Issue 1365003003: switch from ebp to ebx.

Issue 1365003003: switch from ebp to ebx. (Closed)

Created:
5 years, 2 months ago by fbarchard1

Modified:
5 years, 2 months ago

Reviewers:
harryjin

Base URL:
https://chromium.googlesource.com/libyuv/libyuv@master

Target Ref:
refs/heads/master

Project:
libyuv

Visibility:
Public.

More Reviews

Description

switch from ebp to ebx. ebx encodes more efficiently (1 byte less) for most address modes, than ebp. previously it was used for 411 format, but the reader uses pinsrw now avoiding gpr register. BUG=libyuv:488 R=harryjin@google.com Committed: https://chromium.googlesource.com/libyuv/libyuv/+/d4594beefc0986f02949fecb5d2424b1e07254b1

Patch Set 1 #

Patch Set 2 : add version #

Created: 5 years, 2 months ago

Download [raw] [tar.bz2]

	Unified diffs	Side-by-side diffs	Delta from patch set	Stats (+102 lines, -102 lines)			Patch
M	README.chromium	View	1	1 chunk	+1 line, -1 line	0 comments	Download
M	include/libyuv/version.h	View	1	1 chunk	+1 line, -1 line	0 comments	Download
M	source/row_win.cc	View		27 chunks	+100 lines, -100 lines	0 comments	Download

Messages

Total messages: 4 (1 generated)

Expand Messages | Collapse Messages | Show Generated Messages | Hide Generated Messages

fbarchard1

Was _NV12ToARGBRow_AVX2: 00000000: 56 push esi 00000001: 55 push ebp 00000002: 8B 44 24 0C ...

5 years, 2 months ago (2015-09-24 23:24:40 UTC) #2

Was
_NV12ToARGBRow_AVX2:
  00000000: 56                 push        esi
  00000001: 55                 push        ebp
  00000002: 8B 44 24 0C        mov         eax,dword ptr [esp+0Ch]
  00000006: 8B 74 24 10        mov         esi,dword ptr [esp+10h]
  0000000A: 8B 54 24 14        mov         edx,dword ptr [esp+14h]
  0000000E: 8B 6C 24 18        mov         ebp,dword ptr [esp+18h]
  00000012: 8B 4C 24 1C        mov         ecx,dword ptr [esp+1Ch]
  00000016: C5 D5 74 ED        vpcmpeqb    ymm5,ymm5,ymm5
  0000001A: C5 FA 6F 06        vmovdqu     xmm0,xmmword ptr [esi]
  0000001E: 8D 76 10           lea         esi,[esi+10h]
  00000021: C4 E3 FD 00 C0 D8  vpermq      ymm0,ymm0,0D8h
  00000027: C5 FD 61 C0        vpunpcklwd  ymm0,ymm0,ymm0
  0000002B: C5 FA 6F 20        vmovdqu     xmm4,xmmword ptr [eax]
  0000002F: C4 E3 FD 00 E4 D8  vpermq      ymm4,ymm4,0D8h
  00000035: C5 DD 60 E4        vpunpcklbw  ymm4,ymm4,ymm4
  00000039: 8D 40 10           lea         eax,[eax+10h]
  0000003C: C4 E2 7D 04 55 40  vpmaddubsw  ymm2,ymm0,ymmword ptr [ebp+40h]
  00000042: C4 E2 7D 04 4D 20  vpmaddubsw  ymm1,ymm0,ymmword ptr [ebp+20h]
  00000048: C4 E2 7D 04 45 00  vpmaddubsw  ymm0,ymm0,ymmword ptr [ebp]
  0000004E: C5 FE 6F 9D A0 00  vmovdqu     ymm3,ymmword ptr [ebp+000000A0h]
            00 00
  00000056: C5 E5 F9 D2        vpsubw      ymm2,ymm3,ymm2
  0000005A: C5 FE 6F 9D 80 00  vmovdqu     ymm3,ymmword ptr [ebp+00000080h]
            00 00
  00000062: C5 E5 F9 C9        vpsubw      ymm1,ymm3,ymm1
  00000066: C5 FE 6F 5D 60     vmovdqu     ymm3,ymmword ptr [ebp+60h]
  0000006B: C5 E5 F9 C0        vpsubw      ymm0,ymm3,ymm0
  0000006F: C5 DD E4 A5 C0 00  vpmulhuw    ymm4,ymm4,ymmword ptr [ebp+000000C0h]
            00 00
  00000077: C5 FD ED C4        vpaddsw     ymm0,ymm0,ymm4
  0000007B: C5 F5 ED CC        vpaddsw     ymm1,ymm1,ymm4
  0000007F: C5 ED ED D4        vpaddsw     ymm2,ymm2,ymm4
  00000083: C5 FD 71 E0 06     vpsraw      ymm0,ymm0,6
  00000088: C5 F5 71 E1 06     vpsraw      ymm1,ymm1,6
  0000008D: C5 ED 71 E2 06     vpsraw      ymm2,ymm2,6
  00000092: C5 FD 67 C0        vpackuswb   ymm0,ymm0,ymm0
  00000096: C5 F5 67 C9        vpackuswb   ymm1,ymm1,ymm1
  0000009A: C5 ED 67 D2        vpackuswb   ymm2,ymm2,ymm2
  0000009E: C5 FD 60 C1        vpunpcklbw  ymm0,ymm0,ymm1
  000000A2: C4 E3 FD 00 C0 D8  vpermq      ymm0,ymm0,0D8h
  000000A8: C5 ED 60 D5        vpunpcklbw  ymm2,ymm2,ymm5
  000000AC: C4 E3 FD 00 D2 D8  vpermq      ymm2,ymm2,0D8h
  000000B2: C5 FD 61 CA        vpunpcklwd  ymm1,ymm0,ymm2
  000000B6: C5 FD 69 C2        vpunpckhwd  ymm0,ymm0,ymm2
  000000BA: C5 FE 7F 0A        vmovdqu     ymmword ptr [edx],ymm1
  000000BE: C5 FE 7F 42 20     vmovdqu     ymmword ptr [edx+20h],ymm0
  000000C3: 8D 52 40           lea         edx,[edx+40h]
  000000C6: 83 E9 10           sub         ecx,10h
  000000C9: 0F 8F 4B FF FF FF  jg          0000001A
  000000CF: 5D                 pop         ebp
  000000D0: 5E                 pop         esi
  000000D1: C5 F8 77           vzeroupper
  000000D4: C3                 ret

Now
_NV12ToARGBRow_AVX2:
  00000000: 56                 push        esi
  00000001: 53                 push        ebx
  00000002: 8B 44 24 0C        mov         eax,dword ptr [esp+0Ch]
  00000006: 8B 74 24 10        mov         esi,dword ptr [esp+10h]
  0000000A: 8B 54 24 14        mov         edx,dword ptr [esp+14h]
  0000000E: 8B 5C 24 18        mov         ebx,dword ptr [esp+18h]
  00000012: 8B 4C 24 1C        mov         ecx,dword ptr [esp+1Ch]
  00000016: C5 D5 74 ED        vpcmpeqb    ymm5,ymm5,ymm5
  0000001A: C5 FA 6F 06        vmovdqu     xmm0,xmmword ptr [esi]
  0000001E: 8D 76 10           lea         esi,[esi+10h]
  00000021: C4 E3 FD 00 C0 D8  vpermq      ymm0,ymm0,0D8h
  00000027: C5 FD 61 C0        vpunpcklwd  ymm0,ymm0,ymm0
  0000002B: C5 FA 6F 20        vmovdqu     xmm4,xmmword ptr [eax]
  0000002F: C4 E3 FD 00 E4 D8  vpermq      ymm4,ymm4,0D8h
  00000035: C5 DD 60 E4        vpunpcklbw  ymm4,ymm4,ymm4
  00000039: 8D 40 10           lea         eax,[eax+10h]
  0000003C: C4 E2 7D 04 53 40  vpmaddubsw  ymm2,ymm0,ymmword ptr [ebx+40h]
  00000042: C4 E2 7D 04 4B 20  vpmaddubsw  ymm1,ymm0,ymmword ptr [ebx+20h]
  00000048: C4 E2 7D 04 03     vpmaddubsw  ymm0,ymm0,ymmword ptr [ebx]
  0000004D: C5 FE 6F 9B A0 00  vmovdqu     ymm3,ymmword ptr [ebx+000000A0h]
            00 00
  00000055: C5 E5 F9 D2        vpsubw      ymm2,ymm3,ymm2
  00000059: C5 FE 6F 9B 80 00  vmovdqu     ymm3,ymmword ptr [ebx+00000080h]
            00 00
  00000061: C5 E5 F9 C9        vpsubw      ymm1,ymm3,ymm1
  00000065: C5 FE 6F 5B 60     vmovdqu     ymm3,ymmword ptr [ebx+60h]
  0000006A: C5 E5 F9 C0        vpsubw      ymm0,ymm3,ymm0
  0000006E: C5 DD E4 A3 C0 00  vpmulhuw    ymm4,ymm4,ymmword ptr [ebx+000000C0h]
            00 00
  00000076: C5 FD ED C4        vpaddsw     ymm0,ymm0,ymm4
  0000007A: C5 F5 ED CC        vpaddsw     ymm1,ymm1,ymm4
  0000007E: C5 ED ED D4        vpaddsw     ymm2,ymm2,ymm4
  00000082: C5 FD 71 E0 06     vpsraw      ymm0,ymm0,6
  00000087: C5 F5 71 E1 06     vpsraw      ymm1,ymm1,6
  0000008C: C5 ED 71 E2 06     vpsraw      ymm2,ymm2,6
  00000091: C5 FD 67 C0        vpackuswb   ymm0,ymm0,ymm0
  00000095: C5 F5 67 C9        vpackuswb   ymm1,ymm1,ymm1
  00000099: C5 ED 67 D2        vpackuswb   ymm2,ymm2,ymm2
  0000009D: C5 FD 60 C1        vpunpcklbw  ymm0,ymm0,ymm1
  000000A1: C4 E3 FD 00 C0 D8  vpermq      ymm0,ymm0,0D8h
  000000A7: C5 ED 60 D5        vpunpcklbw  ymm2,ymm2,ymm5
  000000AB: C4 E3 FD 00 D2 D8  vpermq      ymm2,ymm2,0D8h
  000000B1: C5 FD 61 CA        vpunpcklwd  ymm1,ymm0,ymm2
  000000B5: C5 FD 69 C2        vpunpckhwd  ymm0,ymm0,ymm2
  000000B9: C5 FE 7F 0A        vmovdqu     ymmword ptr [edx],ymm1
  000000BD: C5 FE 7F 42 20     vmovdqu     ymmword ptr [edx+20h],ymm0
  000000C2: 8D 52 40           lea         edx,[edx+40h]
  000000C5: 83 E9 10           sub         ecx,10h
  000000C8: 0F 8F 4C FF FF FF  jg          0000001A
  000000CE: 5B                 pop         ebx
  000000CF: 5E                 pop         esi
  000000D0: C5 F8 77           vzeroupper
  000000D3: C3                 ret

fbarchard1

5 years, 2 months ago (2015-09-25 00:25:16 UTC) #4

Message was sent while issue was closed.

Committed patchset #1 (id:1) manually as
d4594beefc0986f02949fecb5d2424b1e07254b1 (presubmit successful).

Expand Messages | Collapse Messages | Show Generated Messages | Hide Generated Messages