Re: [PATCH v4] arch/x86: Improve 'rep movs{b|q}' usage in memmove_64.S

Re: [PATCH v4] arch/x86: Improve 'rep movs{b|q}' usage in memmove_64.S

[Noah Goldstein]

On Thu, Feb 10, 2022 at 3:08 AM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
>
> On Tue, Jan 11, 2022 at 9:13 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> >
> > On Fri, Dec 10, 2021 at 12:35 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> > >
> > > On Fri, Nov 19, 2021 at 6:05 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> > > >
> > > > On Fri, Nov 19, 2021 at 4:31 PM David Laight <David.Laight@aculab.com> wrote:
> > > > >
> > > > > From: Noah Goldstein
> > > > > > Sent: 17 November 2021 22:45
> > > > > >
> > > > > > On Wed, Nov 17, 2021 at 4:31 PM David Laight <David.Laight@aculab.com> wrote:
> > > > > > >
> > > > > > > From: Noah Goldstein
> > > > > > > > Sent: 17 November 2021 21:03
> > > > > > > >
> > > > > > > > Add check for "short distance movsb" for forwards FSRM usage and
> > > > > > > > entirely remove backwards 'rep movsq'. Both of these usages hit "slow
> > > > > > > > modes" that are an order of magnitude slower than usual.
> > > > > > > >
> > > > > > > > 'rep movsb' has some noticeable VERY slow modes that the current
> > > > > > > > implementation is either 1) not checking for or 2) intentionally
> > > > > > > > using.
> > > > > > >
> > > > > > > How does this relate to the decision that glibc made a few years
> > > > > > > ago to use backwards 'rep movs' for non-overlapping copies?
> > > > > >
> > > > > > GLIBC doesn't use backwards `rep movs`.  Since the regions are
> > > > > > non-overlapping it just uses forward copy. Backwards `rep movs` is
> > > > > > from setting the direction flag (`std`) and is a very slow byte
> > > > > > copy. For overlapping regions where backwards copy is necessary GLIBC
> > > > > > uses 4x vec copy loop.
> > > > >
> > > > > Try to find this commit 6fb8cbcb58a29fff73eb2101b34caa19a7f88eba
> > > > >
> > > > > Or follow links from https://www.win.tue.nl/~aeb/linux/misc/gcc-semibug.html
> > > > > But I can't find the actual patch.
> > > > >
> > > > > The claims were a massive performance increase for the reverse copy.
> > > > >
> > > >
> > > > I don't think that's referring to optimizations around `rep movs`. It
> > > > appears to be referring to fallout from this patch:
> > > > https://sourceware.org/git/?p=glibc.git;a=commit;h=6fb8cbcb58a29fff73eb2101b34caa19a7f88eba
> > > >
> > > > which broken programs misusing `memcpy` with overlapping regions
> > > > resulting in this fix:
> > > > https://sourceware.org/git/?p=glibc.git;a=commit;h=0354e355014b7bfda32622e0255399d859862fcd
> > > >
> > > > AFAICT support for ERMS was only added around:
> > > > https://sourceware.org/git/?p=glibc.git;a=commit;h=13efa86ece61bf84daca50cab30db1b0902fe2db
> > > >
> > > > Either way GLIBC memcpy/memmove moment most certainly does not
> > > > use backwards `rep movs`:
> > > > https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S;hb=HEAD#l655
> > > >
> > > > as it is very slow.
> > > >
> > > > > The pdf from www.agner.org/optimize may well indicate why some
> > > > > copies are unexpectedly slow due to cache access aliasing.
> > > >
> > > > Even in the `4k` aliasing case `rep movsb` seems to stay within a
> > > > factor of 2 of optimal whereas the `std` backwards `rep movs` loses
> > > > by a factor of 10.
> > > >
> > > > Either way, `4k` aliasing detection is mostly a concern of `memcpy` as
> > > > the direction of copy for `memmove` is a correctness question, not
> > > > an optimization.
> > > >
> > > >
> > > > >
> > > > > I'm pretty sure that Intel cpu (possibly from Ivy bridge onwards)
> > > > > can be persuaded to copy 8 bytes/clock for in-cache data with
> > > > > a fairly simple loop that contains 2 reads (maybe misaligned)
> > > > > and two writes (so 16 bytes per iteration).
> > > > > Extra unrolling just adds extra code top and bottom.
> > > > >
> > > > > You might want a loop like:
> > > > >         1:      mov     0(%rsi, %rcx),%rax
> > > > >                 mov     8(%rsi, %rcx),%rdx
> > > > >                 mov     %rax, 0(%rdi, %rcx)
> > > > >                 mov     %rdx, 8(%rdi, %rcx)
> > > > >                 add     $16, %rcx
> > > > >                 jnz     1b
> > > > >
> > > > >         David
> > > >
> > > > The backwards loop already has 4x unrolled `movq` loop.
> > > ping.
> > ping.
> ping3.

Hi,

Anything I'm missing to get this looked at?

> > > >
> > > > >
> > > > > -
> > > > > Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> > > > > Registration No: 1397386 (Wales)

Re: [PATCH v4] arch/x86: Improve 'rep movs{b|q}' usage in memmove_64.S

[Borislav Petkov]

On Wed, Nov 17, 2021 at 03:02:45PM -0600, Noah Goldstein wrote:
> Add check for "short distance movsb" for forwards FSRM usage and
> entirely remove backwards 'rep movsq'. Both of these usages hit "slow
> modes" that are an order of magnitude slower than usual.
> 
> 'rep movsb' has some noticeable VERY slow modes that the current
> implementation is either 1) not checking for or 2) intentionally
> using.
> 
> All times are in cycles and measuring the throughput of copying 1024
> bytes.

All these claims need to be proven by

 - real benchmarks - not a microbenchmark - where it shows that
 modifications like that are not "in the noise". Others should be able
 to verify those results too.

 - on a bunch of CPUs from different vendors to verify that they don't
   cause performance regressions on any.

HTH.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette