Blink heap fragmentation observed under certain allocation patterns

6 views
Skip to first unread message

Victor Zhavoronkov

unread,
Jul 17, 2026, 3:24:30 AM (13 days ago) Jul 17
to memory-dev
Hi,

We have observed that certain Blink object allocation patterns lead to a seemingly persistent Blink heap fragmentation (~36 MiB allocated vs. 1.4 GiB total heap size when idle after about 15 minutes running).

Is this a correct channel for reporting this?

If yes, what info would be useful aside from repro HTML and trace with `memory-infra` toggled on?

Thanks,
Victor.

Michael Lippautz

unread,
Jul 17, 2026, 3:32:38 AM (13 days ago) Jul 17
to Victor Zhavoronkov, memory-dev, Dominik Inführ
On Fri, Jul 17, 2026 at 9:24 AM 'Victor Zhavoronkov' via memory-dev <memor...@chromium.org> wrote:
Hi,

We have observed that certain Blink object allocation patterns lead to a seemingly persistent Blink heap fragmentation (~36 MiB allocated vs. 1.4 GiB total heap size when idle after about 15 minutes running).

Is this a correct channel for reporting this?


crbug.com/new with component Blink>GarbageCollection  
 
If yes, what info would be useful aside from repro HTML and trace with `memory-infra` toggled on?


Most important are
- repro 
- memory-infra trace
- DevTools heap snapshot on a recent Canary

-Michael
 
Thanks,
Victor.

This e-mail transmission may contain legally privileged and/or CONFIDENTIAL information. Unauthorized use, disclosure, distribution and/or copying of this e-mail or any information it contains is prohibited. If you have received this email in error or are not an intended recipient, please inform Exactpro Systems Limited (“Exactpro”) immediately by return e-mail. Exactpro may collect, process and retain your personal information for its business purposes as per its  Privacy Policy. E-mail transmissions cannot be guaranteed to be secure or error-free as information could be intercepted, corrupted or incomplete, or contain viruses. The sender, therefore, does not accept liability for any viruses, errors or omissions in the contents of this message which arise as a result of e-mail transmission. The views and opinions expressed in this message are those of the individual sender of this message and do not necessarily represent the views and opinions of Exactpro.

Exactpro Systems Limited, Suite 1.10 6-9 Snow Hill, London, England, EC1A 2AY. Registered in England and Wales No 09485548

--
You received this message because you are subscribed to the Google Groups "memory-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to memory-dev+...@chromium.org.
To view this discussion visit https://groups.google.com/a/chromium.org/d/msgid/memory-dev/3eac779c-0557-44d5-970e-64e68d89bdf8n%40chromium.org.

Victor Zhavoronkov

unread,
Jul 17, 2026, 5:56:51 AM (13 days ago) Jul 17
to memory-dev, Michael Lippautz, memory-dev, Dominik Inführ, Victor Zhavoronkov
Reply all
Reply to author
Forward
0 new messages