[CUDACachingAllocator] Better use of blocks with rounding of allocation sizes #74213
Conversation
This file contains hidden or 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
…ion sizes In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) [ghstack-poisoned]
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit 813ca75 (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
CI Flow Status⚛️ CI FlowRuleset - Version:
|
banitag1
added a commit
that referenced
this pull request
Mar 15, 2022
…ion sizes In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) ghstack-source-id: 151355133 Pull Request resolved: #74213
banitag1
changed the title
Contributor
|
cc @mcarilli |
…of allocation sizes" In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) [ghstack-poisoned]
…of allocation sizes" In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) [ghstack-poisoned]
banitag1
added a commit
that referenced
this pull request
Mar 15, 2022
…ion sizes Pull Request resolved: #74213 In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 ghstack-source-id: 151412183 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34868036/)!
…of allocation sizes" In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) [ghstack-poisoned]
…of allocation sizes" In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) [ghstack-poisoned]
…of allocation sizes" In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) [ghstack-poisoned]
banitag1
added a commit
that referenced
this pull request
Mar 15, 2022
…ion sizes Pull Request resolved: #74213 In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 ghstack-source-id: 151446017 Differential Revision: [D34868036](https://our.internmc.facebook.com/intern/diff/D34868036/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34868036/)!
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 16, 2022
…ion sizes (#74213) Summary: Pull Request resolved: #74213 In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity. This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting. For example, if we need to round-up size of1200 and if number of divisions is 4, the size 1200 lies between 1024 and 2048 and if we do 4 divisions between them, the values are 1024, 1280, 1536, and 1792. So the function will return 1280 as the nearest ceiling of power-2 division. env setting: export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4 ghstack-source-id: 151446017 Reviewed By: ezyang Differential Revision: D34868036 fbshipit-source-id: 494785add16e6b37c920dcb5a2b81d4c637b554a
Contributor
|
Hey @banitag1. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
In the current CUDACachingAllocator, the sizes are rounded up in multiple of blocks size of 512, so this works for smaller sizes. However for large sizes, we can have lots of different size blocks in the larger pool. This is problematic when we have variable batch sizes 1001, 1021, 1023 -> all will go to different block size and will create different size of blocks. This will create lots of unused blocks and will waste GPU memory capacity.
This diff adds a rounding approach to allocation size. It rounds up the size to nearest power-of-2 divisions and the power2-division can be changed with env variable setting.
For example, if we need to round-up size of1200 and if number of divisions is 4,
the size 1200 lies between 1024 and 2048 and if we do 4 divisions between
them, the values are 1024, 1280, 1536, and 1792. So the function will
return 1280 as the nearest ceiling of power-2 division.
env setting:
export PYTORCH_CUDA_ALLOC_CONF=roundup_power2_divisions:4
Differential Revision: D34868036