[pytorch] String optimizations related to serialization. #28230
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
This change improves the pickling small data benchmark by roughly 20%. One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. Differential Revision: [D17982181](https://our.internmc.facebook.com/intern/diff/D17982181/) [ghstack-poisoned]
facebook-github-bot
added
the
oncall: jit
This change improves the pickling small data benchmark by roughly 20%. One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. Differential Revision: [D17982181](https://our.internmc.facebook.com/intern/diff/D17982181/) [ghstack-poisoned]
This change improves the pickling small data benchmark by roughly 30% (25.8usec -> 18.05usec on my machine). One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. Differential Revision: [D17982181](https://our.internmc.facebook.com/intern/diff/D17982181/) [ghstack-poisoned]
jjlilley
pushed a commit
that referenced
this pull request
Oct 17, 2019
Pull Request resolved: #28230 This change improves the pickling small data benchmark by roughly 30%. (25.8usec -> 18.05usec). One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. ghstack-source-id: 92114972 Differential Revision: [D17982181](https://our.internmc.facebook.com/intern/diff/D17982181/)
Author
|
(fwiw, pprof trace is attached to D17982181; this diff should be ready for a look, and gets the % of time spent in just std::stringstream-related constructors to under 10%, from 25%+. I don't think we need wholesale std::stringstream changes, but tweaking the hotter paths seems sensible). |
This change improves the pickling small data benchmark by roughly 30% (25.8usec -> 18.05usec on my machine, with pending zip changes also applied). One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. Differential Revision: [D17982181](https://our.internmc.facebook.com/intern/diff/D17982181/) [ghstack-poisoned]
jjlilley
pushed a commit
that referenced
this pull request
Oct 17, 2019
Pull Request resolved: #28230 This change improves the pickling small data benchmark by roughly 30%. (25.8usec -> 18.05usec). One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. ghstack-source-id: 92153103 Differential Revision: [D17982181](https://our.internmc.facebook.com/intern/diff/D17982181/)
driazati
approved these changes
Oct 18, 2019
| data_type << toString(tensor.scalar_type()) << "Storage"; | ||
| pushGlobal("torch", data_type.str()); | ||
| std::string data_type = | ||
| std::string(toString(tensor.scalar_type())).append("Storage"); |
Contributor
There was a problem hiding this comment.
Is there any difference between append and +?
Author
There was a problem hiding this comment.
append() is going to be slightly more efficient because it assumes we can clobber the lhs string, unlike operator+() which maintains the input lhs and rhs args.
(operator+() unconditionally copies both lhs and rhs into a newly-created string. append() just needs to ensure the buffer is big enough and then copies in the rhs. And for small strings like these, if the string size is within the default in-place buffer, there's a chance no resizing is needed)
Contributor
|
This pull request has been merged in 3d74550. |
thiagocrepaldi
pushed a commit
to thiagocrepaldi/pytorch
that referenced
this pull request
Feb 4, 2020
Summary: Pull Request resolved: pytorch#28230 This change improves the pickling small data benchmark by roughly 30%. (25.8usec -> 18.05usec). One of the main issues was that we were spending 25%+ of the cpu profile time in std::[o]stringstream constructors alone. Two main parts - Change some std::stringstream to std::ostringstream, when they showed up on hot-ish paths, and it was trivial to convert them. Roughly 27% of the std::stringstream constructor time is spent building the constituent std::basic_istream. If the istream isn't needed, don't construct it. - For a couple of very hot paths (e.g. Pickler::pushGlobal), just convert to traditional string::append(). std::ostringstream is convenient, but not particularly efficient. ghstack-source-id: 92153103 Test Plan: Benchmarking: buck build mode/opt experimental/jeremyl/c2:SerializationBench Correctness: buck test mode/dev-nosan caffe2/test/... Differential Revision: D17982181 fbshipit-source-id: 7fd4d267293231244c10c1e5b8f4951a7a3d852f
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:
This change improves the pickling small data benchmark by roughly 30%
(25.8usec -> 18.05usec on my machine, with pending zip changes also applied).
One of the main issues was that we were spending 25%+ of the cpu profile
time in std::[o]stringstream constructors alone.
Two main parts
Change some std::stringstream to std::ostringstream, when they
showed up on hot-ish paths, and it was trivial to convert them.
Roughly 27% of the std::stringstream constructor time is spent
building the constituent std::basic_istream. If the istream isn't
needed, don't construct it.
For a couple of very hot paths (e.g. Pickler::pushGlobal), just
convert to traditional string::append(). std::ostringstream is
convenient, but not particularly efficient.
Differential Revision: D17982181