This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Add Prefer_MAP_32BIT_EXEC for Silvermont
- From: Florian Weimer <fweimer at redhat dot com>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: Zack Weinberg <zackw at panix dot com>, Andi Kleen <andi at firstfloor dot org>, Adhemerval Zanella <adhemerval dot zanella at linaro dot org>, GNU C Library <libc-alpha at sourceware dot org>
- Date: Sat, 12 Dec 2015 14:18:25 +0100
- Subject: Re: [PATCH] Add Prefer_MAP_32BIT_EXEC for Silvermont
- Authentication-results: sourceware.org; auth=none
- References: <20151211143706 dot GA7868 at intel dot com> <alpine dot DEB dot 2 dot 10 dot 1512111539300 dot 17023 at digraph dot polyomino dot org dot uk> <CAMe9rOqbqyFw3CMa35vwOEefdFq1xK2Q9hX8GXoGMKVZ-A2y0g at mail dot gmail dot com> <566AF894 dot 4060300 at linaro dot org> <CAMe9rOr-LypZXvq4Y4uwE_JybYoTXctZXMLjo4TH517NnC6omg at mail dot gmail dot com> <566B01BE dot 1070703 at linaro dot org> <CAKCAbMhMArQ9wsXhw2y+Fvv+_3O5i4g8pdDQdWo6_1YxqfVxkQ at mail dot gmail dot com> <CAMe9rOrVjSnhp-EzmAnVBg10wbqk9U4n+hL-3xF5=DPZP5co1A at mail dot gmail dot com> <CAKCAbMhk69hUBbrQ=0j0NDYjRT6R-EK1+F43+Mmi9FwS7epexQ at mail dot gmail dot com> <CAKCAbMhA6x4r6Bhw8cnAavoWzjWsm6WM8JPzrnCsrqxbEswS_g at mail dot gmail dot com> <87egeszoq3 dot fsf at tassilo dot jf dot intel dot com> <CAKCAbMibxh54DGJ6p59ah6jQK=oFkH9LCZo0UDBefQeh1y-5eg at mail dot gmail dot com> <CAMe9rOoyc4MXVz74GmjViora-iCEKitQXbpX9EB+Ln1Q_DAD_A at mail dot gmail dot com>
On 12/11/2015 10:25 PM, H.J. Lu wrote:
> 3% speedup is for my typical workloads, which is running GCC. For
> an artificial benchmark, I got
>
> Old glibc:
>
> [hjl@gnu-slm-1 dlcall]$
> /export/build/gnu/glibc/build-x86_64-linux.old/elf/ld-linux-x86-64.so.2
> --library-path /export/build/gnu/glibc/build-x86_64-linux
> intel64/dlcall
>
> Time for 1000000 calls into dynamic library 1.00 , 22.83 MT,
> 22.83 MT, 22.83 MT, 0 T
>
> New glibc:
>
> [hjl@gnu-slm-1 dlcall]$
> /export/build/gnu/glibc/build-x86_64-linux/elf/ld-linux-x86-64.so.2
> --library-path /export/build/gnu/glibc/build-x86_64-linux
> intel64/dlcall
>
> Time for 1000000 calls into dynamic library 1.00 , 2.87 MT,
> 2.87 MT, 2.87 MT, 0 T
>
> which is 8X speedup.
Ouch. Can't you label those chips as 32-bit-only?
What does your benchmark look like?
Can you run it with PIE? And please also measure the performance of
gettimeofday and sched_getcpu.
I suspect you need kernel support to avoid the performance hit.
Florian