@@ -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
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(-) -- 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