* RE: [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes [not found] <1373675551-18684-1-git-send-email-andi@firstfloor.org> @ 2013-07-15 8:15 ` Kirill A. Shutemov 2013-07-15 12:16 ` Hitoshi Mitake 2013-07-16 10:11 ` Kirill A. Shutemov 1 sibling, 1 reply; 5+ messages in thread From: Kirill A. Shutemov @ 2013-07-15 8:15 UTC (permalink / raw) To: Andi Kleen; +Cc: acme, linux-kernel, Andi Kleen, h.mitake, kirill.shutemov Andi Kleen wrote: > From: Andi Kleen <ak@linux.intel.com> > > glibc calloc() has an optimization to not explicitely memset() > very large calloc allocations that just came from mmap(), > because they are known to be zero. > > This could result in the perf memcpy benchmark reading only from > the zero page, which gives unrealistic results. > > Always call memset explicitly on the source area to avoid this problem. > > Cc: h.mitake@gmail.com > Cc: kirill.shutemov@linux.intel.com > Signed-off-by: Andi Kleen <ak@linux.intel.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> > --- > tools/perf/bench/mem-memcpy.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c > index 93c83e3..690f75f 100644 > --- a/tools/perf/bench/mem-memcpy.c > +++ b/tools/perf/bench/mem-memcpy.c > @@ -117,6 +117,8 @@ static void alloc_mem(void **dst, void **src, size_t length) > *src = zalloc(length); > if (!src) > die("memory allocation failed - maybe length is too large?\n"); > + /* Make sure to always replace the zero pages even if MMAP_THRESH is crossed */ > + memset(src, 0, length); > } > > static u64 do_memcpy_cycle(memcpy_t fn, size_t len, bool prefault) > -- > 1.8.1.4 -- Kirill A. Shutemov ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes 2013-07-15 8:15 ` [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes Kirill A. Shutemov @ 2013-07-15 12:16 ` Hitoshi Mitake 0 siblings, 0 replies; 5+ messages in thread From: Hitoshi Mitake @ 2013-07-15 12:16 UTC (permalink / raw) To: Kirill A. Shutemov Cc: Andi Kleen, acme, linux-kernel, Andi Kleen, h.mitake, kirill.shutemov At Mon, 15 Jul 2013 11:15:32 +0300 (EEST), Kirill A. Shutemov wrote: > > Andi Kleen wrote: > > From: Andi Kleen <ak@linux.intel.com> > > > > glibc calloc() has an optimization to not explicitely memset() > > very large calloc allocations that just came from mmap(), > > because they are known to be zero. > > > > This could result in the perf memcpy benchmark reading only from > > the zero page, which gives unrealistic results. > > > > Always call memset explicitly on the source area to avoid this problem. > > > > Cc: h.mitake@gmail.com > > Cc: kirill.shutemov@linux.intel.com > > Signed-off-by: Andi Kleen <ak@linux.intel.com> > > Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Thanks for this fix, Andi. Acked-by: Hitoshi Mitake <mitake.hitoshi@lab.ntt.co.jp> > > > --- > > tools/perf/bench/mem-memcpy.c | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c > > index 93c83e3..690f75f 100644 > > --- a/tools/perf/bench/mem-memcpy.c > > +++ b/tools/perf/bench/mem-memcpy.c > > @@ -117,6 +117,8 @@ static void alloc_mem(void **dst, void **src, size_t length) > > *src = zalloc(length); > > if (!src) > > die("memory allocation failed - maybe length is too large?\n"); > > + /* Make sure to always replace the zero pages even if MMAP_THRESH is crossed */ > > + memset(src, 0, length); > > } > > > > static u64 do_memcpy_cycle(memcpy_t fn, size_t len, bool prefault) > > -- > > 1.8.1.4 > > -- > Kirill A. Shutemov > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ ^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes [not found] <1373675551-18684-1-git-send-email-andi@firstfloor.org> 2013-07-15 8:15 ` [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes Kirill A. Shutemov @ 2013-07-16 10:11 ` Kirill A. Shutemov 2013-07-16 13:14 ` Hitoshi Mitake 2013-07-16 14:20 ` Andi Kleen 1 sibling, 2 replies; 5+ messages in thread From: Kirill A. Shutemov @ 2013-07-16 10:11 UTC (permalink / raw) To: Andi Kleen; +Cc: acme, linux-kernel, Andi Kleen, h.mitake, kirill.shutemov Andi Kleen wrote: > From: Andi Kleen <ak@linux.intel.com> > > glibc calloc() has an optimization to not explicitely memset() > very large calloc allocations that just came from mmap(), > because they are known to be zero. > > This could result in the perf memcpy benchmark reading only from > the zero page, which gives unrealistic results. > > Always call memset explicitly on the source area to avoid this problem. > > Cc: h.mitake@gmail.com > Cc: kirill.shutemov@linux.intel.com > Signed-off-by: Andi Kleen <ak@linux.intel.com> > --- > tools/perf/bench/mem-memcpy.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c > index 93c83e3..690f75f 100644 > --- a/tools/perf/bench/mem-memcpy.c > +++ b/tools/perf/bench/mem-memcpy.c > @@ -117,6 +117,8 @@ static void alloc_mem(void **dst, void **src, size_t length) > *src = zalloc(length); > if (!src) > die("memory allocation failed - maybe length is too large?\n"); > + /* Make sure to always replace the zero pages even if MMAP_THRESH is crossed */ > + memset(src, 0, length); It should be memset(*src, 0, length) instead. There's fix for wrong memory allocation fail check in v3.11-rc1. -- Kirill A. Shutemov ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes 2013-07-16 10:11 ` Kirill A. Shutemov @ 2013-07-16 13:14 ` Hitoshi Mitake 2013-07-16 14:20 ` Andi Kleen 1 sibling, 0 replies; 5+ messages in thread From: Hitoshi Mitake @ 2013-07-16 13:14 UTC (permalink / raw) To: Kirill A. Shutemov Cc: Andi Kleen, acme, linux-kernel, Andi Kleen, h.mitake, kirill.shutemov At Tue, 16 Jul 2013 13:11:00 +0300 (EEST), Kirill A. Shutemov wrote: > > Andi Kleen wrote: > > From: Andi Kleen <ak@linux.intel.com> > > > > glibc calloc() has an optimization to not explicitely memset() > > very large calloc allocations that just came from mmap(), > > because they are known to be zero. > > > > This could result in the perf memcpy benchmark reading only from > > the zero page, which gives unrealistic results. > > > > Always call memset explicitly on the source area to avoid this problem. > > > > Cc: h.mitake@gmail.com > > Cc: kirill.shutemov@linux.intel.com > > Signed-off-by: Andi Kleen <ak@linux.intel.com> > > --- > > tools/perf/bench/mem-memcpy.c | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c > > index 93c83e3..690f75f 100644 > > --- a/tools/perf/bench/mem-memcpy.c > > +++ b/tools/perf/bench/mem-memcpy.c > > @@ -117,6 +117,8 @@ static void alloc_mem(void **dst, void **src, size_t length) > > *src = zalloc(length); > > if (!src) > > die("memory allocation failed - maybe length is too large?\n"); > > + /* Make sure to always replace the zero pages even if MMAP_THRESH is crossed */ > > + memset(src, 0, length); > > It should be memset(*src, 0, length) instead. > > There's fix for wrong memory allocation fail check in v3.11-rc1. I couldn't catch it, thanks. The memset() should be fixed. Thanks, Hitoshi ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes 2013-07-16 10:11 ` Kirill A. Shutemov 2013-07-16 13:14 ` Hitoshi Mitake @ 2013-07-16 14:20 ` Andi Kleen 1 sibling, 0 replies; 5+ messages in thread From: Andi Kleen @ 2013-07-16 14:20 UTC (permalink / raw) To: Kirill A. Shutemov Cc: Andi Kleen, acme, linux-kernel, Andi Kleen, h.mitake, kirill.shutemov > > @@ -117,6 +117,8 @@ static void alloc_mem(void **dst, void **src, size_t length) > > *src = zalloc(length); > > if (!src) > > die("memory allocation failed - maybe length is too large?\n"); > > + /* Make sure to always replace the zero pages even if MMAP_THRESH is crossed */ > > + memset(src, 0, length); > > It should be memset(*src, 0, length) instead. Thanks for catching. I wonder why it fixed the problem for me. > There's fix for wrong memory allocation fail check in v3.11-rc1. Doubt it matters. -Andi -- ak@linux.intel.com -- Speaking for myself only ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2013-07-16 14:20 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <1373675551-18684-1-git-send-email-andi@firstfloor.org>
2013-07-15 8:15 ` [PATCH] perf, tools, bench: Fix memcpy benchmark for large sizes Kirill A. Shutemov
2013-07-15 12:16 ` Hitoshi Mitake
2013-07-16 10:11 ` Kirill A. Shutemov
2013-07-16 13:14 ` Hitoshi Mitake
2013-07-16 14:20 ` Andi Kleen
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox