Messages in this thread | | | From | sunrui <> | Subject | Re: [PATCH] arm64/lib: improve CRC32 performance for deep pipelines | Date | Wed, 28 Nov 2018 03:01:22 +0000 |
| |
The code execute on my system performance improves from # time ./old
real 0m6.645s user 0m6.644s sys 0m0.001s
to
# time ./new
real 0m5.565s user 0m5.565s sys 0m0.000s
> Improve the performance of the crc32() asm routines by getting rid of most of the branches and small sized loads on the common path. > > Instead, use a branchless code path involving overlapping 16 byte loads to process the first (length % 32) bytes, and process the remainder using a loop that processes 32 bytes at a time. > > Tested using the following test program: > > #include <stdlib.h> > > extern void crc32_le(unsigned short, char const*, int); > > int main(void) > { > static const char buf[4096]; > > srand(20181126); > > for (int i = 0; i < 100 * 1000 * 1000; i++) > crc32_le(0, buf, rand() % 1024); > > return 0; > } > > On Cortex-A53 and Cortex-A57, the performance regresses but only very slightly. On Cortex-A72 however, the performance improves from > > $ time ./crc32 > > real 0m10.149s > user 0m10.149s > sys 0m0.000s > > to > > $ time ./crc32 > > real 0m7.915s > user 0m7.915s > sys 0m0.000s > > Cc: Rui Sun <sunrui26@huawei.com> > Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> > --- > Cortex-A57 tcrypt results after the patch. > > I ran Rui's code [0] as well. On Cortex-A57, performance regresses a bit more (but not dramatically). On Cortex-A72, it executes at > > $ time ./crc32 > > real 0m9.625s > user 0m9.625s > sys 0m0.000s > > Rui, can you please benchmark this code on your system as well? > > [0] https://lore.kernel.org/lkml/1542612560-10089-1-git-send-email-sunrui26@huawei.com/ > > arch/arm64/lib/crc32.S | 54 ++++++++++++++++++-- > 1 file changed, 49 insertions(+), 5 deletions(-) > > diff --git a/arch/arm64/lib/crc32.S b/arch/arm64/lib/crc32.S index 5bc1e85b4e1c..f132f2a7522e 100644 > --- a/arch/arm64/lib/crc32.S > +++ b/arch/arm64/lib/crc32.S > @@ -15,15 +15,59 @@ > .cpu generic+crc > > .macro __crc32, c > -0: subs x2, x2, #16 > - b.mi 8f > - ldp x3, x4, [x1], #16 > + cmp x2, #16 > + b.lt 8f // less than 16 bytes > + > + and x7, x2, #0x1f > + and x2, x2, #~0x1f > + cbz x7, 32f // multiple of 32 bytes > + > + and x8, x7, #0xf > + ldp x3, x4, [x1] > + add x8, x8, x1 > + add x1, x1, x7 > + ldp x5, x6, [x8] > CPU_BE( rev x3, x3 ) > CPU_BE( rev x4, x4 ) > +CPU_BE( rev x5, x5 ) > +CPU_BE( rev x6, x6 ) > + > + tst x7, #8 > + crc32\c\()x w8, w0, x3 > + csel x3, x3, x4, eq > + csel w0, w0, w8, eq > + tst x7, #4 > + lsr x4, x3, #32 > + crc32\c\()w w8, w0, w3 > + csel x3, x3, x4, eq > + csel w0, w0, w8, eq > + tst x7, #2 > + lsr w4, w3, #16 > + crc32\c\()h w8, w0, w3 > + csel w3, w3, w4, eq > + csel w0, w0, w8, eq > + tst x7, #1 > + crc32\c\()b w8, w0, w3 > + csel w0, w0, w8, eq > + tst x7, #16 > + crc32\c\()x w8, w0, x5 > + crc32\c\()x w8, w8, x6 > + csel w0, w0, w8, eq > + cbz x2, 0f > + > +32: ldp x3, x4, [x1], #32 > + sub x2, x2, #32 > + ldp x5, x6, [x1, #-16] > +CPU_BE( rev x3, x3 ) > +CPU_BE( rev x4, x4 ) > +CPU_BE( rev x5, x5 ) > +CPU_BE( rev x6, x6 ) > crc32\c\()x w0, w0, x3 > crc32\c\()x w0, w0, x4 > - b.ne 0b > - ret > + crc32\c\()x w0, w0, x5 > + crc32\c\()x w0, w0, x6 > + cbnz x2, 32b > +0: ret > > 8: tbz x2, #3, 4f > ldr x3, [x1], #8 > -- > 2.19.1 > > > BEFORE testing speed of async crc32c (crc32c-generic) > tcrypt: test 0 ( 16 byte blocks, 16 bytes per update, 1 updates): 35416299 opers/sec, 566660784 bytes/sec > tcrypt: test 1 ( 64 byte blocks, 16 bytes per update, 4 updates): 5342888 opers/sec, 341944832 bytes/sec > tcrypt: test 2 ( 64 byte blocks, 64 bytes per update, 1 updates): 30056634 opers/sec, 1923624576 bytes/sec > tcrypt: test 3 ( 256 byte blocks, 16 bytes per update, 16 updates): 1543567 opers/sec, 395153152 bytes/sec > tcrypt: test 4 ( 256 byte blocks, 64 bytes per update, 4 updates): 4865198 opers/sec, 1245490688 bytes/sec > tcrypt: test 5 ( 256 byte blocks, 256 bytes per update, 1 updates): 12709474 opers/sec, 3253625344 bytes/sec > tcrypt: test 6 ( 1024 byte blocks, 16 bytes per update, 64 updates): 401746 opers/sec, 411387904 bytes/sec > tcrypt: test 7 ( 1024 byte blocks, 256 bytes per update, 4 updates): 2576764 opers/sec, 2638606336 bytes/sec > tcrypt: test 8 ( 1024 byte blocks, 1024 bytes per update, 1 updates): 4464109 opers/sec, 4571247616 bytes/sec > tcrypt: test 9 ( 2048 byte blocks, 16 bytes per update, 128 updates): 202236 opers/sec, 414179328 bytes/sec > tcrypt: test 10 ( 2048 byte blocks, 256 bytes per update, 8 updates): 1344017 opers/sec, 2752546816 bytes/sec > tcrypt: test 11 ( 2048 byte blocks, 1024 bytes per update, 2 updates): 2000544 opers/sec, 4097114112 bytes/sec > tcrypt: test 12 ( 2048 byte blocks, 2048 bytes per update, 1 updates): 2395890 opers/sec, 4906782720 bytes/sec > tcrypt: test 13 ( 4096 byte blocks, 16 bytes per update, 256 updates): 101569 opers/sec, 416026624 bytes/sec > tcrypt: test 14 ( 4096 byte blocks, 256 bytes per update, 16 updates): 687876 opers/sec, 2817540096 bytes/sec > tcrypt: test 15 ( 4096 byte blocks, 1024 bytes per update, 4 updates): 1029042 opers/sec, 4214956032 bytes/sec > tcrypt: test 16 ( 4096 byte blocks, 4096 bytes per update, 1 updates): 1206227 opers/sec, 4940705792 bytes/sec > tcrypt: test 17 ( 8192 byte blocks, 16 bytes per update, 512 updates): 50842 opers/sec, 416497664 bytes/sec > tcrypt: test 18 ( 8192 byte blocks, 256 bytes per update, 32 updates): 347779 opers/sec, 2849005568 bytes/sec > tcrypt: test 19 ( 8192 byte blocks, 1024 bytes per update, 8 updates): 525054 opers/sec, 4301242368 bytes/sec > tcrypt: test 20 ( 8192 byte blocks, 4096 bytes per update, 2 updates): 600919 opers/sec, 4922728448 bytes/sec > tcrypt: test 21 ( 8192 byte blocks, 8192 bytes per update, 1 updates): 606954 opers/sec, 4972167168 bytes/sec > > AFTER testing speed of async crc32c (crc32c-generic) > tcrypt: test 0 ( 16 byte blocks, 16 bytes per update, 1 updates): 30535173 opers/sec, 488562768 bytes/sec > tcrypt: test 1 ( 64 byte blocks, 16 bytes per update, 4 updates): 4798401 opers/sec, 307097664 bytes/sec > tcrypt: test 2 ( 64 byte blocks, 64 bytes per update, 1 updates): 30061075 opers/sec, 1923908800 bytes/sec > tcrypt: test 3 ( 256 byte blocks, 16 bytes per update, 16 updates): 1359905 opers/sec, 348135680 bytes/sec > tcrypt: test 4 ( 256 byte blocks, 64 bytes per update, 4 updates): 4862043 opers/sec, 1244683008 bytes/sec > tcrypt: test 5 ( 256 byte blocks, 256 bytes per update, 1 updates): 14375092 opers/sec, 3680023552 bytes/sec > tcrypt: test 6 ( 1024 byte blocks, 16 bytes per update, 64 updates): 351936 opers/sec, 360382464 bytes/sec > tcrypt: test 7 ( 1024 byte blocks, 256 bytes per update, 4 updates): 2665564 opers/sec, 2729537536 bytes/sec > tcrypt: test 8 ( 1024 byte blocks, 1024 bytes per update, 1 updates): 4467924 opers/sec, 4575154176 bytes/sec > tcrypt: test 9 ( 2048 byte blocks, 16 bytes per update, 128 updates): 177021 opers/sec, 362539008 bytes/sec > tcrypt: test 10 ( 2048 byte blocks, 256 bytes per update, 8 updates): 1414689 opers/sec, 2897283072 bytes/sec > tcrypt: test 11 ( 2048 byte blocks, 1024 bytes per update, 2 updates): 1995413 opers/sec, 4086605824 bytes/sec > tcrypt: test 12 ( 2048 byte blocks, 2048 bytes per update, 1 updates): 2393630 opers/sec, 4902154240 bytes/sec > tcrypt: test 13 ( 4096 byte blocks, 16 bytes per update, 256 updates): 88758 opers/sec, 363552768 bytes/sec > tcrypt: test 14 ( 4096 byte blocks, 256 bytes per update, 16 updates): 731752 opers/sec, 2997256192 bytes/sec > tcrypt: test 15 ( 4096 byte blocks, 1024 bytes per update, 4 updates): 1030393 opers/sec, 4220489728 bytes/sec > tcrypt: test 16 ( 4096 byte blocks, 4096 bytes per update, 1 updates): 1205718 opers/sec, 4938620928 bytes/sec > tcrypt: test 17 ( 8192 byte blocks, 16 bytes per update, 512 updates): 44450 opers/sec, 364134400 bytes/sec > tcrypt: test 18 ( 8192 byte blocks, 256 bytes per update, 32 updates): 373236 opers/sec, 3057549312 bytes/sec > tcrypt: test 19 ( 8192 byte blocks, 1024 bytes per update, 8 updates): 524905 opers/sec, 4300021760 bytes/sec > tcrypt: test 20 ( 8192 byte blocks, 4096 bytes per update, 2 updates): 601242 opers/sec, 4925374464 bytes/sec > tcrypt: test 21 ( 8192 byte blocks, 8192 bytes per update, 1 updates): 606769 opers/sec, 4970651648 bytes/sec
| |