[SharedCache] Optimize parsing and applying of slide info #6321
Conversation
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
Slide info is parsed and applied on the main thread, below the `SharedCache` constructor, when a shared cache is opened. Time spent applying the slide is time that the main thread is blocked. These changes eliminate some unnecessary overhead so what work remains is dominated by kernel work (paging in data, copying pages when the first modification is made). The changes are: 1. Read slide pointers via `ReadULong` rather than the variable-length `Read` method. The compiler isn't able to eliminate the call to `memcpy` in the variable-length `Read` function, and the function call overhead is noticeable given the small size of the read and number of times it is called. 2. Remove the file member from `MappingInfo`. Slide info is applied for a single file at a time so there's no reason to track the file it belongs to. This removes unnecesssary reference counting on the `std::shared_ptr` that holds the `MMappedFileAccessor`.
bdash
force-pushed
the
dsc-optimize-slide-info-parsing
branch
from
a554405 to
aa22af9
Compare
|
Updated after 27eb30d removed use of the |
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Outdated suggestions cannot be applied.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.
Slide info is parsed and applied on the main thread, below the
SharedCacheconstructor, when a shared cache is opened. Time spent applying the slide is time that the main thread is blocked.These changes eliminate some unnecessary overhead so what work remains is dominated by kernel work (paging in data, copying pages when the first modification is made).
The changes are:
ReadULongrather than the variable-lengthReadmethod. The compiler isn't able to eliminate the call tomemcpyin the variable-lengthReadfunction, and the function call overhead is noticeable given the small size of the read and number of times it is called.MappingInfo. Slide info is applied for a single file at a time so there's no reason to track the file it belongs to. This removes unnecesssary reference counting on thestd::shared_ptrthat holds theMMappedFileAccessor.