This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH] Simple malloc benchtest.
- From: Siddhesh Poyarekar <siddhesh at redhat dot com>
- To: OndÅej BÃlka <neleai at seznam dot cz>
- Cc: GNU C Library <libc-alpha at sourceware dot org>
- Date: Tue, 31 Dec 2013 18:16:52 +0530
- Subject: Re: [PATCH] Simple malloc benchtest.
- Authentication-results: sourceware.org; auth=none
- References: <20131221153303 dot GA8420 at domone dot podge> <20131223090627 dot GF4979 at spoyarek dot pnq dot redhat dot com> <20131223095034 dot GA20816 at domone> <20131223110912 dot GG4979 at spoyarek dot pnq dot redhat dot com> <20131223133157 dot GB20816 at domone> <CAAHN_R2+N432brpObQ9RK4cJgjcqvpf+gtF7J4b1wwMShFQu7A at mail dot gmail dot com> <20131228011928 dot GB15270 at domone> <20131231060303 dot GA5374 at spoyarek dot pnq dot redhat dot com> <20131231113841 dot GA17877 at domone>
On Tue, Dec 31, 2013 at 12:38:41PM +0100, OndÅej BÃlka wrote:
> > I wonder how relevant the comparison in memory usage is in that blog
> > post, given that the bloated 'memory usage' may actually just be
> > address space usage due to the per-thread arenas, where most of each
> > arena is unused and doesn't even have any pages backing the address
> > space. In other words, I believe more information is needed from that
> > test.
> Still when average user discovers that we use more memory than others it
> does not look good.
There's a very significant problem of user education here though,
especially when they're looking at address space usage and screaming
memory usage. I've seen first hand some such users complaining about
address space usage and then upon knowing the difference between
address space and memory, going back happier with the performance.