mirror/gmsm
1
0
Fork 0
mirror of https://github.com/emmansun/gmsm.git synced 2025-04-27 04:36:19 +08:00
Code Issues Packages Projects Releases Wiki Activity

MAGIC - optimize sm2P256SqrInternal

Browse Source
This commit is contained in:
emmansun 2021-02-12 12:59:25 +08:00
parent 885cb37c5a
commit aa6b436714
1 changed files with 72 additions and 84 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

156
sm2/p256_asm_amd64.s
View File

@ -1753,93 +1753,81 @@ TEXT sm2P256SqrInternal(SB),NOSPLIT,$0
ADCQ mul0, t2 ADCQ mul0, t2
ADCQ DX, t3 ADCQ DX, t3
// First reduction step // First reduction step
MOVQ p256p<>+0x08(SB), mul0 MOVQ acc0, hlp
MULQ acc0
ADDQ acc0, acc1 MOVQ hlp, mul0
ADCQ $0, mul1 MOVQ hlp, mul1
ADDQ mul0, acc1 SHLQ $32, mul0
ADCQ $0, mul1 SHRQ $32, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x010(SB), mul0 ADDQ hlp, acc1
MULQ acc0 ADCQ $0, acc2
ADDQ hlp, acc2 SUBQ mul0, acc1
ADCQ $0, mul1 SBBQ mul1, acc2
ADDQ mul0, acc2
ADCQ $0, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x018(SB), mul0
MULQ acc0
ADDQ hlp, acc3 ADDQ hlp, acc3
ADCQ $0, mul1 ADCQ $0, acc0
ADDQ mul0, acc3 SUBQ mul0, acc3
ADCQ $0, mul1 SBBQ mul1, acc0
MOVQ mul1, acc0 SUBQ hlp, acc3
SBBQ $0, acc0
// Second reduction step // Second reduction step
MOVQ p256p<>+0x08(SB), mul0 MOVQ acc1, hlp
MULQ acc1
ADDQ acc1, acc2 MOVQ hlp, mul0
ADCQ $0, mul1 MOVQ hlp, mul1
ADDQ mul0, acc2 SHLQ $32, mul0
ADCQ $0, mul1 SHRQ $32, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x010(SB), mul0
MULQ acc1
ADDQ hlp, acc3
ADCQ $0, mul1
ADDQ mul0, acc3
ADCQ $0, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x018(SB), mul0
MULQ acc1
ADDQ hlp, acc0
ADCQ $0, mul1
ADDQ mul0, acc0
ADCQ $0, mul1
MOVQ mul1, acc1
// Third reduction step
MOVQ p256p<>+0x08(SB), mul0
MULQ acc2
ADDQ acc2, acc3
ADCQ $0, mul1
ADDQ mul0, acc3
ADCQ $0, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x010(SB), mul0
MULQ acc2
ADDQ hlp, acc0
ADCQ $0, mul1
ADDQ mul0, acc0
ADCQ $0, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x018(SB), mul0
MULQ acc2
ADDQ hlp, acc1
ADCQ $0, mul1
ADDQ mul0, acc1
ADCQ $0, mul1
MOVQ mul1, acc2
// Last reduction step
MOVQ p256p<>+0x08(SB), mul0
MULQ acc3
ADDQ acc3, acc0
ADCQ $0, mul1
ADDQ mul0, acc0
ADCQ $0, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x010(SB), mul0
MULQ acc3
ADDQ hlp, acc1
ADCQ $0, mul1
ADDQ mul0, acc1
ADCQ $0, mul1
MOVQ mul1, hlp
MOVQ p256p<>+0x018(SB), mul0
MULQ acc3
ADDQ hlp, acc2 ADDQ hlp, acc2
ADCQ $0, mul1 ADCQ $0, acc3
ADDQ mul0, acc2 SUBQ mul0, acc2
ADCQ $0, mul1 SBBQ mul1, acc3
MOVQ mul1, acc3
ADDQ hlp, acc0
ADCQ $0, acc1
SUBQ mul0, acc0
SBBQ mul1, acc1
SUBQ hlp, acc0
SBBQ $0, acc1
// Third reduction step
MOVQ acc2, hlp
MOVQ hlp, mul0
MOVQ hlp, mul1
SHLQ $32, mul0
SHRQ $32, mul1
ADDQ hlp, acc3
ADCQ $0, acc0
SUBQ mul0, acc3
SBBQ mul1, acc0
ADDQ hlp, acc1
ADCQ $0, acc2
SUBQ mul0, acc1
SBBQ mul1, acc2
SUBQ hlp, acc1
SBBQ $0, acc2
// Last reduction step
MOVQ acc3, hlp
MOVQ hlp, mul0
MOVQ hlp, mul1
SHLQ $32, mul0
SHRQ $32, mul1
ADDQ hlp, acc0
ADCQ $0, acc1
SUBQ mul0, acc0
SBBQ mul1, acc1
ADDQ hlp, acc2
ADCQ $0, acc3
SUBQ mul0, acc2
SBBQ mul1, acc3
SUBQ hlp, acc2
SBBQ $0, acc3
MOVQ $0, BP MOVQ $0, BP
// Add bits [511:256] of the result // Add bits [511:256] of the result
ADCQ acc0, t0 ADCQ acc0, t0