Faster IndexOf for arm64 #67811
Merged
Faster IndexOf for arm64 #67811
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
|
Tagging subscribers to this area: @dotnet/area-system-memory Issue DetailsUse a slightly faster "no matches" check in Benchmark: public class Benchmarks
{
public IEnumerable<object[]> TestData()
{
// Small inputs which are handled via SIMD
yield return new object[] { "12345678", '1' };
yield return new object[] { "123456789", '9' };
yield return new object[] { "1234567812345678", '1' };
yield return new object[] { "1234567812345678", '0' };
yield return new object[] { "Console WriteLine Hello World", 'o' };
yield return new object[] { "Console WriteLine Hello World", 'd' };
// Large inputs
yield return new object[] { new string('x', 64), 'y' };
yield return new object[] { new string('x', 200), 'y' };
yield return new object[] { new string('x', 1000), 'y' };
yield return new object[] { new string('x', 1000000), 'y' };
}
[Benchmark]
[ArgumentsSource(nameof(TestData))]
public int IndexOf_byte(string str, char c)
{
return MemoryMarshal.Cast<char,byte>(str.AsSpan()).IndexOf((byte)c);
}
[Benchmark]
[ArgumentsSource(nameof(TestData))]
public int IndexOf_char(string str, char c)
{
return str.AsSpan().IndexOf(c);
}
}IndexOf_byteIndexOf_charup +40% faster for large inputs can 0.5-1ns regress in the worst case - if input is found in the very first vector.
|
EgorBo
force-pushed
the
arm-spanhelpers-fast
branch
from
9e59caa to
8ab651d
Compare
Contributor
|
Performance improvement on Windows arm64: dotnet/perf-autofiling-issues#4732 |
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Use a slightly faster "no matches" check in
IndexOfandIndexOfAny.Benchmark:
IndexOf_byte
IndexOf_char
up to +40% faster for large inputs can 0.5-1ns regress in the worst case - if input is found in the very first vector.
There are weird regressions
but it goes away in FullPGO mode (and the PR becomes faster than the base as expected in this case) - it means the loop is not properly aligned by default. So the regression will be fixed once we update *.mibc for corelib (IndexOf definitely participates in the trainings)