Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

WIP: Re-add "-fno-rounding-math" and "-fno-signaling-nans" options in JIT #6755

Closed

Conversation

babsingh
Copy link
Contributor

@babsingh babsingh commented Aug 16, 2019

The above compiler options were removed in #4156.

Removing the above compiler options causes a performance regression.
Symptoms of the performance regression: bursts of high throughput mixed
with periods of lower throughput. These symptoms cause a 5-7% throughput
regression.

Reintroducing the above compiler options fixes the throughput
regression.

Signed-off-by: Babneet Singh sbabneet@ca.ibm.com

The above compiler options were removed in
eclipse-openj9#4156.

Removing the above compiler options causes a performance regression.
Symptoms of the performance regression: bursts of high throughput mixed
with periods of lower throughput. These symptoms cause a 5-7% throughput
regression.

Reintroducing the above compiler options fixes the throughput
regression.

Signed-off-by: Babneet Singh <sbabneet@ca.ibm.com>
@babsingh babsingh changed the title Re-add "-fno-rounding-math" and "-fno-signaling-nans" options in JIT WIP: Re-add "-fno-rounding-math" and "-fno-signaling-nans" options in JIT Aug 16, 2019
@AdamBrousseau
Copy link
Contributor

Jenkins copyright check

@andrewcraik
Copy link
Contributor

There is a lot of discussion in the PR that removed the flags #4156 which says that the -fno-rounding-math and -fno-signaling-nans are default on gcc 4.7 and up - what gcc is causing the problem? Our default GCC should be above where they became default?

@babsingh
Copy link
Contributor Author

babsingh commented Aug 26, 2019

@andrewcraik The performance issue is still under investigation (WIP). This change caught my eye in the git diff between the last good and first bad SDKs. I haven't received conclusive data that this change may resolve the performance issue. So, I am closing this pull request.

@babsingh babsingh closed this Aug 26, 2019
@babsingh babsingh deleted the enable_jit_gcc_optimizations branch August 17, 2020 02:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants