Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[show-gpus] Optimize show-gpus for GCP and reorganize TPU #3113

Merged
merged 14 commits into from
Feb 12, 2024

Conversation

Michaelvll
Copy link
Collaborator

@Michaelvll Michaelvll commented Feb 7, 2024

Fixes #2953.

This is to reimplement the fix in #3083 reverted in #3110. This PR makes the implementation have similar speed as before with correct results.

This PR:

time sky show-gpus
1.470s

time sky show-gpus --cloud gcp
0.991s

time sky show-gpus v100
1.431s

time sky show-gpus v100 --cloud gcp
1.041s

time sky show-gpus a100
1.629s
sky show-gpus tpu
GPU         QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v2-8    1    GCP    (attachable)   -           -      -         $ 4.500       $ 1.350            us-central1   
tpu-v2-32   1    GCP    (attachable)   -           -      -         $ 24.000      $ 7.200            europe-west4  
tpu-v2-128  1    GCP    (attachable)   -           -      -         $ 96.000      $ 28.800           europe-west4  
tpu-v2-256  1    GCP    (attachable)   -           -      -         $ 192.000     $ 57.600           europe-west4  
tpu-v2-512  1    GCP    (attachable)   -           -      -         $ 384.000     $ 115.200          europe-west4  
...

Master 905dd9c

time sky show-gpus
1.833s

time sky show-gpus --cloud gcp
1.332s

time sky show-gpus v100
1.368s

time sky show-gpus v100 --cloud gcp
0.975s

time sky show-gpus a100
1.430s
 time sky show-gpus tpu
GPU         QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v2-128  1    GCP    TPU-VM         -           -      -         $ 96.000      $ 28.800           europe-west4  

GPU         QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v2-256  1    GCP    TPU-VM         -           -      -         $ 192.000     $ 57.600           europe-west4  

GPU        QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v2-32  1    GCP    TPU-VM         -           -      -         $ 24.000      $ 7.200            europe-west4  

GPU         QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v2-512  1    GCP    TPU-VM         -           -      -         $ 384.000     $ 115.200          europe-west4  

GPU       QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION       
tpu-v2-8  1    GCP    TPU-VM         -           -      -         $ 4.500       $ 1.350            us-central1  

GPU          QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v3-1024  1    GCP    TPU-VM         -           -      -         $ 1024.000    $ 307.200          europe-west4  

GPU         QTY  CLOUD  INSTANCE_TYPE  DEVICE_MEM  vCPUs  HOST_MEM  HOURLY_PRICE  HOURLY_SPOT_PRICE  REGION        
tpu-v3-128  1    GCP    TPU-VM         -           -      -         $ 128.000     $ 38.400           europe-west4 
...

Tested (run the relevant ones):

  • Code formatting: bash format.sh
  • Any manual or new tests for this PR (please specify below)
  • All smoke tests: pytest tests/test_smoke.py
  • Relevant individual smoke tests: pytest tests/test_smoke.py::test_fill_in_the_name
  • Backward compatibility tests: bash tests/backward_comaptibility_tests.sh

@Michaelvll Michaelvll changed the title [show-gpus] Optimize show gpus gcp [show-gpus] Optimize show-gpus for gcp Feb 7, 2024
@Michaelvll Michaelvll changed the title [show-gpus] Optimize show-gpus for gcp [show-gpus] Optimize show-gpus for GCP and reorganize TPU Feb 7, 2024
Copy link
Collaborator

@cblmemo cblmemo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks! It looks great to me. Left several nits on readability & minor issues

sky/clouds/service_catalog/gcp_catalog.py Outdated Show resolved Hide resolved
sky/clouds/service_catalog/gcp_catalog.py Outdated Show resolved Hide resolved
sky/clouds/service_catalog/gcp_catalog.py Show resolved Hide resolved
sky/clouds/service_catalog/gcp_catalog.py Outdated Show resolved Hide resolved
sky/clouds/service_catalog/gcp_catalog.py Outdated Show resolved Hide resolved
sky/clouds/service_catalog/gcp_catalog.py Show resolved Hide resolved
@Michaelvll Michaelvll merged commit 3f0ad21 into master Feb 12, 2024
19 checks passed
@Michaelvll Michaelvll deleted the optimize-show-gpus-gcp branch February 12, 2024 17:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[Catalog] Inconsistency for GCP, A100:16 instance
2 participants