[BACK]Return to decode_i586_.s CVS log [TXT][DIR] Up to [cvs.NetBSD.org] / pkgsrc / audio / mpg123 / files

File: [cvs.NetBSD.org] / pkgsrc / audio / mpg123 / files / decode_i586_.s (download)

Revision 1.2, Tue May 31 22:13:24 2005 UTC (14 years, 8 months ago) by rillig
Branch: MAIN
CVS Tags: pkgsrc-2019Q4-base, pkgsrc-2019Q4, pkgsrc-2019Q3-base, pkgsrc-2019Q3, pkgsrc-2019Q2-base, pkgsrc-2019Q2, pkgsrc-2019Q1-base, pkgsrc-2019Q1, pkgsrc-2018Q4-base, pkgsrc-2018Q4, pkgsrc-2018Q3-base, pkgsrc-2018Q3, pkgsrc-2018Q2-base, pkgsrc-2018Q2, pkgsrc-2018Q1-base, pkgsrc-2018Q1, pkgsrc-2017Q4-base, pkgsrc-2017Q4, pkgsrc-2017Q3-base, pkgsrc-2017Q3, pkgsrc-2017Q2-base, pkgsrc-2017Q2, pkgsrc-2017Q1-base, pkgsrc-2017Q1, pkgsrc-2016Q4-base, pkgsrc-2016Q4, pkgsrc-2016Q3-base, pkgsrc-2016Q3, pkgsrc-2016Q2-base, pkgsrc-2016Q2, pkgsrc-2016Q1-base, pkgsrc-2016Q1, pkgsrc-2015Q4-base, pkgsrc-2015Q4, pkgsrc-2015Q3-base, pkgsrc-2015Q3, pkgsrc-2015Q2-base, pkgsrc-2015Q2, pkgsrc-2015Q1-base, pkgsrc-2015Q1, pkgsrc-2014Q4-base, pkgsrc-2014Q4, pkgsrc-2014Q3-base, pkgsrc-2014Q3, pkgsrc-2014Q2-base, pkgsrc-2014Q2, pkgsrc-2014Q1-base, pkgsrc-2014Q1, pkgsrc-2013Q4-base, pkgsrc-2013Q4, pkgsrc-2013Q3-base, pkgsrc-2013Q3, pkgsrc-2013Q2-base, pkgsrc-2013Q2, pkgsrc-2013Q1-base, pkgsrc-2013Q1, pkgsrc-2012Q4-base, pkgsrc-2012Q4, pkgsrc-2012Q3-base, pkgsrc-2012Q3, pkgsrc-2012Q2-base, pkgsrc-2012Q2, pkgsrc-2012Q1-base, pkgsrc-2012Q1, pkgsrc-2011Q4-base, pkgsrc-2011Q4, pkgsrc-2011Q3-base, pkgsrc-2011Q3, pkgsrc-2011Q2-base, pkgsrc-2011Q2, pkgsrc-2011Q1-base, pkgsrc-2011Q1, pkgsrc-2010Q4-base, pkgsrc-2010Q4, pkgsrc-2010Q3-base, pkgsrc-2010Q3, pkgsrc-2010Q2-base, pkgsrc-2010Q2, pkgsrc-2010Q1-base, pkgsrc-2010Q1, pkgsrc-2009Q4-base, pkgsrc-2009Q4, pkgsrc-2009Q3-base, pkgsrc-2009Q3, pkgsrc-2009Q2-base, pkgsrc-2009Q2, pkgsrc-2009Q1-base, pkgsrc-2009Q1, pkgsrc-2008Q4-base, pkgsrc-2008Q4, pkgsrc-2008Q3-base, pkgsrc-2008Q3, pkgsrc-2008Q2-base, pkgsrc-2008Q2, pkgsrc-2008Q1-base, pkgsrc-2008Q1, pkgsrc-2007Q4-base, pkgsrc-2007Q4, pkgsrc-2007Q3-base, pkgsrc-2007Q3, pkgsrc-2007Q2-base, pkgsrc-2007Q2, pkgsrc-2007Q1-base, pkgsrc-2007Q1, pkgsrc-2006Q4-base, pkgsrc-2006Q4, pkgsrc-2006Q3-base, pkgsrc-2006Q3, pkgsrc-2006Q2-base, pkgsrc-2006Q2, pkgsrc-2006Q1-base, pkgsrc-2006Q1, pkgsrc-2005Q4-base, pkgsrc-2005Q4, pkgsrc-2005Q3-base, pkgsrc-2005Q3, pkgsrc-2005Q2-base, pkgsrc-2005Q2, cwrapper, cube-native-xorg-base, cube-native-xorg, HEAD
Changes since 1.1: +1 -2 lines

Removed trailing empty line.

/ 
/ synth_1to1 works the same way as the c version of this
/ file.  only two types of changes have been made:
/ - reordered floating point instructions to
/   prevent pipline stalls
/ - made WRITE_SAMPLE use integer instead of
/   (slower) floating point
/ all kinds of x86 processors should benefit from these
/ modifications.
/
/ useful sources of information on optimizing x86 code include:
/
/     Intel Architecture Optimization Manual
/     http://www.intel.com/design/pentium/manuals/242816.htm
/
/     Cyrix 6x86 Instruction Set Summary
/     ftp://ftp.cyrix.com/6x86/6x-dbch6.pdf
/
/     AMD-K5 Processor Software Development
/     http://www.amd.com/products/cpg/techdocs/appnotes/20007e.pdf
/
/ Stefan Bieschewski <stb@acm.org>
/
/ $Id: decode_i586_.s,v 1.2 2005/05/31 22:13:24 rillig Exp $
/
.bss
	.comm	buffs,4352
.data
	.align 2
bo:
	.long 1
.text
	.align 3
.LC0:
	.long 0x0,0x40dfffc0
	.align 3
.LC1:
	.long 0x0,0xc0e00000
.text
	.align 3
.globl _synth_1to1_pent
_synth_1to1_pent:
	subl $12,%esp
	pushl %ebp
	pushl %edi
	pushl %esi
	pushl %ebx
	movl 32(%esp),%eax
	movl 40(%esp),%esi
	xorl %edi,%edi
	movl bo,%ebp
	cmpl %edi,36(%esp)
	jne .L48
	decl %ebp
	andl $15,%ebp
	movl %ebp,bo
	movl $buffs,%ecx
	jmp .L49
.L48:
	addl $2,%esi
	movl $buffs+2176,%ecx
.L49:
	testl $1,%ebp
	je .L50
	movl %ecx,%ebx
	movl %ebp,16(%esp)
	pushl %eax
	movl 20(%esp),%edx
	leal (%ebx,%edx,4),%eax
	pushl %eax
	movl 24(%esp),%eax
	incl %eax
	andl $15,%eax
	leal 1088(,%eax,4),%eax
	addl %ebx,%eax
	jmp .L74
.L50:
	leal 1088(%ecx),%ebx
	leal 1(%ebp),%edx
	movl %edx,16(%esp)
	pushl %eax
	leal 1092(%ecx,%ebp,4),%eax
	pushl %eax
	leal (%ecx,%ebp,4),%eax
.L74:
	pushl %eax
	call _dct64
	addl $12,%esp
	movl 16(%esp),%edx
	leal 0(,%edx,4),%edx
	movl $_decwin+64,%eax
	movl %eax,%ecx
	subl %edx,%ecx
	movl $16,%ebp
.L55:
	flds (%ecx)
	fmuls (%ebx)
	flds 4(%ecx)
	fmuls 4(%ebx)
	fxch %st(1)
	flds 8(%ecx)
	fmuls 8(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 12(%ecx)
	fmuls 12(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 16(%ecx)
	fmuls 16(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 20(%ecx)
	fmuls 20(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 24(%ecx)
	fmuls 24(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 28(%ecx)
	fmuls 28(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 32(%ecx)
	fmuls 32(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 36(%ecx)
	fmuls 36(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 40(%ecx)
	fmuls 40(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 44(%ecx)
	fmuls 44(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 48(%ecx)
	fmuls 48(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 52(%ecx)
	fmuls 52(%ebx)
	fxch %st(2)         
	faddp %st,%st(1)
	flds 56(%ecx)
	fmuls 56(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds 60(%ecx)
	fmuls 60(%ebx)
	fxch %st(2)
	subl $4,%esp
	faddp %st,%st(1)
	fxch %st(1)
	fsubrp %st,%st(1)
	fistpl (%esp)
	popl %eax
	cmpl $32767,%eax
	jg 1f
	cmpl $-32768,%eax
	jl 2f
	movw %ax,(%esi)
	jmp 4f
1:	movw $32767,(%esi)
	jmp 3f
2:	movw $-32768,(%esi)
3:	incl %edi
4:
.L54:
	addl $64,%ebx
	subl $-128,%ecx
	addl $4,%esi
	decl %ebp
	jnz .L55
	flds (%ecx)
	fmuls (%ebx)
	flds 8(%ecx)
	fmuls 8(%ebx)
	flds 16(%ecx)
	fmuls 16(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 24(%ecx)
	fmuls 24(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 32(%ecx)
	fmuls 32(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 40(%ecx)
	fmuls 40(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 48(%ecx)
	fmuls 48(%ebx)
	fxch %st(2)
	faddp %st,%st(1)
	flds 56(%ecx)
	fmuls 56(%ebx)
	fxch %st(2)
	subl $4,%esp
	faddp %st,%st(1)
	fxch %st(1)
	faddp %st,%st(1)
	fistpl (%esp)
	popl %eax
	cmpl $32767,%eax
	jg 1f
	cmpl $-32768,%eax
	jl 2f
	movw %ax,(%esi)
	jmp 4f
1:	movw $32767,(%esi)
	jmp 3f
2:	movw $-32768,(%esi)
3:	incl %edi
4:
.L62:
	addl $-64,%ebx
	addl $4,%esi
	movl 16(%esp),%edx
	leal -128(%ecx,%edx,8),%ecx
	movl $15,%ebp
.L68:
	flds -4(%ecx)
	fchs
	fmuls (%ebx)
	flds -8(%ecx)
	fmuls 4(%ebx)
	fxch %st(1)
	flds -12(%ecx)
	fmuls 8(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -16(%ecx)
	fmuls 12(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -20(%ecx)
	fmuls 16(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -24(%ecx)
	fmuls 20(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -28(%ecx)
	fmuls 24(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -32(%ecx)
	fmuls 28(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -36(%ecx)
	fmuls 32(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -40(%ecx)
	fmuls 36(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -44(%ecx)
	fmuls 40(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -48(%ecx)
	fmuls 44(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -52(%ecx)
	fmuls 48(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -56(%ecx)
	fmuls 52(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds -60(%ecx)
	fmuls 56(%ebx)
	fxch %st(2)
	fsubrp %st,%st(1)
	flds (%ecx)
	fmuls 60(%ebx)
	fxch %st(2)
	subl $4,%esp
	fsubrp %st,%st(1)
	fxch %st(1)
	fsubrp %st,%st(1)
	fistpl (%esp)
	popl %eax
	cmpl $32767,%eax
	jg 1f
	cmpl $-32768,%eax
	jl 2f
	movw %ax,(%esi)
	jmp 4f
1:	movw $32767,(%esi)
	jmp 3f
2:	movw $-32768,(%esi)
3:	incl %edi
4:
.L67:
	addl $-64,%ebx
	addl $-128,%ecx
	addl $4,%esi
	decl %ebp
	jnz .L68
	movl %edi,%eax
	popl %ebx
	popl %esi
	popl %edi
	popl %ebp
	addl $12,%esp
	ret