bench: benchmark several common WASI scenarios #5274
Merged
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
In order to properly understand the impact of providing thread-safe implmentations of WASI contexts (bytecodealliance#5235), we need benchmarks that measure the current performance of WASI calls using Wiggle. This change adds several common WASI scenarios as WAT files (see `benches/wasi/*.wat`) and benchmarks them with `criterion`. Using `criterion`'s `iter_custom`, the WAT file runs the desired number of benchmark iterations internally and the total duration of the runs is divided to get the average time for each loop iteration. Why WAT? When compiling these benchmarks from Rust to `wasm32-wasi`, the output files are large, contain other WASI imports than the desired ones, and overall it is difficult to tell if we are measuring what we expect. By hand-writing the WAT, it is (slightly) more clear what each benchmark is doing.
|
cc: @alexcrichton, @pchickey: I would like to add benchmarks for "scan a directory" and "read a file" but I don't know how to do that yet... any pointers to how to use |
jameysharp
approved these changes
Nov 15, 2022
15 tasks
|
Not much I can say about using fd_read and path_open from using |
|
I'm going to merge this as-is and figure out the next scenarios in a future PR. |
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Nov 21, 2022
This follows up on bytecodealliance#5274 to add several more scenarios with which to benchmark WASI performance: - `open-file.wat`: opens and closes a file - `read-file.wat`: opens a file, reads 4K bytes from it, then closes it - `read-dir.wat`: reads a directory's entries Each benchmark is hand-crafted WAT to more clearly control what WASI calls are made. As with bytecodealliance#5274, these modules' sole entry point takes a parameter indicating the number of iterations to run in order to use `criterion`'s `iter_custom` feature.
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Nov 21, 2022
This follows up on bytecodealliance#5274 to add several more scenarios with which to benchmark WASI performance: - `open-file.wat`: opens and closes a file - `read-file.wat`: opens a file, reads 4K bytes from it, then closes it - `read-dir.wat`: reads a directory's entries Each benchmark is hand-crafted WAT to more clearly control what WASI calls are made. As with bytecodealliance#5274, these modules' sole entry point takes a parameter indicating the number of iterations to run in order to use `criterion`'s `iter_custom` feature.
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Nov 21, 2022
This follows up on bytecodealliance#5274 to add several more scenarios with which to benchmark WASI performance: - `open-file.wat`: opens and closes a file - `read-file.wat`: opens a file, reads 4K bytes from it, then closes it - `read-dir.wat`: reads a directory's entries Each benchmark is hand-crafted WAT to more clearly control what WASI calls are made. As with bytecodealliance#5274, these modules' sole entry point takes a parameter indicating the number of iterations to run in order to use `criterion`'s `iter_custom` feature.
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Nov 21, 2022
This follows up on bytecodealliance#5274 to add several more scenarios with which to benchmark WASI performance: - `open-file.wat`: opens and closes a file - `read-file.wat`: opens a file, reads 4K bytes from it, then closes it - `read-dir.wat`: reads a directory's entries Each benchmark is hand-crafted WAT to more clearly control what WASI calls are made. As with bytecodealliance#5274, these modules' sole entry point takes a parameter indicating the number of iterations to run in order to use `criterion`'s `iter_custom` feature.
abrown
added a commit
that referenced
this pull request
Nov 22, 2022
* bench: add more WASI benchmarks This follows up on #5274 to add several more scenarios with which to benchmark WASI performance: - `open-file.wat`: opens and closes a file - `read-file.wat`: opens a file, reads 4K bytes from it, then closes it - `read-dir.wat`: reads a directory's entries Each benchmark is hand-crafted WAT to more clearly control what WASI calls are made. As with #5274, these modules' sole entry point takes a parameter indicating the number of iterations to run in order to use `criterion`'s `iter_custom` feature. * fix: reduce expected size of directory entries
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.
In order to properly understand the impact of providing thread-safe
implmentations of WASI contexts (#5235), we need benchmarks that measure
the current performance of WASI calls using Wiggle. This change adds
several common WASI scenarios as WAT files (see
benches/wasi/*.wat)and benchmarks them with
criterion. Usingcriterion'siter_custom,the WAT file runs the desired number of benchmark iterations internally
and the total duration of the runs is divided to get the average time
for each loop iteration.
Why WAT? When compiling these benchmarks from Rust to
wasm32-wasi, theoutput files are large, contain other WASI imports than the desired
ones, and overall it is difficult to tell if we are measuring what we
expect. By hand-writing the WAT, it is (slightly) more clear what each
benchmark is doing.