📄 gemv_n_sse2_opteron.s
字号:
addpd %xmm12, %xmm4 movlpd 8 * SIZE(A2, LDA, 2), %xmm12 movhpd 9 * SIZE(A2, LDA, 2), %xmm12 mulpd %xmm2, %xmm8 addpd %xmm13, %xmm5 movlpd 10 * SIZE(A2, LDA, 2), %xmm13 movhpd 11 * SIZE(A2, LDA, 2), %xmm13 mulpd %xmm2, %xmm9 addpd %xmm14, %xmm6 movlpd 12 * SIZE(A2, LDA, 2), %xmm14 movhpd 13 * SIZE(A2, LDA, 2), %xmm14 mulpd %xmm2, %xmm10 addpd %xmm15, %xmm7 movlpd 14 * SIZE(A2, LDA, 2), %xmm15 movhpd 15 * SIZE(A2, LDA, 2), %xmm15 mulpd %xmm2, %xmm11 addpd %xmm8, %xmm4 mulpd %xmm3, %xmm12 addpd %xmm9, %xmm5 mulpd %xmm3, %xmm13 addpd %xmm10, %xmm6 mulpd %xmm3, %xmm14 addpd %xmm11, %xmm7 mulpd %xmm3, %xmm15 addpd %xmm12, %xmm4 addpd %xmm13, %xmm5 addpd %xmm14, %xmm6 addpd %xmm15, %xmm7 movlpd %xmm4, 8 * SIZE(Y1) movhpd %xmm4, 9 * SIZE(Y1) movlpd %xmm5, 10 * SIZE(Y1) movhpd %xmm5, 11 * SIZE(Y1) movlpd %xmm6, 12 * SIZE(Y1) movhpd %xmm6, 13 * SIZE(Y1) movlpd %xmm7, 14 * SIZE(Y1) movhpd %xmm7, 15 * SIZE(Y1) subq $-16 * SIZE, A1 subq $-16 * SIZE, A2 subq $-16 * SIZE, Y1 ALIGN_3.L14: testq $8, M je .L15 movlpd 0 * SIZE(A1), %xmm8 movhpd 1 * SIZE(A1), %xmm8 movlpd 2 * SIZE(A1), %xmm9 movhpd 3 * SIZE(A1), %xmm9 movlpd 4 * SIZE(A1), %xmm10 movhpd 5 * SIZE(A1), %xmm10 movlpd 6 * SIZE(A1), %xmm11 movhpd 7 * SIZE(A1), %xmm11 movlpd 0 * SIZE(A2), %xmm12 movhpd 1 * SIZE(A2), %xmm12 movlpd 2 * SIZE(A2), %xmm13 movhpd 3 * SIZE(A2), %xmm13 movlpd 4 * SIZE(A2), %xmm14 movhpd 5 * SIZE(A2), %xmm14 movlpd 6 * SIZE(A2), %xmm15 movhpd 7 * SIZE(A2), %xmm15 movlpd 0 * SIZE(Y1), %xmm4 movhpd 1 * SIZE(Y1), %xmm4 movlpd 2 * SIZE(Y1), %xmm5 movhpd 3 * SIZE(Y1), %xmm5 movlpd 4 * SIZE(Y1), %xmm6 movhpd 5 * SIZE(Y1), %xmm6 movlpd 6 * SIZE(Y1), %xmm7 movhpd 7 * SIZE(Y1), %xmm7 mulpd %xmm0, %xmm8 mulpd %xmm0, %xmm9 mulpd %xmm0, %xmm10 mulpd %xmm0, %xmm11 addpd %xmm8, %xmm4 movlpd 0 * SIZE(A1, LDA, 2), %xmm8 movhpd 1 * SIZE(A1, LDA, 2), %xmm8 mulpd %xmm1, %xmm12 addpd %xmm9, %xmm5 movlpd 2 * SIZE(A1, LDA, 2), %xmm9 movhpd 3 * SIZE(A1, LDA, 2), %xmm9 mulpd %xmm1, %xmm13 addpd %xmm10, %xmm6 movlpd 4 * SIZE(A1, LDA, 2), %xmm10 movhpd 5 * SIZE(A1, LDA, 2), %xmm10 mulpd %xmm1, %xmm14 addpd %xmm11, %xmm7 movlpd 6 * SIZE(A1, LDA, 2), %xmm11 movhpd 7 * SIZE(A1, LDA, 2), %xmm11 mulpd %xmm1, %xmm15 addpd %xmm12, %xmm4 movlpd 0 * SIZE(A2, LDA, 2), %xmm12 movhpd 1 * SIZE(A2, LDA, 2), %xmm12 mulpd %xmm2, %xmm8 addpd %xmm13, %xmm5 movlpd 2 * SIZE(A2, LDA, 2), %xmm13 movhpd 3 * SIZE(A2, LDA, 2), %xmm13 mulpd %xmm2, %xmm9 addpd %xmm14, %xmm6 movlpd 4 * SIZE(A2, LDA, 2), %xmm14 movhpd 5 * SIZE(A2, LDA, 2), %xmm14 mulpd %xmm2, %xmm10 addpd %xmm15, %xmm7 movlpd 6 * SIZE(A2, LDA, 2), %xmm15 movhpd 7 * SIZE(A2, LDA, 2), %xmm15 mulpd %xmm2, %xmm11 addpd %xmm8, %xmm4 mulpd %xmm3, %xmm12 addpd %xmm9, %xmm5 mulpd %xmm3, %xmm13 addpd %xmm10, %xmm6 mulpd %xmm3, %xmm14 addpd %xmm11, %xmm7 mulpd %xmm3, %xmm15 addpd %xmm12, %xmm4 mulpd %xmm0, %xmm8 addpd %xmm13, %xmm5 mulpd %xmm0, %xmm9 addpd %xmm14, %xmm6 mulpd %xmm0, %xmm10 addpd %xmm15, %xmm7 mulpd %xmm0, %xmm11 movlpd %xmm4, 0 * SIZE(Y1) movhpd %xmm4, 1 * SIZE(Y1) movlpd %xmm5, 2 * SIZE(Y1) movhpd %xmm5, 3 * SIZE(Y1) movlpd %xmm6, 4 * SIZE(Y1) movhpd %xmm6, 5 * SIZE(Y1) movlpd %xmm7, 6 * SIZE(Y1) movhpd %xmm7, 7 * SIZE(Y1) addq $8 * SIZE, A1 addq $8 * SIZE, A2 addq $8 * SIZE, Y1 ALIGN_3.L15: testq $4, M je .L16 movlpd 0 * SIZE(A1), %xmm8 movhpd 1 * SIZE(A1), %xmm8 movlpd 2 * SIZE(A1), %xmm9 movhpd 3 * SIZE(A1), %xmm9 movlpd 0 * SIZE(A2), %xmm10 movhpd 1 * SIZE(A2), %xmm10 movlpd 2 * SIZE(A2), %xmm11 movhpd 3 * SIZE(A2), %xmm11 movlpd 0 * SIZE(A1, LDA, 2), %xmm12 movhpd 1 * SIZE(A1, LDA, 2), %xmm12 movlpd 2 * SIZE(A1, LDA, 2), %xmm13 movhpd 3 * SIZE(A1, LDA, 2), %xmm13 movlpd 0 * SIZE(A2, LDA, 2), %xmm14 movhpd 1 * SIZE(A2, LDA, 2), %xmm14 movlpd 2 * SIZE(A2, LDA, 2), %xmm15 movhpd 3 * SIZE(A2, LDA, 2), %xmm15 movlpd 0 * SIZE(Y1), %xmm4 movhpd 1 * SIZE(Y1), %xmm4 movlpd 2 * SIZE(Y1), %xmm5 movhpd 3 * SIZE(Y1), %xmm5 mulpd %xmm0, %xmm8 mulpd %xmm0, %xmm9 mulpd %xmm1, %xmm10 mulpd %xmm1, %xmm11 mulpd %xmm2, %xmm12 mulpd %xmm2, %xmm13 mulpd %xmm3, %xmm14 mulpd %xmm3, %xmm15 addpd %xmm8, %xmm4 addpd %xmm9, %xmm5 addpd %xmm10, %xmm4 addpd %xmm11, %xmm5 addpd %xmm12, %xmm4 addpd %xmm13, %xmm5 addpd %xmm14, %xmm4 addpd %xmm15, %xmm5 movlpd %xmm4, 0 * SIZE(Y1) movhpd %xmm4, 1 * SIZE(Y1) movlpd %xmm5, 2 * SIZE(Y1) movhpd %xmm5, 3 * SIZE(Y1) addq $4 * SIZE, A1 addq $4 * SIZE, A2 addq $4 * SIZE, Y1 ALIGN_3.L16: testq $2, M je .L17 movlpd 0 * SIZE(A1), %xmm8 movhpd 1 * SIZE(A1), %xmm8 movlpd 0 * SIZE(A2), %xmm10 movhpd 1 * SIZE(A2), %xmm10 movlpd 0 * SIZE(A1, LDA, 2), %xmm12 movhpd 1 * SIZE(A1, LDA, 2), %xmm12 movlpd 0 * SIZE(A2, LDA, 2), %xmm14 movhpd 1 * SIZE(A2, LDA, 2), %xmm14 movlpd 0 * SIZE(Y1), %xmm4 movhpd 1 * SIZE(Y1), %xmm4 mulpd %xmm0, %xmm8 mulpd %xmm1, %xmm10 mulpd %xmm2, %xmm12 mulpd %xmm3, %xmm14 addpd %xmm8, %xmm4 addpd %xmm10, %xmm4 addpd %xmm12, %xmm4 addpd %xmm14, %xmm4 movlpd %xmm4, 0 * SIZE(Y1) movhpd %xmm4, 1 * SIZE(Y1) addq $2 * SIZE, A1 addq $2 * SIZE, A2 addq $2 * SIZE, Y1 ALIGN_3.L17: testq $1, M je .L19 movlpd 0 * SIZE(Y1), %xmm4 movlpd 0 * SIZE(A1), %xmm8 movlpd 0 * SIZE(A2), %xmm9 movlpd 0 * SIZE(A1, LDA, 2), %xmm10 movlpd 0 * SIZE(A2, LDA, 2), %xmm11 mulsd %xmm0, %xmm8 mulsd %xmm1, %xmm9 mulsd %xmm2, %xmm10 mulsd %xmm3, %xmm11 addsd %xmm8, %xmm4 addsd %xmm9, %xmm4 addsd %xmm10, %xmm4 addsd %xmm11, %xmm4 movlpd %xmm4, 0 * SIZE(Y1) addq $1 * SIZE, A1 addq $1 * SIZE, A2 addq $1 * SIZE, Y1 ALIGN_3.L19: decq J jg .L11 ALIGN_3.L20: testq $2, N je .L30 movq YY, Y1 movq A, A1 leaq (A, LDA, 1), A2 leaq (A, LDA, 2), A movlpd (X), %xmm0 addq INCX, X movlpd (X), %xmm1 addq INCX, X mulsd STACK_ALPHA, %xmm0 mulsd STACK_ALPHA, %xmm1 unpcklpd %xmm0, %xmm0 unpcklpd %xmm1, %xmm1 movq M, I sarq $4, I jle .L25 movlpd 0 * SIZE(A1), %xmm8 movhpd 1 * SIZE(A1), %xmm8 movlpd 2 * SIZE(A1), %xmm9 movhpd 3 * SIZE(A1), %xmm9 movlpd 4 * SIZE(A1), %xmm10 movhpd 5 * SIZE(A1), %xmm10 movlpd 6 * SIZE(A1), %xmm11 movhpd 7 * SIZE(A1), %xmm11 movlpd 0 * SIZE(A2), %xmm12 movhpd 1 * SIZE(A2), %xmm12 movlpd 2 * SIZE(A2), %xmm13 movhpd 3 * SIZE(A2), %xmm13 movlpd 4 * SIZE(A2), %xmm14 movhpd 5 * SIZE(A2), %xmm14 movlpd 6 * SIZE(A2), %xmm15 movhpd 7 * SIZE(A2), %xmm15 movlpd 0 * SIZE(Y1), %xmm4 movhpd 1 * SIZE(Y1), %xmm4 movlpd 2 * SIZE(Y1), %xmm5 movhpd 3 * SIZE(Y1), %xmm5 movlpd 4 * SIZE(Y1), %xmm6 movhpd 5 * SIZE(Y1), %xmm6 movlpd 6 * SIZE(Y1), %xmm7 movhpd 7 * SIZE(Y1), %xmm7 mulpd %xmm0, %xmm8 mulpd %xmm0, %xmm9 mulpd %xmm0, %xmm10 mulpd %xmm0, %xmm11 decq I jle .L22 ALIGN_3.L21: PREFETCH PREFETCHSIZE * SIZE(A1) addpd %xmm8, %xmm4 movlpd 8 * SIZE(A1), %xmm8 movhpd 9 * SIZE(A1), %xmm8 mulpd %xmm1, %xmm12 addpd %xmm9, %xmm5 movlpd 10 * SIZE(A1), %xmm9 movhpd 11 * SIZE(A1), %xmm9 mulpd %xmm1, %xmm13 addpd %xmm10, %xmm6 movlpd 12 * SIZE(A1), %xmm10 movhpd 13 * SIZE(A1), %xmm10 mulpd %xmm1, %xmm14 addpd %xmm11, %xmm7 movlpd 14 * SIZE(A1), %xmm11 movhpd 15 * SIZE(A1), %xmm11 mulpd %xmm1, %xmm15 PREFETCH PREFETCHSIZE * SIZE(Y1) addpd %xmm12, %xmm4 movlpd 8 * SIZE(A2), %xmm12 movhpd 9 * SIZE(A2), %xmm12 mulpd %xmm0, %xmm8 addpd %xmm13, %xmm5 movlpd 10 * SIZE(A2), %xmm13 movhpd 11 * SIZE(A2), %xmm13 mulpd %xmm0, %xmm9 addpd %xmm14, %xmm6 movlpd 12 * SIZE(A2), %xmm14 movhpd 13 * SIZE(A2), %xmm14 mulpd %xmm0, %xmm10 addpd %xmm15, %xmm7 movlpd 14 * SIZE(A2), %xmm15 movhpd 15 * SIZE(A2), %xmm15 mulpd %xmm0, %xmm11 movlpd %xmm4, 0 * SIZE(Y1) movlpd 8 * SIZE(Y1), %xmm4 movhpd %xmm4, 1 * SIZE(Y1) movhpd 9 * SIZE(Y1), %xmm4 movlpd %xmm5, 2 * SIZE(Y1) movlpd 10 * SIZE(Y1), %xmm5 movhpd %xmm5, 3 * SIZE(Y1) movhpd 11 * SIZE(Y1), %xmm5 movlpd %xmm6, 4 * SIZE(Y1) movlpd 12 * SIZE(Y1), %xmm6 movhpd %xmm6, 5 * SIZE(Y1) movhpd 13 * SIZE(Y1), %xmm6 movlpd %xmm7, 6 * SIZE(Y1) movlpd 14 * SIZE(Y1), %xmm7 movhpd %xmm7, 7 * SIZE(Y1) movhpd 15 * SIZE(Y1), %xmm7 PREFETCH PREFETCHSIZE * SIZE(A2) addpd %xmm8, %xmm4 movlpd 16 * SIZE(A1), %xmm8 movhpd 17 * SIZE(A1), %xmm8 mulpd %xmm1, %xmm12 addpd %xmm9, %xmm5 movlpd 18 * SIZE(A1), %xmm9 movhpd 19 * SIZE(A1), %xmm9 mulpd %xmm1, %xmm13 addpd %xmm10, %xmm6 movlpd 20 * SIZE(A1), %xmm10 movhpd 21 * SIZE(A1), %xmm10 mulpd %xmm1, %xmm14 addpd %xmm11, %xmm7 movlpd 22 * SIZE(A1), %xmm11 movhpd 23 * SIZE(A1), %xmm11 mulpd %xmm1, %xmm15 addpd %xmm12, %xmm4 movlpd 16 * SIZE(A2), %xmm12 movhpd 17 * SIZE(A2), %xmm12 mulpd %xmm0, %xmm8 addpd %xmm13, %xmm5 movlpd 18 * SIZE(A2), %xmm13 movhpd 19 * SIZE(A2), %xmm13 mulpd %xmm0, %xmm9 addpd %xmm14, %xmm6 movlpd 20 * SIZE(A2), %xmm14 movhpd 21 * SIZE(A2), %xmm14 mulpd %xmm0, %xmm10 addpd %xmm15, %xmm7 movlpd 22 * SIZE(A2), %xmm15 movhpd 23 * SIZE(A2), %xmm15 mulpd %xmm0, %xmm11 movlpd %xmm4, 8 * SIZE(Y1) movlpd 16 * SIZE(Y1), %xmm4 movhpd %xmm4, 9 * SIZE(Y1) movhpd 17 * SIZE(Y1), %xmm4 movlpd %xmm5, 10 * SIZE(Y1) movlpd 18 * SIZE(Y1), %xmm5 movhpd %xmm5, 11 * SIZE(Y1) movhpd 19 * SIZE(Y1), %xmm5 movlpd %xmm6, 12 * SIZE(Y1) movlpd 20 * SIZE(Y1), %xmm6 movhpd %xmm6, 13 * SIZE(Y1) movhpd 21 * SIZE(Y1), %xmm6 movlpd %xmm7, 14 * SIZE(Y1) movlpd 22 * SIZE(Y1), %xmm7 movhpd %xmm7, 15 * SIZE(Y1) movhpd 23 * SIZE(Y1), %xmm7 subq $-16 * SIZE, A1 subq $-16 * SIZE, A2 subq $-16 * SIZE, Y1 decq I jg .L21 ALIGN_3.L22: addpd %xmm8, %xmm4 movlpd 8 * SIZE(A1), %xmm8 movhpd 9 * SIZE(A1), %xmm8 mulpd %xmm1, %xmm12 addpd %xmm9, %xmm5 movlpd 10 * SIZE(A1), %xmm9 movhpd 11 * SIZE(A1), %xmm9 mulpd %xmm1, %xmm13 addpd %xmm10, %xmm6 movlpd 12 * SIZE(A1), %xmm10 movhpd 13 * SIZE(A1), %xmm10 mulpd %xmm1, %xmm14 addpd %xmm11, %xmm7 movlpd 14 * SIZE(A1), %xmm11 movhpd 15 * SIZE(A1), %xmm11 mulpd %xmm1, %xmm15 addpd %xmm12, %xmm4 movlpd 8 * SIZE(A2), %xmm12 movhpd 9 * SIZE(A2), %xmm12 mulpd %xmm0, %xmm8 addpd %xmm13, %xmm5 movlpd 10 * SIZE(A2), %xmm13 movhpd 11 * SIZE(A2), %xmm13 mulpd %xmm0, %xmm9 addpd %xmm14, %xmm6 movlpd 12 * SIZE(A2), %xmm14 movhpd 13 * SIZE(A2), %xmm14 mulpd %xmm0, %xmm10 addpd %xmm15, %xmm7 movlpd 14 * SIZE(A2), %xmm15 movhpd 15 * SIZE(A2), %xmm15 mulpd %xmm0, %xmm11 movlpd %xmm4, 0 * SIZE(Y1) movlpd 8 * SIZE(Y1), %xmm4 movhpd %xmm4, 1 * SIZE(Y1) movhpd 9 * SIZE(Y1), %xmm4 movlpd %xmm5, 2 * SIZE(Y1) movlpd 10 * SIZE(Y1), %xmm5 movhpd %xmm5, 3 * SIZE(Y1) movhpd 11 * SIZE(Y1), %xmm5 movlpd %xmm6, 4 * SIZE(Y1) movlpd 12 * SIZE(Y1), %xmm6 movhpd %xmm6, 5 * SIZE(Y1) movhpd 13 * SIZE(Y1), %xmm6 movlpd %xmm7, 6 * SIZE(Y1) movlpd 14 * SIZE(Y1), %xmm7 movhpd %xmm7, 7 * SIZE(Y1) movhpd 15 * SIZE(Y1), %xmm7 addpd %xmm8, %xmm4 mulpd %xmm1, %xmm12 addpd %xmm9, %xmm5 mulpd %xmm1, %xmm13 addpd %xmm10, %xmm6 mulpd %xmm1, %xmm14 addpd %xmm11, %xmm7 mulpd %xmm1, %xmm15 addpd %xmm12, %xmm4 addpd %xmm13, %xmm5 addpd %xmm14, %xmm6 addpd %xmm15, %xmm7 movlpd %xmm4, 8 * SIZE(Y1) movhpd %xmm4, 9 * SIZE(Y1) movlpd %xmm5, 10 * SIZE(Y1) movhpd %xmm5, 11 * SIZE(Y1) movlpd %xmm6, 12 * SIZE(Y1) movhpd %xmm6, 13 * SIZE(Y1)
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -