This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] always inline alloc_perturb.
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: David Miller <davem at davemloft dot net>
- Cc: roland at hack dot frob dot com, libc-alpha at sourceware dot org
- Date: Fri, 19 Dec 2014 16:14:09 +0100
- Subject: Re: [PATCH] always inline alloc_perturb.
- Authentication-results: sourceware.org; auth=none
- References: <20141212111856 dot GA8768 at domone> <20141212 dot 093417 dot 507023024907812119 dot davem at davemloft dot net> <20141212155051 dot GA7878 at domone> <20141212 dot 105326 dot 1674361735913519795 dot davem at davemloft dot net>
On Fri, Dec 12, 2014 at 10:53:26AM -0500, David Miller wrote:
> From: OndÅej BÃlka <neleai@seznam.cz>
> Date: Fri, 12 Dec 2014 16:50:51 +0100
>
> > And is worse performance in next ten years until that issues are fixed
> > acceptable?
>
> In moderate amounts, it's no big deal.
You have no guarantee that its just moderate amount. From your warnings
it looks that compiler decided never inline anything. So how much does
malloc_benchtest differ after you build with ./configure --disable-werror?
On x86 simulating that with noinline gives 10% performance degradation
which as half of total time spend in libc is in malloc+variants.