Recent drop in JetStream 2 Score

[Suraj Sharma]

Hi,

There is a recent drop in overall score on JetStream 2  (~4.5%) from Chrome v79 to v80. 

There are 3 tests that are significantly hurting our score(Scores from my local machine):

Tests	Chrome 79 Score	Chrome 80 Score
Async-fs	244	72
crypto-sha1-SP	356	141
crypto-md5-SP	430	179

Note: Higher Score indicate better performance

On further investigation, it seems turning on v8_enable_pointer_compression OR v8_enable_31bit_smis_on_64bit_arch significantly hurt our performance in these cases. 

Here are the results from my local machine before and after turning off both flags on the x64 release build of d8.

Tests	Current Execution time (ms)	Execution time without the flags (ms)	Improvement
Async-fs	120290.64	36429.46	230%
crypto-sha1-SP	2869.39	1041.91	175%
crypto-md5-SP	2690.55	1157	133%

Have we seen this before? Does anyone know why we are seeing this?

Thanks

Suraj.

[Santiago Aboy Solanes]

In Chrome 80 we enabled pointer compression by default, which enables 31 bit smi by default.

Have you tested both in Chrome 79 and 80 with v8_enable_31bit_smis_on_64bit_arch = true? (please use v8_enable_pointer_compression = false for 80)

--
--
v8-dev mailing list
v8-...@googlegroups.com
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to v8-dev+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/741085bd-0131-42e6-bc53-30522c4f27dd%40googlegroups.com.

[Igor Sheludko]

Hi! 

Thank you for the report!

31-bit smis are known to slow down some Smi-size sensitive benchmarks and we already addressed some of the regressions here https://bugs.chromium.org/p/v8/issues/detail?id=8344.

Certainly we are planning to take another look over all regressions but I'm not sure that all 31-bit Smi related ones are fixable.

To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/CAFKEYJjVVs5ss%2B%3DsQc%3DUQBxY4gJvLu8u4p3S2EBbQWGggyisDw%40mail.gmail.com.

--

Igor Sheludko

Software Engineer

ish...@google.com

Google Germany GmbH

Erika-Mann-Str. 33

80636 München

Geschäftsführer: Paul Manicle, Halimah DeLaine Prado

Registergericht und -nummer: Hamburg, HRB 86891

Sitz der Gesellschaft: Hamburg

Diese E-Mail ist vertraulich. Falls sie diese fälschlicherweise erhalten haben sollten, leiten Sie diese bitte nicht an jemand anderes weiter, löschen Sie alle Kopien und Anhänge davon und lassen Sie mich bitte wissen, dass die E-Mail an die falsche Person gesendet wurde.

    

This e-mail is confidential. If you received this communication by mistake, please don't forward it to anyone else, please erase all copies and attachments, and please let me know that it has gone to the wrong person.

[Suraj Sharma]

Thanks for the response.

Yes, I had verified earlier that the regression had something to do with pointer compression and 31 bit Smi both.

I re-ran the tests on x64 release build of d8 using the latest v8 master on my local machine

Here are the results:

Tests

Current Execution time (ms)

Only Pointer Compression turned off (ms)	Only 31-bit Smi turned off (ms)	Both flags turned off (ms)
Async-fs	115944.46	140734.31	112070.59	30315.62
crypto-sha1-SP	2371.06	2361.47	2379.39	791.12
crypto-md5-SP	2347.19	2352.17	2294.41	875.07

Note: I ensured that turning on pointer compression doesn't turn on 31 bit Smi by default.

Please let me know if you need any more info.

Thanks

Suraj

To unsubscribe from this group and stop receiving emails from it, send an email to v8-...@googlegroups.com.

[Suraj Sharma]

Hello Igor

Thanks for the update.

To unsubscribe from this group and stop receiving emails from it, send an email to v8-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/741085bd-0131-42e6-bc53-30522c4f27dd%40googlegroups.com.

--
--
v8-dev mailing list
v8-...@googlegroups.com
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.

To unsubscribe from this group and stop receiving emails from it, send an email to v8-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/CAFKEYJjVVs5ss%2B%3DsQc%3DUQBxY4gJvLu8u4p3S2EBbQWGggyisDw%40mail.gmail.com.