forked from mlcommons/algorithmic-efficiency
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
add some flags to sub runner, control memory, max pct steps + scripts
- Loading branch information
1 parent
adc5ea9
commit 68735ea
Showing
8 changed files
with
334 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,42 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/jax/a100/auto_run_array.sh | ||
|
|
||
| num_jobs=8 | ||
|
|
||
| # Job specific vars | ||
| workload_or_id=$(Process) | ||
| # workload_or_id=imagenet_vit | ||
| framework=jax | ||
| submission=prize_qualification_baselines/external_tuning/jax_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=a100_yesTF32_04 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 1000000 | ||
| request_cpus = 24 | ||
| request_gpus = 8 | ||
| requirements = (TARGET.CUDADeviceName == "NVIDIA A100-SXM4-80GB") | ||
| # requirements = (TARGET.CUDADeviceName == "NVIDIA A100-SXM4-40GB") | ||
|
|
||
| queue $(num_jobs) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,41 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/jax/v100/auto_run_array.sh | ||
|
|
||
| num_jobs=8 | ||
|
|
||
| # Job specific vars | ||
| workload_or_id=$(Process) | ||
| # workload_or_id=imagenet_vit | ||
| framework=jax | ||
| submission=prize_qualification_baselines/external_tuning/jax_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=v100_04 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 700000 | ||
| request_cpus = 24 | ||
| request_gpus = 8 | ||
| requirements = (TARGET.CUDADeviceName == "Tesla V100-SXM2-32GB") | ||
|
|
||
| queue $(num_jobs) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,45 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/pytorch/2xa100/auto_run_array.sh | ||
|
|
||
| num_jobs=8 | ||
|
|
||
| # Job specific vars | ||
| workload_or_id=$(Process) | ||
| # workload_or_id=criteo1tb | ||
| framework=pytorch | ||
| submission=prize_qualification_baselines/external_tuning/pytorch_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=a10040GB_x2_noTF32_01 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
| allow_tf_32=0 | ||
| halve_cuda_mem=0 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) \ | ||
| $(allow_tf_32) \ | ||
| $(halve_cuda_mem) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 250000 | ||
| request_cpus = 12 | ||
| request_gpus = 2 | ||
| requirements = (TARGET.CUDADeviceName == "NVIDIA A100-SXM4-40GB") | ||
|
|
||
| queue $(num_jobs) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,45 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/pytorch/4xa100/auto_run_array.sh | ||
|
|
||
| num_jobs=8 | ||
|
|
||
| # Job specific vars | ||
| workload_or_id=$(Process) | ||
| # workload_or_id=imagenet_resnet | ||
| framework=pytorch | ||
| submission=prize_qualification_baselines/external_tuning/pytorch_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=a100_x4_yesTF32_10 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
| allow_tf_32=1 | ||
| eval_num_workers=4 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) \ | ||
| $(allow_tf_32) \ | ||
| $(eval_num_workers) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 700000 | ||
| request_cpus = 36 | ||
| request_gpus = 4 | ||
| requirements = (TARGET.CUDADeviceName == "NVIDIA A100-SXM4-40GB") | ||
|
|
||
| queue $(num_jobs) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,43 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/pytorch/8xa100/auto_run_array.sh | ||
|
|
||
| num_jobs=8 | ||
|
|
||
| # Job specific vars | ||
| workload_or_id=$(Process) | ||
| # workload_or_id=imagenet_vit | ||
| framework=pytorch | ||
| submission=prize_qualification_baselines/external_tuning/pytorch_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=a100_x8_noTF32_10 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
| allow_tf_32=0 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) \ | ||
| $(allow_tf_32) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 700000 | ||
| request_cpus = 36 | ||
| request_gpus = 8 | ||
| requirements = (TARGET.CUDADeviceName == "NVIDIA A100-SXM4-40GB") | ||
|
|
||
| queue $(num_jobs) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,41 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/pytorch/v100/auto_run_array.sh | ||
|
|
||
| num_jobs=8 | ||
|
|
||
| # Job specific vars | ||
| workload_or_id=$(Process) | ||
| # workload_or_id=imagenet_resnet | ||
| framework=pytorch | ||
| submission=prize_qualification_baselines/external_tuning/pytorch_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=v100_10 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 700000 | ||
| request_cpus = 36 | ||
| request_gpus = 8 | ||
| requirements = (TARGET.CUDADeviceName == "Tesla V100-SXM2-32GB") | ||
|
|
||
| queue $(num_jobs) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,45 @@ | ||
| # Define variables | ||
| LOGS_DIR=/fast/najroldi/logs/algoperf | ||
| EXE=/home/najroldi/algorithmic-efficiency/script/pytorch/v100_test/auto_run_array.sh | ||
|
|
||
| num_jobs=1 | ||
|
|
||
| # Job specific vars | ||
| # workload_or_id=$(Process) | ||
| workload_or_id=imagenet_resnet | ||
| framework=pytorch | ||
| submission=prize_qualification_baselines/external_tuning/pytorch_nadamw_full_budget.py | ||
| search_space=prize_qualification_baselines/external_tuning/tuning_search_space.json | ||
| name=v100_test_05 | ||
| study=1 | ||
| num_tuning_trials=1 | ||
| rng_seed=96 | ||
| eval_workers=2 | ||
| omp_threads=2 | ||
|
|
||
| # Args | ||
| executable = $(EXE) | ||
| arguments = \ | ||
| $(workload_or_id) \ | ||
| $(framework) \ | ||
| $(submission) \ | ||
| $(search_space) \ | ||
| $(name) \ | ||
| $(study) \ | ||
| $(num_tuning_trials) \ | ||
| $(rng_seed) \ | ||
| $(eval_workers) \ | ||
| $(omp_threads) | ||
|
|
||
| # Logs | ||
| error = $(LOGS_DIR)/err/job.$(Cluster).$(Process).err | ||
| output = $(LOGS_DIR)/out/job.$(Cluster).$(Process).out | ||
| log = $(LOGS_DIR)/log/job.$(Cluster).$(Process).log | ||
|
|
||
| # Specs | ||
| request_memory = 700000 | ||
| request_cpus = 36 | ||
| request_gpus = 8 | ||
| requirements = (TARGET.CUDADeviceName == "Tesla V100-SXM2-32GB") | ||
|
|
||
| queue $(num_jobs) |
Oops, something went wrong.