Messages in this thread | | | Date | Tue, 1 Apr 2008 10:43:08 +0200 | From | Ingo Molnar <> | Subject | Re: spinlocks -- why are releases inlined and acquires are not? |
| |
* Jiri Kosina <jkosina@suse.cz> wrote:
> In fact we have received report from one of our users that he is > seeing approximately 15% performance degradation of mmap() when > spinlocks are not inlined. I am going to do some performance > measurements myself shortly, as it seems quite strange, but while at > it, I have noticed the aforementioned asymetry in spinlock.h, so I > just wanted to know if there is any particular reason behind that.
inlining decisions almost never have effects of that order of magnitude - especially on new CPUs, so that 15% looks quite suspicious to me. If it's real then an easy-to-run testcase would be nice.
Ingo
| |