This patch cleanups the multiarch bzero for powerpc64 by remove
the multiarch objects and use instead the the memset embedded
implementation presented in each multiarch optimization. The
code generate is essentially the same, but the TB_TOCLESS (which
is not essential).
http://sourceware.org/ml/libc-alpha/2013-08/msg00104.html
One of the things I noticed when looking at power7 timing is that rlwimi
is cracked and the two resulting insns have a register dependency.
That makes it a little slower than the equivalent rldimi.
* sysdeps/powerpc/powerpc64/memset.S: Replace rlwimi with
insrdi. Formatting.
* sysdeps/powerpc/powerpc64/power4/memset.S: Likewise.
* sysdeps/powerpc/powerpc64/power6/memset.S: Likewise.
* sysdeps/powerpc/powerpc64/power7/memset.S: Likewise.
* sysdeps/powerpc/powerpc32/power4/memset.S: Likewise.
* sysdeps/powerpc/powerpc32/power6/memset.S: Likewise.
* sysdeps/powerpc/powerpc32/power7/memset.S: Likewise.
http://sourceware.org/ml/libc-alpha/2013-08/msg00103.html
LIttle-endian support for memcpy. I spent some time cleaning up the
64-bit power7 memcpy, in order to avoid the extra alignment traps
power7 takes for little-endian. It probably would have been better
to copy the linux kernel version of memcpy.
* sysdeps/powerpc/powerpc32/power4/memcpy.S: Add little endian support.
* sysdeps/powerpc/powerpc32/power6/memcpy.S: Likewise.
* sysdeps/powerpc/powerpc32/power7/memcpy.S: Likewise.
* sysdeps/powerpc/powerpc32/power7/mempcpy.S: Likewise.
* sysdeps/powerpc/powerpc64/memcpy.S: Likewise.
* sysdeps/powerpc/powerpc64/power4/memcpy.S: Likewise.
* sysdeps/powerpc/powerpc64/power6/memcpy.S: Likewise.
* sysdeps/powerpc/powerpc64/power7/memcpy.S: Likewise.
* sysdeps/powerpc/powerpc64/power7/mempcpy.S: Likewise. Make better
use of regs. Use power7 mtocrf. Tidy function tails.
Initially based on the versions found in wcsmbs/* ; these files have
been changed by hand unrolling, and adding some additional variables
to allow some read-ahead to occur, which then relieves some of the
wait-for-increment/wait-for-load/wait-for-compare-results pressure
that was slowing down every iteration through the while-loop.
For 64-bit Power7, These changes give an approx 20% throughput boost
for the wcschr and wcsrchr functions; and approx 40% boost for the
wcscpy function. 32-bit improvements appear to be slightly better
with ~ %30 and ~ %45 respectively. Results for Power6 closely match
those for power7.
This patch tries to organize the implies files for ppc, since there are
a number of processors and most of them are compatible with each other
(backwards compatible).
Having in mind that we start the search for processor-specific files in
the sysdeps/unix/sysv/linux tree
(sysdeps/unix/sysv/linux/powerpc/powerpc[32|64]/[processor]/fpu to be
exact), we would like to grab any linux-specific code from that tree
prior to going through the other tree (sysdeps/powerpc/...).
For that, i removed the Implies files that were originally inside the
fpu directories and placed then in the non-fpu directories (still inside
the unix/sysv/linux tree). If no processor-specific/linux-specific files
could be found, we "imply" the other tree's (sysdeps/powerpc/...) fpu
directory for that specific processor AND also the non-fpu directory for
that same tree.
If, again, no processor-specific code is found, we read another Implies
file that will point to the most compatible processor that we should
grab code from, and so on, until we reach the power4 processor.
So, in summary, the Implies files will live inside these directories
now:
* sysdeps/unix/sysv/linux/powerpc/powerpc[32|64]/[processor]
* sysdeps/powerpc/powerpc[32|64]/[processor]
Practical example of the order we will use to pick power6-specific code
with the new structure.
sysdeps/unix/sysv/linux/powerpc/powerpc[32|64]/power6/fpu ->
sysdeps/unix/sysv/linux/powerpc/powerpc[32|64]/power6 ->
sysdeps/powerpc/powerpc[32|64]/power6/fpu ->
sysdeps/powerpc/powerpc[32|64]/power6 ->
sysdeps/powerpc/powerpc[32|64]/power5+/fpu ->
sysdeps/powerpc/powerpc[32|64]/power5+ ->
sysdeps/powerpc/powerpc[32|64]/power5/fpu ->
sysdeps/powerpc/powerpc[32|64]/power5 ->
sysdeps/powerpc/powerpc[32|64]/power4/fpu ->
sysdeps/powerpc/powerpc[32|64]/power4 (from here, it'll go to the
generic path as usual)
* sysdeps/powerpc/powerpc32/power5/fpu/s_isnanf.S: New file.
* sysdeps/powerpc/powerpc32/power6/fpu/s_isnan.S: New file.
* sysdeps/powerpc/powerpc32/power6/fpu/s_isnanf.S: New file.
* sysdeps/powerpc/powerpc64/power5/fpu/s_isnan.S: New file.
* sysdeps/powerpc/powerpc64/power6/fpu/s_isnan.S: New file.
* sysdeps/powerpc/powerpc64/power6x/fpu/s_isnan.S: New file.
* sysdeps/powerpc/powerpc32/power5/fpu/Implies: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/Implies: New file.
* sysdeps/powerpc/powerpc32/power6/fpu/Implies: New file.
* sysdeps/powerpc/powerpc32/power6x/fpu/Implies: New file.
* sysdeps/powerpc/powerpc64/970/fpu/Implies: New file.
* sysdeps/powerpc/powerpc64/power5/fpu/Implies: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/Implies: New file.
* sysdeps/powerpc/powerpc64/power6/fpu/Implies: New file.
* sysdeps/powerpc/powerpc64/power6x/fpu/Implies: New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/970/fpu/Implies: New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power4/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power5/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power5+/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power6/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/power6x/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/970/fpu/Implies: New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power4/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power5/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power5+/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power6/fpu/Implies:
New file.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/power6x/fpu/Implies:
New file.
2007-05-31 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llrint.S: Move.
* sysdeps/powerpc/powerpc32/power4/fpu/s_llrint.S: To here.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llrintf.S: Move.
* sysdeps/powerpc/powerpc32/power4/fpu/s_llrintf.S: To here.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llround.S: Move.
* sysdeps/powerpc/powerpc32/power4/fpu/s_llround.S: To here.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llroundf.S: Move.
* sysdeps/powerpc/powerpc32/power4/fpu/s_llroundf.S: To here.
2007-05-22 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power5+/fpu/s_round.S
(LONG_DOUBLE_COMPAT): Specify correct version, GLIBC_2_1.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_trunc.S
(LONG_DOUBLE_COMPAT): Specify correct version, GLIBC_2_1.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_round.S
(LONG_DOUBLE_COMPAT): Specify correct version, GLIBC_2_1.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_trunc.S
(LONG_DOUBLE_COMPAT): Specify correct version, GLIBC_2_1.
2007-05-21 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power4/fpu/slowexp.c: New file.
* sysdeps/powerpc/powerpc32/power4/fpu/w_sqrt.c: New file.
* sysdeps/powerpc/powerpc64/power4/fpu/slowexp.c: New file.
* sysdeps/powerpc/powerpc64/power4/fpu/w_sqrt.c: New file.
2007-03-15 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llrint.S
[LONG_DOUBLE_COMPAT]: Add compat_symbol for llrintl@@GLIBC_2_1.
2006-02-13 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power6/fpu/s_llrint.S: New File
* sysdeps/powerpc/powerpc32/power6/fpu/s_llrintf.S: New File
* sysdeps/powerpc/powerpc32/power6/fpu/s_llround.S: New File
* sysdeps/powerpc/powerpc32/power6/fpu/s_llroundf.S: New File
2006-10-20 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power4/fpu/slowpow.c: New file.
* sysdeps/powerpc/powerpc64/power4/fpu/slowpow.c: New file.
2006-10-03 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llround.S: New file.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llroundf.S: New file.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/Makefile: Moved.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/mpa.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/Makefile: To here.
* sysdeps/powerpc/powerpc32/power4/fpu/mpa.c: Likewise.
2006-09-29 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power6x/fpu/s_lrint.S: New file.
* sysdeps/powerpc/powerpc32/power6x/fpu/s_lround.S: New file.
* sysdeps/powerpc/powerpc64/power6x/fpu/s_llrint.S: New file.
* sysdeps/powerpc/powerpc64/power6x/fpu/s_llround.S: New file.
2006-09-28 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power5+/fpu/s_llround.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_llroundf.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_lround.S: New file.
* sysdeps/powerpc/powerpc32/power6x/fpu/Implies: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_llround.S: New file.
* sysdeps/powerpc/powerpc64/power6x/fpu/Implies: New file.
2006-08-31 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/powerpc64/fpu/Makefile: New file.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/mpa.c: New file.
* sysdeps/powerpc/powerpc64/power4/fpu/Makefile: New file.
* sysdeps/powerpc/powerpc64/power4/fpu/mpa.c: New file.
2006-06-15 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power5+/fpu/s_ceil.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_ceilf.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_floor.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_floorf.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_round.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_roundf.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_trunc.S: New file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_truncf.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_ceil.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_ceilf.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_floor.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_floorf.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_round.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_roundf.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_trunc.S: New file.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_truncf.S: New file.
2006-03-20 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llrint.S: New file.
* sysdeps/powerpc/powerpc32/powerpc64/fpu/s_llrintf.S: New file.
2007-06-01 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power6/memset.S: New file.
* sysdeps/powerpc/powerpc64/power6/memset.S: New file.
2007-05-31 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/970/Implies: New file.
* sysdeps/powerpc/powerpc32/power5/Implies: New file.
* sysdeps/powerpc/powerpc32/power5+/Implies: New file.
* sysdeps/powerpc/powerpc32/power6/Implies: New file.
* sysdeps/powerpc/powerpc32/power6x/Implies: New file.
* sysdeps/powerpc/powerpc64/970/Implies: New file.
* sysdeps/powerpc/powerpc64/power5/Implies: New file.
* sysdeps/powerpc/powerpc64/power5+/Implies: New file.
* sysdeps/powerpc/powerpc64/power6/Implies: New file.
* sysdeps/powerpc/powerpc64/power6x/Implies: New file.
2007-05-21 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power4/memset.S: New file
2007-03-13 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc64/memcpy.S: Improve aligned loop to minimize
branch miss-predicts. Ensure that cache line crossing does not impact
dispatch grouping.
2006-12-13 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc64/power4/memcopy.h: Replace with include
"../../powerpc32/power4/memcopy.h".
* sysdeps/powerpc/powerpc64/power4/wordcopy.c: Replace with include
"../../powerpc32/power4/wordcopy.c".
2006-10-03 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/powerpc64/Makefile: Moved.
* sysdeps/powerpc/powerpc32/powerpc64/memcopy.h: Likewise.
* sysdeps/powerpc/powerpc32/powerpc64/wordcopy.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/Makefile: To here.
* sysdeps/powerpc/powerpc32/power4/memcopy.h: Likewise.
* sysdeps/powerpc/powerpc32/power4/wordcopy.c: Likewise.
2006-09-10 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power6/memcpy.S: New file.
2006-08-31 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power6/wordcopy.c: New file.
* sysdeps/powerpc/powerpc32/powerpc64/Makefile: New file.
* sysdeps/powerpc/powerpc32/powerpc64/memcopy.h: New file.
* sysdeps/powerpc/powerpc32/powerpc64/wordcopy.c: New file.
* sysdeps/powerpc/powerpc64/power4/Makefile: New file.
* sysdeps/powerpc/powerpc64/power4/memcopy.h: New file.
* sysdeps/powerpc/powerpc64/power4/wordcopy.c: New file.
* sysdeps/powerpc/powerpc64/power6/wordcopy.c: New file.
2006-07-06 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc64/power6/memcpy.S: New file.
2006-03-20 Steven Munroe <sjmunroe@us.ibm.com>
* sysdeps/powerpc/powerpc32/power4/memcmp.S: New file.
* sysdeps/powerpc/powerpc32/power4/memcpy.S: New file.
* sysdeps/powerpc/powerpc32/power4/memset.S: New file.
* sysdeps/powerpc/powerpc32/power4/strncmp.S: New file.
* sysdeps/powerpc/powerpc64/power4/memcmp.S: New file.
* sysdeps/powerpc/powerpc64/power4/memcpy.S: New file.
* sysdeps/powerpc/powerpc64/power4/strncmp.S: New file.
