search

steleman / address-sanitizer

Automatically exported from code.google.com/p/address-sanitizer
0 stars 0 forks source link

LLVM sanitizer-bootstrap bot failing #350

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
I see multiple ld processes, each takes from 1 to 4 Gb RAM (RSS). Heavy 
swapping (the machine has 24Gb RAM).

No idea where this comes from. We don't instrument ld, and on my machine 
(non-bootstrap) ld takes ~300Mb for the same targets.

One of the targets is clang-format.

Any ideas?

Original issue reported on code.google.com by euge...@google.com on 2 Oct 2014 at 7:37

GoogleCodeExporter commented 9 years ago 
Don't know. Maybe, instrumented binaries with debug info are just much larger, 
so linking them requires more memory.

Original comment by samso...@google.com on 2 Oct 2014 at 7:56

GoogleCodeExporter commented 9 years ago 
We could use this "job pool" feature of cmake+ninja:
http://www.cmake.org/cmake/help/v3.0/prop_gbl/JOB_POOLS.html

Will probably need to upgrade both cmake and ninja.

Original comment by euge...@google.com on 3 Oct 2014 at 8:58

GoogleCodeExporter commented 9 years ago 
It's become a lot worse since UBSan was enabled on the bot.
It used to fail once a day or two, now it's constantly red.

Original comment by euge...@google.com on 3 Oct 2014 at 2:35