This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Remove atomic operations from malloc.c
- From: Will Newton <will dot newton at linaro dot org>
- To: Leonhard Holz <leonhard dot holz at web dot de>
- Cc: libc-alpha <libc-alpha at sourceware dot org>
- Date: Wed, 11 Feb 2015 09:25:17 +0000
- Subject: Re: [PATCH] Remove atomic operations from malloc.c
- Authentication-results: sourceware.org; auth=none
- References: <54DB130F dot 9070300 at web dot de> <CANu=DmjP0kQV2ivYduq_B3EYPU-FCjWah0+6jyYC2_CxbURmww at mail dot gmail dot com> <54DB1ACB dot 2050307 at web dot de>
On 11 February 2015 at 09:03, Leonhard Holz <leonhard.holz@web.de> wrote:
>>> Current implementation
>>> threads time per iteration
>>> 1 116.709
>>> 8 705.080
>>> 16 1394.74
>>> 32 2923.03
>>>
>>> Without atomics
>>> threads time per iteration
>>> 1 112.541
>>> 8 715.897
>>> 16 1403.67
>>> 32 2881.30
>>
>>
>> Did you try running the malloc benchtest?
>>
>
> Yes, these are the numbers from bench-malloc-thread.
Good to know it has been useful to someone. ;-)
Benchmarking malloc is notoriously hard due to the wide variety of
workloads so it would be good to also run some other more system type
benchmark e.g. sysbench in order to convince people that their real
world workloads won't regress.
--
Will Newton
Toolchain Working Group, Linaro