Skip to content
This repository was archived by the owner on Jan 23, 2023. It is now read-only.

Reduce SeekUnroll iter count #8861

Merged
merged 1 commit into from
Jan 10, 2017

Conversation

michellemcdaniel
Copy link

On non-hardware accelerated platforms, reduce the inner iteration count in
SeekUnroll. This test was timing out in x86 release builds because the
iteration count was too high.

@michellemcdaniel
Copy link
Author

@JosephTremoulet PTAL
/cc: @dotnet/jit-contrib

@michellemcdaniel
Copy link
Author

Fixes #8838

Copy link

@JosephTremoulet JosephTremoulet left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM modulo a couple nits/questions, thanks.

@@ -14,7 +14,7 @@
[assembly: OptimizeForBenchmarks]
[assembly: MeasureInstructionsRetired]

public static class SeekUnroll
public class SeekUnroll

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nit: This bit seems unnecessary since you've made the field static

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done.

@@ -143,6 +143,11 @@ public static bool Test(object boxedIndex)
public static int Main()
{
int failures = 0;
if (!Vector.IsHardwareAccelerated)
{

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. Might want to add a comment explaining this.
  2. Have you verified that this fixes the problem when run through the test harness? It just now occurred to me that maybe Main doesn't execute when xunit-perf runs the test? (also, I'm not sure if our CI is running this through xunit-perf or not... presumably if we do want to track the performance of this test on a non-accelerated platform we'd want some value between 1 and 100000000...)

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added a comment and upped the value to 100000. We can up it again later if we need/want to. I've tested this inside runtest.cmd, and it runs successfully.

On non-hardware accelerated platforms, reduce the inner iteration count in
SeekUnroll. This test was timing out in x86 release builds because the
iteration count was too high.
@michellemcdaniel michellemcdaniel merged commit 7dd3478 into dotnet:master Jan 10, 2017
@michellemcdaniel michellemcdaniel deleted the updateSeekUnroll branch January 10, 2017 17:38
@karelz karelz modified the milestone: 2.0.0 Aug 28, 2017
picenka21 pushed a commit to picenka21/runtime that referenced this pull request Feb 18, 2022
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants