Bug 177038 – Add an option to run-jsc-stress-tests to limit tests variations to a basic set

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Michael Saboff]

run-jsc-stress-tests has a --quick option, but that runs a minimal number of tests.  It would be good to have something between the minimal --quick set and the full set of  test variation.

[Michael Saboff]

Created attachment 321012 [details]
Patch

[JF Bastien]

Comment on attachment 321012 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=321012&action=review

r=me

> JSTests/ChangeLog:8
> +        * stress/unshiftCountSlowCase-correct-postCapacity.js: Disabled this test on ARM64 iOS devices

Do you know how much memory is needed? Would be good to know when to enable again.

[Michael Saboff]

(In reply to JF Bastien from comment #2)
> Comment on attachment 321012 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=321012&action=review
> 
> r=me
> 
> > JSTests/ChangeLog:8
> > +        * stress/unshiftCountSlowCase-correct-postCapacity.js: Disabled this test on ARM64 iOS devices
> 
> Do you know how much memory is needed? Would be good to know when to enable
> again.

It ran out of memory after using 800-850MB.  On a MacBook Pro I ran the the same test and saw a high usage of 5.4GB as reported by vmmap --dirty.

[WebKit Commit Bot]

Comment on attachment 321012 [details]
Patch

Clearing flags on attachment: 321012

Committed r222135: <http://trac.webkit.org/changeset/222135>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Saam Barati]

Comment on attachment 321012 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=321012&action=review

> Tools/Scripts/run-jsc-stress-tests:767
> +        return if !$isFTLPlatform

Why this change?

[Michael Saboff]

Comment on attachment 321012 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=321012&action=review

>> Tools/Scripts/run-jsc-stress-tests:767
>> +        return if !$isFTLPlatform
> 
> Why this change?

This is a bug fix I added.  The statement was doing the opposite of what was intended as the following tests are FTL tests.

[Radar WebKit Bug Importer]

<rdar://problem/34693287>