[inductor] let codegen not rely on node order #107320
Closed
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
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/107320
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 067f47c with merge base 138e289 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
See a lot of CI failures. Convert to draft for now. Will change it back once I fix them |
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
|
Test failures fixed. Ready for review now. |
peterbell10
reviewed
Aug 18, 2023
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
jansel
reviewed
Aug 23, 2023
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
shunting314
commented
Aug 24, 2023
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
peterbell10
requested changes
Aug 25, 2023
shunting314
commented
Aug 25, 2023
|
Feel free to re-request review once peters comments are addressed |
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (`fuse_nodes_once`) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done). Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a): ``` benchmark one fusion decision -> codegen -> xSchedulerNode.last_usage -> node order -> all fusions have completed ``` Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx peterbell10 ipiszy ngimel yf225 chenyang78 kadeng muchulee8 aakhundov [ghstack-poisoned]
jansel
approved these changes
Aug 29, 2023
peterbell10
approved these changes
Aug 29, 2023
|
@pytorchbot merge |
Merge failedReason: This PR needs a If not, please add the To add a label, you can comment to pytorchbot, for example For more information, see Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot label "topic: not user facing" |
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
ezyang
added a commit
that referenced
this pull request
Sep 17, 2023
This reverts commit 556bfe7. [ghstack-poisoned]
ezyang
added a commit
that referenced
this pull request
Sep 17, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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):
We'd like to benchmark fusion (either for autotuning or for gathering data to find some patterns that can guide optimizations). There is a deadlock here that prevents us from doing this: to benchmark fusion, we need do codegen before all the fusions are done. However currently codegen rely on xSchedulerNode.last_usage information to decide which buffers are not needed at all and thus don't even need to be allocated/written (Scheduler.removed_buffers tracks this). xSchedulerNode.last_usage information can only be computed once the order of all the nodes have been decided. But each fusion pass (
fuse_nodes_once) can also change node orders. So we know the final node orders only after all the fusions have completed. That blocks us from doing codegen during fusion (before all fusion are done).Here I just show the above with a chain of dependencies to make it easier to understand (a -> b means a depends on b, or b has to happen before a):
Actually we only need to decide if a buffer has only local usages (if yes, it's a candidate for removing). This can be decided if we know what are all the users for each buffer. We can avoid using xSchedulerNode.last_usage in this case.
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @Xia-Weiwen @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @ngimel @yf225 @chenyang78 @kadeng @muchulee8 @aakhundov