From e1e14bc59f04e11e5e32523cce4d1f7902bc89a4 Mon Sep 17 00:00:00 2001 From: eb Date: Mon, 23 Jun 2008 23:03:43 +0000 Subject: [PATCH] Backed out [8669]. Better fix for removing cruft from opreport git-svn-id: http://gnuradio.org/svn/gnuradio/trunk@8685 221aa14e-8319-0410-a670-987f0aec2ac5 --- .../src/lib/filter/3dnow_float_dotprod_really_simple.S | 4 ++-- gnuradio-core/src/lib/filter/3dnow_float_dotprod_simple.S | 4 ++-- gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow.S | 4 ++-- gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow64.S | 4 ++-- gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext.S | 4 ++-- .../src/lib/filter/ccomplex_dotprod_3dnowext64.S | 4 ++-- gnuradio-core/src/lib/filter/ccomplex_dotprod_sse.S | 4 ++-- gnuradio-core/src/lib/filter/ccomplex_dotprod_sse64.S | 4 ++-- gnuradio-core/src/lib/filter/complex_dotprod_3dnow.S | 4 ++-- gnuradio-core/src/lib/filter/complex_dotprod_3dnow64.S | 4 ++-- gnuradio-core/src/lib/filter/complex_dotprod_3dnowext.S | 4 ++-- gnuradio-core/src/lib/filter/complex_dotprod_3dnowext64.S | 4 ++-- gnuradio-core/src/lib/filter/complex_dotprod_sse.S | 8 ++++---- gnuradio-core/src/lib/filter/complex_dotprod_sse64.S | 8 ++++---- gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow.S | 4 ++-- gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow64.S | 4 ++-- gnuradio-core/src/lib/filter/fcomplex_dotprod_sse.S | 8 ++++---- gnuradio-core/src/lib/filter/fcomplex_dotprod_sse64.S | 8 ++++---- gnuradio-core/src/lib/filter/float_dotprod_3dnow.S | 4 ++-- gnuradio-core/src/lib/filter/float_dotprod_3dnow64.S | 4 ++-- gnuradio-core/src/lib/filter/float_dotprod_sse.S | 8 ++++---- gnuradio-core/src/lib/filter/float_dotprod_sse64.S | 8 ++++---- 22 files changed, 56 insertions(+), 56 deletions(-) diff --git a/gnuradio-core/src/lib/filter/3dnow_float_dotprod_really_simple.S b/gnuradio-core/src/lib/filter/3dnow_float_dotprod_really_simple.S index b5f435b6..6050425a 100644 --- a/gnuradio-core/src/lib/filter/3dnow_float_dotprod_really_simple.S +++ b/gnuradio-core/src/lib/filter/3dnow_float_dotprod_really_simple.S @@ -72,14 +72,14 @@ sse_float_dotprod: shll $1, %ecx # count * 2 .p2align 4 -.3dnow_float_dotprod_really_simple_loop1: +.Loop1: movq (%eax), %mm0 pfmul (%edx), %mm0 pfadd %mm0, %mm4 addl $8, %edx addl $8, %eax decl %ecx - jne .3dnow_float_dotprod_really_simple_loop1 + jne .Loop1 # at this point mm4 contains partial sums diff --git a/gnuradio-core/src/lib/filter/3dnow_float_dotprod_simple.S b/gnuradio-core/src/lib/filter/3dnow_float_dotprod_simple.S index 91370bb1..1097728c 100644 --- a/gnuradio-core/src/lib/filter/3dnow_float_dotprod_simple.S +++ b/gnuradio-core/src/lib/filter/3dnow_float_dotprod_simple.S @@ -72,7 +72,7 @@ sse_float_dotprod: pxor %mm5, %mm5 # mm5 = 0 0 .p2align 4 -.3dnow_float_dotprod_simple_loop1: +.Loop1: movq 0(%eax), %mm0 movq 8(%eax), %mm1 @@ -85,7 +85,7 @@ sse_float_dotprod: addl $16, %edx addl $16, %eax decl %ecx - jne .3dnow_float_dotprod_simple_loop1 + jne .Loop1 # at this point mm4 and mm5 contain partial sums diff --git a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow.S b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow.S index 51a09bad..26f5830f 100644 --- a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow.S +++ b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow.S @@ -91,7 +91,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnow): # .p2align 4 -.ccomplex_dotprod_3dnow_loop1: +.Loop1: # complex prod: C += A * B, w/ temp Z, mmPN=$80000000 # @@ -165,7 +165,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnow): .L1_test: decl %ecx - jge .ccomplex_dotprod_3dnow_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Let's see if original n_2_ccomplex_blocks was odd. diff --git a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow64.S b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow64.S index bc9859b2..feaf1cb0 100644 --- a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow64.S +++ b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnow64.S @@ -91,7 +91,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnow): # .p2align 4 -.ccomplex_dotprod_3dnow64_loop1: +.Loop1: # complex prod: C += A * B, w/ temp Z, mmPN=$80000000 # @@ -165,7 +165,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnow): .L1_test: dec %rax - jge .ccomplex_dotprod_3dnow64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Let's see if original n_2_ccomplex_blocks was odd. diff --git a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext.S b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext.S index ffe4d597..1a09cc3e 100644 --- a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext.S +++ b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext.S @@ -89,7 +89,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnowext): # .p2align 4 -.ccomplex_dotprod_3dnowext_loop1: +.Loop1: # complex prod: C += A * B, w/ temp Z # @@ -144,7 +144,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnowext): .L1_test: decl %ecx - jge .ccomplex_dotprod_3dnowext_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Let's see if original n_2_ccomplex_blocks was odd. diff --git a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext64.S b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext64.S index 7fb4c306..75f5af28 100644 --- a/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext64.S +++ b/gnuradio-core/src/lib/filter/ccomplex_dotprod_3dnowext64.S @@ -89,7 +89,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnowext): # .p2align 4 -.ccomplex_dotprod_3dnowext64_loop1: +.Loop1: # complex prod: C += A * B, w/ temp Z # @@ -144,7 +144,7 @@ GLOB_SYMB(ccomplex_dotprod_3dnowext): .L1_test: dec %rax - jge .ccomplex_dotprod_3dnowext64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Let's see if original n_2_ccomplex_blocks was odd. diff --git a/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse.S b/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse.S index 5107faec..a06ce4d3 100644 --- a/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse.S +++ b/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse.S @@ -85,7 +85,7 @@ GLOB_SYMB(ccomplex_dotprod_sse): # .p2align 4 -.ccomplex_dotprod_sse_loop1: +.Loop1: # complex prod: C += A * B, w/ temp Z & Y (or B), xmmPN=$0x8000000080000000 # @@ -141,7 +141,7 @@ GLOB_SYMB(ccomplex_dotprod_sse): .L1_test: decl %ecx - jge .ccomplex_dotprod_sse_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Let's sse if original n_2_ccomplex_blocks was odd. diff --git a/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse64.S b/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse64.S index b9c3bad0..6da4f2eb 100644 --- a/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse64.S +++ b/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse64.S @@ -85,7 +85,7 @@ GLOB_SYMB(ccomplex_dotprod_sse): # .p2align 4 -.ccomplex_dotprod_sse64_loop1: +.Loop1: # complex prod: C += A * B, w/ temp Z & Y (or B), xmmPN=$0x8000000080000000 # @@ -141,7 +141,7 @@ GLOB_SYMB(ccomplex_dotprod_sse): .L1_test: dec %rax - jge .ccomplex_dotprod_sse64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Let's sse if original n_2_ccomplex_blocks was odd. diff --git a/gnuradio-core/src/lib/filter/complex_dotprod_3dnow.S b/gnuradio-core/src/lib/filter/complex_dotprod_3dnow.S index a2222f26..6e46bf7b 100644 --- a/gnuradio-core/src/lib/filter/complex_dotprod_3dnow.S +++ b/gnuradio-core/src/lib/filter/complex_dotprod_3dnow.S @@ -100,7 +100,7 @@ GLOB_SYMB(complex_dotprod_3dnow): # .p2align 4 -.complex_dotprod_3dnow_loop1: +.Loop1: pfmul 0(%edx), %mm0 pfadd %mm2, %mm6 @@ -145,7 +145,7 @@ GLOB_SYMB(complex_dotprod_3dnow): .L1_test: decl %ecx - jge .complex_dotprod_3dnow_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/complex_dotprod_3dnow64.S b/gnuradio-core/src/lib/filter/complex_dotprod_3dnow64.S index d67448d8..fe978455 100644 --- a/gnuradio-core/src/lib/filter/complex_dotprod_3dnow64.S +++ b/gnuradio-core/src/lib/filter/complex_dotprod_3dnow64.S @@ -100,7 +100,7 @@ GLOB_SYMB(complex_dotprod_3dnow): # .p2align 4 -.complex_dotprod_3dnow64_loop1: +.Loop1: pfmul 0(%rsi), %mm0 pfadd %mm2, %mm6 @@ -145,7 +145,7 @@ GLOB_SYMB(complex_dotprod_3dnow): .L1_test: dec %rax - jge .complex_dotprod_3dnow64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext.S b/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext.S index 03c93170..122fd2b5 100644 --- a/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext.S +++ b/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext.S @@ -93,7 +93,7 @@ GLOB_SYMB(complex_dotprod_3dnowext): # .p2align 4 -.complex_dotprod_3dnowext_loop1: +.Loop1: pfmul 0(%edx), %mm0 pfadd %mm2, %mm6 @@ -125,7 +125,7 @@ GLOB_SYMB(complex_dotprod_3dnowext): .L1_test: decl %ecx - jge .complex_dotprod_3dnowext_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext64.S b/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext64.S index 1f4da3c5..94905497 100644 --- a/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext64.S +++ b/gnuradio-core/src/lib/filter/complex_dotprod_3dnowext64.S @@ -93,7 +93,7 @@ GLOB_SYMB(complex_dotprod_3dnowext): # .p2align 4 -.complex_dotprod_3dnowext64_loop1: +.Loop1: pfmul 0(%rsi), %mm0 pfadd %mm2, %mm6 @@ -125,7 +125,7 @@ GLOB_SYMB(complex_dotprod_3dnowext): .L1_test: dec %rax - jge .complex_dotprod_3dnowext64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/complex_dotprod_sse.S b/gnuradio-core/src/lib/filter/complex_dotprod_sse.S index f7448486..ff5ef5d0 100644 --- a/gnuradio-core/src/lib/filter/complex_dotprod_sse.S +++ b/gnuradio-core/src/lib/filter/complex_dotprod_sse.S @@ -81,7 +81,7 @@ GLOB_SYMB(complex_dotprod_sse): jmp .L1_test .p2align 4 -.complex_dotprod_sse_loop1: +.Loop1: pxor %mm0, %mm0 punpcklwd 0(%eax), %mm0 @@ -95,7 +95,7 @@ GLOB_SYMB(complex_dotprod_sse): addps %xmm0, %xmm4 .L1_test: decl %ecx - jge .complex_dotprod_sse_loop1 + jge .Loop1 # set up for primary loop which is unrolled 4 times @@ -129,7 +129,7 @@ GLOB_SYMB(complex_dotprod_sse): # hence enter loop at top .p2align 4 -.complex_dotprod_sse_loop2: +.Loop2: mulps (%edx), %xmm0 addps %xmm2, %xmm6 @@ -169,7 +169,7 @@ GLOB_SYMB(complex_dotprod_sse): addl $0x40, %edx addl $0x10, %eax decl %ecx - jne .complex_dotprod_sse_loop2 + jne .Loop2 # OK, now we've done with all the multiplies, but # we still need to handle the unaccumulated diff --git a/gnuradio-core/src/lib/filter/complex_dotprod_sse64.S b/gnuradio-core/src/lib/filter/complex_dotprod_sse64.S index 3096da20..2e945654 100644 --- a/gnuradio-core/src/lib/filter/complex_dotprod_sse64.S +++ b/gnuradio-core/src/lib/filter/complex_dotprod_sse64.S @@ -80,7 +80,7 @@ GLOB_SYMB(complex_dotprod_sse): jmp .L1_test .p2align 4 -.complex_dotprod_sse64_loop1: +.Loop1: pxor %mm0, %mm0 punpcklwd 0(%rdi), %mm0 @@ -94,7 +94,7 @@ GLOB_SYMB(complex_dotprod_sse): addps %xmm0, %xmm4 .L1_test: dec %rax - jge .complex_dotprod_sse64_loop1 + jge .Loop1 # set up for primary loop which is unrolled 4 times @@ -127,7 +127,7 @@ GLOB_SYMB(complex_dotprod_sse): # hence enter loop at top .p2align 4 -.complex_dotprod_sse64_loop2: +.Loop2: mulps (%rsi), %xmm0 addps %xmm2, %xmm6 @@ -167,7 +167,7 @@ GLOB_SYMB(complex_dotprod_sse): add $0x40, %rsi add $0x10, %rdi dec %rdx - jne .complex_dotprod_sse64_loop2 + jne .Loop2 # OK, now we've done with all the multiplies, but # we still need to handle the unaccumulated diff --git a/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow.S b/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow.S index 7aa11fe5..8c49909c 100644 --- a/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow.S +++ b/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow.S @@ -92,7 +92,7 @@ GLOB_SYMB(fcomplex_dotprod_3dnow): # .p2align 4 -.fcomplex_dotprod_3dnow_loop1: +.Loop1: pfmul 0(%edx), %mm0 pfadd %mm2, %mm6 @@ -129,7 +129,7 @@ GLOB_SYMB(fcomplex_dotprod_3dnow): .L1_test: decl %ecx - jge .fcomplex_dotprod_3dnow_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow64.S b/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow64.S index c4566153..2924d91a 100644 --- a/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow64.S +++ b/gnuradio-core/src/lib/filter/fcomplex_dotprod_3dnow64.S @@ -91,7 +91,7 @@ GLOB_SYMB(fcomplex_dotprod_3dnow): # .p2align 4 -.fcomplex_dotprod_3dnow64_loop1: +.Loop1: pfmul 0(%rsi), %mm0 pfadd %mm2, %mm6 @@ -128,7 +128,7 @@ GLOB_SYMB(fcomplex_dotprod_3dnow): .L1_test: dec %rax - jge .fcomplex_dotprod_3dnow64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse.S b/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse.S index 98563f90..e390c275 100644 --- a/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse.S +++ b/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse.S @@ -81,7 +81,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): jmp .L1_test .p2align 4 -.fcomplex_dotprod_sse_loop1: +.Loop1: movlps 0(%eax), %xmm0 shufps $0x50, %xmm0, %xmm0 # b01010000 @@ -92,7 +92,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): addps %xmm0, %xmm4 .L1_test: decl %ecx - jge .fcomplex_dotprod_sse_loop1 + jge .Loop1 # set up for primary loop which is unrolled 4 times @@ -120,7 +120,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): # hence enter loop at top .p2align 4 -.fcomplex_dotprod_sse_loop2: +.Loop2: addps %xmm2, %xmm6 movlps 0x10(%eax), %xmm2 @@ -152,7 +152,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): addl $0x40, %edx addl $0x20, %eax decl %ecx - jne .fcomplex_dotprod_sse_loop2 + jne .Loop2 # OK, now we've done with all the multiplies, but # we still need to handle the unaccumulated diff --git a/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse64.S b/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse64.S index b260bffc..9d985a19 100644 --- a/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse64.S +++ b/gnuradio-core/src/lib/filter/fcomplex_dotprod_sse64.S @@ -79,7 +79,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): jmp .L1_test .p2align 4 -.fcomplex_dotprod_sse64_loop1: +.Loop1: movlps 0(%rdi), %xmm0 shufps $0x50, %xmm0, %xmm0 # b01010000 @@ -90,7 +90,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): addps %xmm0, %xmm4 .L1_test: dec %rax - jge .fcomplex_dotprod_sse64_loop1 + jge .Loop1 # set up for primary loop which is unrolled 4 times @@ -117,7 +117,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): # hence enter loop at top .p2align 4 -.fcomplex_dotprod_sse64_loop2: +.Loop2: addps %xmm2, %xmm6 movlps 0x10(%rdi), %xmm2 @@ -149,7 +149,7 @@ GLOB_SYMB(fcomplex_dotprod_sse): add $0x40, %rsi add $0x20, %rdi dec %rdx - jne .fcomplex_dotprod_sse64_loop2 + jne .Loop2 # OK, now we've done with all the multiplies, but # we still need to handle the unaccumulated diff --git a/gnuradio-core/src/lib/filter/float_dotprod_3dnow.S b/gnuradio-core/src/lib/filter/float_dotprod_3dnow.S index a5da9a85..9e2479cb 100644 --- a/gnuradio-core/src/lib/filter/float_dotprod_3dnow.S +++ b/gnuradio-core/src/lib/filter/float_dotprod_3dnow.S @@ -86,7 +86,7 @@ GLOB_SYMB(float_dotprod_3dnow): # .p2align 4 -.float_dotprod_3dnow_loop1: +.Loop1: pfmul 0(%edx), %mm0 pfadd %mm2, %mm6 movq 16(%eax), %mm2 @@ -107,7 +107,7 @@ GLOB_SYMB(float_dotprod_3dnow): addl $32, %eax .L1_test: decl %ecx - jge .float_dotprod_3dnow_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/float_dotprod_3dnow64.S b/gnuradio-core/src/lib/filter/float_dotprod_3dnow64.S index a19e007c..8558f564 100644 --- a/gnuradio-core/src/lib/filter/float_dotprod_3dnow64.S +++ b/gnuradio-core/src/lib/filter/float_dotprod_3dnow64.S @@ -85,7 +85,7 @@ GLOB_SYMB(float_dotprod_3dnow): # .p2align 4 -.float_dotprod_3dnow64_loop1: +.Loop1: pfmul 0(%rdi), %mm0 pfadd %mm2, %mm6 movq 16(%rsi), %mm2 @@ -106,7 +106,7 @@ GLOB_SYMB(float_dotprod_3dnow): add $32, %rsi .L1_test: dec %rax - jge .float_dotprod_3dnow64_loop1 + jge .Loop1 # We've handled the bulk of multiplies up to here. # Now accumulate the final two additions and see if original diff --git a/gnuradio-core/src/lib/filter/float_dotprod_sse.S b/gnuradio-core/src/lib/filter/float_dotprod_sse.S index a8f03b83..31b0f383 100644 --- a/gnuradio-core/src/lib/filter/float_dotprod_sse.S +++ b/gnuradio-core/src/lib/filter/float_dotprod_sse.S @@ -79,7 +79,7 @@ GLOB_SYMB(float_dotprod_sse): jmp .L1_test .p2align 4 -.float_dotprod_sse_loop1: +.Loop1: movaps (%eax), %xmm0 mulps (%edx), %xmm0 addl $0x10, %edx @@ -87,7 +87,7 @@ GLOB_SYMB(float_dotprod_sse): addps %xmm0, %xmm4 .L1_test: decl %ecx - jge .float_dotprod_sse_loop1 + jge .Loop1 # set up for primary loop which is unrolled 4 times @@ -110,7 +110,7 @@ GLOB_SYMB(float_dotprod_sse): # hence enter loop at top .p2align 4 -.float_dotprod_sse_loop2: +.Loop2: mulps (%edx), %xmm0 addps %xmm2, %xmm6 movaps 0x20(%eax), %xmm2 @@ -130,7 +130,7 @@ GLOB_SYMB(float_dotprod_sse): addl $0x40, %edx addl $0x40, %eax decl %ecx - jne .float_dotprod_sse_loop2 + jne .Loop2 # OK, now we've done with all the multiplies, but # we still need to handle the unaccumulated diff --git a/gnuradio-core/src/lib/filter/float_dotprod_sse64.S b/gnuradio-core/src/lib/filter/float_dotprod_sse64.S index 696f2cc0..4c5ad95c 100644 --- a/gnuradio-core/src/lib/filter/float_dotprod_sse64.S +++ b/gnuradio-core/src/lib/filter/float_dotprod_sse64.S @@ -77,7 +77,7 @@ GLOB_SYMB(float_dotprod_sse): jmp .L1_test .p2align 4 -.float_dotprod_sse64_loop1: +.Loop1: movaps (%rsi), %xmm0 mulps (%rdi), %xmm0 add $0x10, %rdi @@ -85,7 +85,7 @@ GLOB_SYMB(float_dotprod_sse): addps %xmm0, %xmm4 .L1_test: dec %rax - jge .float_dotprod_sse64_loop1 + jge .Loop1 # set up for primary loop which is unrolled 4 times @@ -107,7 +107,7 @@ GLOB_SYMB(float_dotprod_sse): # hence enter loop at top .p2align 4 -.float_dotprod_sse64_loop2: +.Loop2: mulps (%rdi), %xmm0 addps %xmm2, %xmm6 movaps 0x20(%rsi), %xmm2 @@ -127,7 +127,7 @@ GLOB_SYMB(float_dotprod_sse): add $0x40, %rdi add $0x40, %rsi dec %rdx - jne .float_dotprod_sse64_loop2 + jne .Loop2 # OK, now we've done with all the multiplies, but # we still need to handle the unaccumulated -- 2.30.2