: Run LLaMA 7B on 6 GPUs - run: HSA=1 python3 examples/llama.py --gen 1 --size 7B --shard 6 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_six_gpu.txt + run: AMD=1 python3 examples/llama.py --gen 1 --size 7B --shard 6 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_six_gpu.txt - name: Run LLaMA-2 70B - run: HSA=1 python3 examples/llama.py --gen 2 --size 70B --shard 6 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_2_70B.txt + run: AMD=1 python3 examples/llama.py --gen 2 --size 70B --shard 6 --prompt "Hello." --count 10 --temperature 0 --timing | tee llama_2_70B.txt - name: Run Mixtral 8x7B - run: time HSA=1 python3 examples/mixtral.py --temperature 0 --count 10 --timing | tee mixtral.txt + run: time AMD=1 python3 examples/mixtral.py --temperature 0 --count 10 --timing | tee mixtral.txt - name: Run GPT2 run: | - HSA=1 JIT=0 python3 examples/gpt2.py --prompt "Hello." --count 10 --temperature 0 --timing | tee gpt2_unjitted.txt - HSA=1 JIT=1 python3 examples/gpt2.py --prompt "Hello." --count 10 --temperature 0 --timing | tee gpt2_jitted.txt + AMD=1 JIT=0 python3 examples/gpt2.py --prompt "Hello." --count 10 --temperature 0 --timing | tee gpt2_unjitted.txt + AMD=1 JIT=1 python3 examples/gpt2.py --prompt "Hello." --count 10 --temperature 0 --timing | tee gpt2_jitted.txt - uses: actions/upload-artifact@v4 with: name: Speed (AMD) @@ -356,23 +353,23 @@ jobs: mkdir -p extra/datasets ln -s /raid/datasets/imagenet extra/datasets/imagenet - name: Train MNIST - run: time PYTHONPATH=. HSA=1 TARGET_EVAL_ACC_PCT=97.3 python3 examples/beautiful_mnist.py | tee beautiful_mnist.txt + run: time PYTHONPATH=. AMD=1 TARGET_EVAL_ACC_PCT=97.3 python3 examples/beautiful_mnist.py | tee beautiful_mnist.txt - name: Run 10 CIFAR training steps - run: HSA=1 STEPS=10 python3 examples/hlb_cifar10.py | tee train_cifar.txt + run: AMD=1 STEPS=10 python3 examples/hlb_cifar10.py | tee train_cifar.txt - name: Run 10 CIFAR training steps w HALF - run: HSA=1 STEPS=10 DEFAULT_FLOAT=HALF python3 examples/hlb_cifar10.py | tee train_cifar_half.txt + run: AMD=1 STEPS=10 DEFAULT_FLOAT=HALF python3 examples/hlb_cifar10.py | tee train_cifar_half.txt - name: Run 10 CIFAR training steps w BF16 - run: HSA=1 STEPS=10 DEFAULT_FLOAT=BFLOAT16 python3 examples/hlb_cifar10.py | tee train_cifar_bf16.txt + run: AMD=1 STEPS=10 DEFAULT_FLOAT=BFLOAT16 python3 examples/hlb_cifar10.py | tee train_cifar_bf16.txt - name: Run full CIFAR training w 1 GPU - run: time HSA=1 DEFAULT_FLOAT=HALF LATEWINO=1 STEPS=1000 TARGET_EVAL_ACC_PCT=93.3 python3 examples/hlb_cifar10.py | tee train_cifar_one_gpu.txt + run: time AMD=1 DEFAULT_FLOAT=HALF LATEWINO=1 STEPS=1000 TARGET_EVAL_ACC_PCT=93.3 python3 examples/hlb_cifar10.py | tee train_cifar_one_gpu.txt - name: Run full CIFAR training steps w 6 GPUS - run: time HSA=1 DEFAULT_FLOAT=HALF STEPS=350 BS=1536 GPUS=6 TARGET_EVAL_ACC_PCT=93.3 python3 examples/hlb_cifar10.py | tee train_cifar_six_gpu.txt + run: time AMD=1 DEFAULT_FLOAT=HALF STEPS=350 BS=1536 GPUS=6 TARGET_EVAL_ACC_PCT=93.3 python3 examples/hlb_cifar10.py | tee train_cifar_six_gpu.txt - name: Run MLPerf resnet eval - run: time HSA=1 MODEL=resnet python3 examples/mlperf/model_eval.py + run: time AMD=1 MODEL=resnet python3 examples/mlperf/model_eval.py - name: Run 10 MLPerf ResNet50 training steps (1 gpu) - run: HSA=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=256 GPUS=1 MODEL=resnet python3 examples/mlperf/model_train.py | tee train_resnet_one_gpu.txt + run: AMD=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=256 GPUS=1 MODEL=resnet python3 examples/mlperf/model_train.py | tee train_resnet_one_gpu.txt - name: Run 10 MLPerf ResNet50 training steps (6 gpu) - run: HSA=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=1536 GPUS=6 MODEL=resnet python3 examples/mlperf/model_train.py | tee train_resnet.txt + run: AMD=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=1536 GPUS=6 MODEL=resnet python3 examples/mlperf/model_train.py | tee train_resnet.txt - uses: actions/upload-artifact@v4 with: name: Speed (AMD Training)