zgemv_n.s

Optimized GotoBLAS libraries

	(p15) LDFD	f49  = [AO3], 1 * SIZE
	(p15) FMA	f84 = f8,  f36, f84
	;;
	(p14) LDFD	f50  = [AO4], 1 * SIZE
	(p15) FMA	f85 = f9,  f36, f85
	;;
	(p14) LDFD	f51  = [AO4], 1 * SIZE
	(p14) ADD3	f80 = f9,  f33, f80
	;;
	(p14) LDFD	f52  = [AO4], 1 * SIZE
	(p14) ADD4	f81 = f8,  f33, f81
	;;
	(p14) LDFD	f53  = [AO4], 1 * SIZE
	(p14) ADD3	f82 = f9,  f35, f82
	;;
	(p15) LDFD	f54  = [AO4], 1 * SIZE
	(p14) ADD4	f83 = f8,  f35, f83
	;;
	(p15) LDFD	f55  = [AO4], 1 * SIZE
	(p15) ADD3	f84 = f9,  f37, f84
	(p15) ADD4	f85 = f8,  f37, f85
	;;
	(p14) FMA	f80 = f10, f38, f80
	(p14) FMA	f81 = f11, f38, f81
	(p14) FMA	f82 = f10, f40, f82
	(p14) FMA	f83 = f11, f40, f83
	(p15) FMA	f84 = f10, f42, f84
	(p15) FMA	f85 = f11, f42, f85
	;;
	(p14) ADD3	f80 = f11, f39, f80
	(p14) ADD4	f81 = f10, f39, f81
	(p14) ADD3	f82 = f11, f41, f82
	(p14) ADD4	f83 = f10, f41, f83
	(p15) ADD3	f84 = f11, f43, f84
	(p15) ADD4	f85 = f10, f43, f85
	;;
	(p14) FMA	f80 = f12, f44, f80
	(p14) FMA	f81 = f13, f44, f81
	(p14) FMA	f82 = f12, f46, f82
	(p14) FMA	f83 = f13, f46, f83
	(p15) FMA	f84 = f12, f48, f84
	(p15) FMA	f85 = f13, f48, f85
	;;
	(p14) ADD3	f80 = f13, f45, f80
	(p14) ADD4	f81 = f12, f45, f81
	(p14) ADD3	f82 = f13, f47, f82
	(p14) ADD4	f83 = f12, f47, f83
	(p15) ADD3	f84 = f13, f49, f84
	(p15) ADD4	f85 = f12, f49, f85
	;;
	(p14) FMA	f80 = f14, f50, f80
	(p14) FMA	f81 = f15, f50, f81
	(p14) FMA	f82 = f14, f52, f82
	(p14) FMA	f83 = f15, f52, f83
	(p15) FMA	f84 = f14, f54, f84
	(p15) FMA	f85 = f15, f54, f85
	;;
	(p14) ADD3	f80 = f15, f51, f80
	(p14) ADD4	f81 = f14, f51, f81
	(p14) ADD3	f82 = f15, f53, f82
	(p14) ADD4	f83 = f14, f53, f83
	(p15) ADD3	f84 = f15, f55, f84
	(p15) ADD4	f85 = f14, f55, f85
	;;
	(p14) STFD	[YST1] = f80, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f81, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f82, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f83, 1 * SIZE
	;;
	(p15) STFD	[YST1] = f84, 1 * SIZE
	;;
	(p15) STFD	[YST1] = f85, 1 * SIZE
	;;

.L30:
	{ .mmi
	mov	YLD1 = YY
	adds	YLD2 = 4 * SIZE, YY
	tbit.z	p6, p0 = N, 1
	}
	;;
	{ .mmb
	mov	YST1 = YY
	adds	YST2 = 4 * SIZE, YY
	(p6) br.cond.dpnt .L40
	}
	;;
	LDFD	f32 = [X], SIZE
	mov	AO1 = A
	mov	pr.rot= 0
	;;
	LDFD	f33 = [X], INCXM1
	add	AO2 = A, LDA
	shr	I = M, 2
	;;
	LDFD	f34 = [X], SIZE
	shladd	A   = LDA, 1, A
	tbit.nz	p14, p0 = M, 1
	;;
	LDFD	f35 = [X], INCXM1
	cmp.eq	p6, p0 = 0, I
	;;
	FMPY	f8  = ALPHA_R, f32
	adds	AO9  = 4 * SIZE, AO1
	FMPY	f9  = ALPHA_I, f32
	adds	AO10 = 4 * SIZE, AO2
	FMPY	f10 = ALPHA_R, f34
	mov	ar.ec= 2
	FMPY	f11 = ALPHA_I, f34
	;;
	adds	PREB   = RPREFETCH * SIZE, YLD1
	adds	I = -1, I
	ADD1	f8  = ALPHA_I, f33, f8
	adds	RPRE1  = RPREFETCH * SIZE, AO1
	ADD2	f9  = ALPHA_R, f33, f9
	adds	RPRE2  = (RPREFETCH + 8) * SIZE, AO2
	ADD1	f10 = ALPHA_I, f35, f10
	ADD2	f11 = ALPHA_R, f35, f11
	;;
	{ .mib
	cmp.eq	p16, p0 = r0, r0
	mov	ar.lc = I
	(p6) br.cond.dpnt .L35
	}
	;;
	.align 16

.L32:
	{ .mfi
	(p17) LDFD	f41  = [AO2], 1 * SIZE
	(p17) FMA	f101 = f8,   f33, f101
	(p16) tbit.nz.unc	p12, p13 = I, 0
	}
	{ .mfi
	(p17) LDFD	f45  = [AO10], 1 * SIZE
	(p17) FMA	f113 = f8,   f37, f113
	}
	;;
	{ .mfi
	(p17) LDFD	f42  = [AO2], 1 * SIZE
	(p17) FMA	f104 = f9,   f33, f104
	(p16) adds	I = -1, I
	}
	{ .mfi
	(p17) LDFD	f46  = [AO10], 1 * SIZE
	(p17) FMA	f116 = f9,   f37, f116
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f102, 1 * SIZE
	(p18) STFD	[YST2] = f114, 1 * SIZE
	(p17) FMA	f107 = f8,   f35, f107
	}
	{ .mmf
	(p17) LDFD	f43  = [AO2], 1 * SIZE
	(p17) LDFD	f47  = [AO10], 1 * SIZE
	(p17) FMA	f119 = f8,   f39, f119
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f105, 1 * SIZE
	(p18) STFD	[YST2] = f117, 1 * SIZE
	(p17) FMA	f110 = f9,   f35, f110
	}
	{ .mmf
	(p17) LDFD	f44  = [AO2], 5 * SIZE
	(p17) LDFD	f48  = [AO10], 5 * SIZE
	(p17) FMA	f122 = f9,   f39, f122
	}
	;;
	{ .mfi
	(p12) lfetch.excl.nt2	[PREB],  16 * SIZE
	(p17) ADD3	f101 = f9,   f34, f101
	}
	{ .mfi
	(p17) ADD3	f113 = f9,   f38, f113
	}
	;;
	{ .mfi
	(p16) LDFD	f100 = [YLD1], 1 * SIZE
	(p17) ADD4	f104 = f8,   f34, f104
	}
	{ .mfi
	(p16) LDFD	f112 = [YLD2], 1 * SIZE
	(p17) ADD4	f116 = f8,   f38, f116
	}
	;;
	{ .mfi
	(p16) LDFD	f103 = [YLD1], 1 * SIZE
	(p17) ADD3	f107 = f9,   f36, f107
	}
	{ .mfi
	(p16) LDFD	f115 = [YLD2], 1 * SIZE
	(p17) ADD3	f119 = f9,   f40, f119
	}
	;;
	{ .mfi
	(p12) PREFETCH	[RPRE1], 16 * SIZE
	(p17) ADD4	f110 = f8,   f36, f110
	}
	{ .mfi
	(p17) ADD4	f122 = f8,   f40, f122
	}
	;;
	{ .mfi
	(p16) LDFD	f32  = [AO1], 1 * SIZE
	(p17) FMA	f101 = f10,   f41, f101
	}
	{ .mfi
	(p16) LDFD	f36  = [AO9], 1 * SIZE
	(p17) FMA	f113 = f10,   f45, f113
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f108, 1 * SIZE
	(p18) STFD	[YST2] = f120, 1 * SIZE
	(p17) FMA	f104 = f11,   f41, f104
	}
	{ .mmf
	(p16) LDFD	f33  = [AO1], 1 * SIZE
	(p16) LDFD	f37  = [AO9], 1 * SIZE
	(p17) FMA	f116 = f11,   f45, f116
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f111, 5 * SIZE
	(p18) STFD	[YST2] = f123, 5 * SIZE
	(p17) FMA	f107 = f10,   f43, f107
	}
	{ .mmf
	(p16) LDFD	f34  = [AO1], 1 * SIZE
	(p16) LDFD	f38  = [AO9], 1 * SIZE
	(p17) FMA	f119 = f10,   f47, f119
	}
	;;
	{ .mfi
	(p16) LDFD	f35  = [AO1], 5 * SIZE
	(p17) FMA	f110 = f11,   f43, f110
	}
	{ .mfi
	(p16) LDFD	f39  = [AO9], 5 * SIZE
	(p17) FMA	f122 = f11,   f47, f122
	}
	;;
	{ .mfi
	(p16) LDFD	f106 = [YLD1], 1 * SIZE
	(p17) ADD3	f101 = f11,   f42, f101
	}
	{ .mfi
	(p16) LDFD	f118 = [YLD2], 1 * SIZE
	(p17) ADD3	f113 = f11,   f46, f113
	}
	;;
	{ .mfi
	(p16) LDFD	f109 = [YLD1], 5 * SIZE
	(p17) ADD4	f104 = f10,   f42, f104
	}
	{ .mfi
	(p16) LDFD	f121 = [YLD2], 5 * SIZE
	(p17) ADD4	f116 = f10,   f46, f116
	}
	;;
	{ .mfi
	(p17) ADD3	f107 = f11,   f44, f107
	}
	{ .mfi
	(p17) ADD3	f119 = f11,   f48, f119
	}
	;;
	{ .mfi
	(p13) PREFETCH	[RPRE2], 16 * SIZE
	(p17) ADD4	f110 = f10,   f44, f110
	}
	{ .mfb
	(p17) ADD4	f122 = f10,   f48, f122
	br.ctop.sptk.few .L32
	}
	;;
	.align 16

.L35:
	{ .mmi
	(p18) STFD	[YST1] = f102, 1 * SIZE
	(p18) STFD	[YST2] = f114, 1 * SIZE
	tbit.nz	p15, p0 = M, 0
	}
	{ .mmi
	(p14) LDFD	f32  = [AO1], 1 * SIZE
	(p14) LDFD	f80  = [YLD1], 1 * SIZE
	}
	;;
	{ .mmi
	(p18) STFD	[YST1] = f105, 1 * SIZE
	(p18) STFD	[YST2] = f117, 1 * SIZE
	}
	{ .mmi
	(p14) LDFD	f33  = [AO1], 1 * SIZE
	(p14) LDFD	f81  = [YLD1], 1 * SIZE
	}
	;;
	{ .mmi
	(p18) STFD	[YST1] = f108, 1 * SIZE
	(p18) STFD	[YST2] = f120, 1 * SIZE
	}
	{ .mmi
	(p14) LDFD	f34  = [AO1], 1 * SIZE
	(p14) LDFD	f82  = [YLD1], 1 * SIZE
	}
	;;
	{ .mmi
	(p18) STFD	[YST1] = f111, 5 * SIZE
	(p18) STFD	[YST2] = f123, 5 * SIZE
	}
	{ .mmi
	(p14) LDFD	f35  = [AO1], 1 * SIZE
	(p14) LDFD	f83  = [YLD1], 1 * SIZE
	}
	;;
	(p15) LDFD	f36  = [AO1], 1 * SIZE
	(p15) LDFD	f84  = [YLD1], 1 * SIZE
	;;
	(p15) LDFD	f37  = [AO1], 1 * SIZE
	(p15) LDFD	f85  = [YLD1], 1 * SIZE
	;;
	(p14) LDFD	f38  = [AO2], 1 * SIZE
	(p14) FMA	f80 = f8,  f32, f80
	;;
	(p14) LDFD	f39  = [AO2], 1 * SIZE
	(p14) FMA	f81 = f9,  f32, f81
	;;
	(p14) LDFD	f40  = [AO2], 1 * SIZE
	(p14) FMA	f82 = f8,  f34, f82
	;;
	(p14) LDFD	f41  = [AO2], 1 * SIZE
	(p14) FMA	f83 = f9,  f34, f83
	;;
	(p15) LDFD	f42  = [AO2], 1 * SIZE
	(p15) FMA	f84 = f8,  f36, f84
	;;
	(p15) LDFD	f43  = [AO2], 1 * SIZE
	(p15) FMA	f85 = f9,  f36, f85
	;;
	(p14) ADD3	f80 = f9,  f33, f80
	(p14) ADD4	f81 = f8,  f33, f81
	(p14) ADD3	f82 = f9,  f35, f82
	(p14) ADD4	f83 = f8,  f35, f83
	(p15) ADD3	f84 = f9,  f37, f84
	(p15) ADD4	f85 = f8,  f37, f85
	;;
	(p14) FMA	f80 = f10, f38, f80
	(p14) FMA	f81 = f11, f38, f81
	(p14) FMA	f82 = f10, f40, f82
	(p14) FMA	f83 = f11, f40, f83
	(p15) FMA	f84 = f10, f42, f84
	(p15) FMA	f85 = f11, f42, f85
	;;
	(p14) ADD3	f80 = f11, f39, f80
	(p14) ADD4	f81 = f10, f39, f81
	(p14) ADD3	f82 = f11, f41, f82
	(p14) ADD4	f83 = f10, f41, f83
	(p15) ADD3	f84 = f11, f43, f84
	(p15) ADD4	f85 = f10, f43, f85
	;;
	(p14) STFD	[YST1] = f80, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f81, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f82, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f83, 1 * SIZE
	;;
	(p15) STFD	[YST1] = f84, 1 * SIZE
	;;
	(p15) STFD	[YST1] = f85, 1 * SIZE
	;;

.L40:
	{ .mmi
	mov	YLD1 = YY
	adds	YLD2 = 4 * SIZE, YY
	tbit.z	p6, p0 = N, 0
	}
	{ .mmb
	mov	YST1 = YY
	adds	YST2 = 4 * SIZE, YY
	(p6) br.cond.dpnt .L990
	}
	;;
	LDFD	f32 = [X], SIZE
	mov	AO1 = A
	adds	AO9 = 4 * SIZE, A
	;;
	LDFD	f33 = [X], INCXM1
	add	A   = A, LDA
	mov	pr.rot= 0
	;;
	{ .mfi
	adds	PREB   = RPREFETCH * SIZE, YLD1
	FMPY	f8  = ALPHA_R, f32
	mov	ar.ec= 2
	}
	{ .mfi
	adds	RPRE1  = RPREFETCH * SIZE, AO1
	FMPY	f9  = ALPHA_I, f32
	shr	I = M, 2
	}
	;;
	{ .mmf
	cmp.eq	p6, p0 = 0, I
	cmp.eq	p16, p0 = r0, r0
	ADD1	f8  = ALPHA_I, f33, f8
	}
	{ .mfi
	adds	I = -1, I
	ADD2	f9  = ALPHA_R, f33, f9
	tbit.nz	p14, p0 = M, 1
	}
	;;
	{ .mib
	nop	__LINE__
	mov	ar.lc = I
	(p6) br.cond.dpnt .L45
	}
	;;
	.align 16

.L42:
	{ .mmf
	(p16) LDFD	f100 = [YLD1], 1 * SIZE
	(p16) LDFD	f112 = [YLD2], 1 * SIZE
	(p17) FMA	f101 = f8,   f33, f101
	}
	{ .mmf
	(p16) LDFD	f32  = [AO1], 1 * SIZE
	(p16) LDFD	f44  = [AO9], 1 * SIZE
	(p17) FMA	f113 = f8,   f45, f113
	}
	;;
	{ .mmf
	(p16) LDFD	f103 = [YLD1], 1 * SIZE
	(p16) LDFD	f115 = [YLD2], 1 * SIZE
	(p17) FMA	f104 = f9,   f33, f104
	}
	{ .mmf
	(p16) LDFD	f35  = [AO1], 1 * SIZE
	(p16) LDFD	f47  = [AO9], 1 * SIZE
	(p17) FMA	f116 = f9,   f45, f116
	}
	;;
	{ .mmf
	(p16) LDFD	f106 = [YLD1], 1 * SIZE
	(p16) LDFD	f118 = [YLD2], 1 * SIZE
	(p17) FMA	f107 = f8,   f39, f107
	}
	{ .mmf
	(p16) LDFD	f38  = [AO1], 1 * SIZE
	(p16) LDFD	f50  = [AO9], 1 * SIZE
	(p17) FMA	f119 = f8,   f51, f119
	}
	;;
	{ .mmf
	(p16) LDFD	f109 = [YLD1], 5 * SIZE
	(p16) LDFD	f121 = [YLD2], 5 * SIZE
	(p17) FMA	f110 = f9,   f39, f110
	}
	{ .mmf
	(p16) LDFD	f41  = [AO1], 5 * SIZE
	(p16) LDFD	f53  = [AO9], 5 * SIZE
	(p17) FMA	f122 = f9,   f51, f122
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f102, 1 * SIZE
	(p18) STFD	[YST2] = f114, 1 * SIZE
	(p17) ADD3	f101 = f9,   f36, f101
	}
	{ .mfi
	(p17) ADD3	f113 = f9,   f48, f113
	(p16) tbit.nz.unc	p12, p13 = I, 0
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f105, 1 * SIZE
	(p18) STFD	[YST2] = f117, 1 * SIZE
	(p17) ADD4	f104 = f8,   f36, f104
	}
	{ .mfi
	(p12) PREFETCH	[RPRE1], 16 * SIZE
	(p17) ADD4	f116 = f8,   f48, f116
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f108, 1 * SIZE
	(p18) STFD	[YST2] = f120, 1 * SIZE
	(p17) ADD3	f107 = f9,   f42, f107
	}
	{ .mfi
	(p13) lfetch.excl.nt2	[PREB],  16 * SIZE
	(p17) ADD3	f119 = f9,   f54, f119
	}
	;;
	{ .mmf
	(p18) STFD	[YST1] = f111, 5 * SIZE
	(p18) STFD	[YST2] = f123, 5 * SIZE
	(p17) ADD4	f110 = f8,   f42, f110
	}
	{ .mfb
	(p17) ADD4	f122 = f8,   f54, f122
	br.ctop.sptk.few .L42
	}
	;;
	.align 16

.L45:
	{ .mmi
	(p18) STFD	[YST1] = f102, 1 * SIZE
	(p18) STFD	[YST2] = f114, 1 * SIZE
	tbit.nz	p15, p0 = M, 0
	}
	{ .mmi
	(p14) LDFD	f32  = [AO1], 1 * SIZE
	(p14) LDFD	f80  = [YLD1], 1 * SIZE
	}
	;;
	{ .mmi
	(p18) STFD	[YST1] = f105, 1 * SIZE
	(p18) STFD	[YST2] = f117, 1 * SIZE
	}
	{ .mmi
	(p14) LDFD	f33  = [AO1], 1 * SIZE
	(p14) LDFD	f81  = [YLD1], 1 * SIZE
	}
	;;
	{ .mmi
	(p18) STFD	[YST1] = f108, 1 * SIZE
	(p18) STFD	[YST2] = f120, 1 * SIZE
	}
	{ .mmi
	(p14) LDFD	f34  = [AO1], 1 * SIZE
	(p14) LDFD	f82  = [YLD1], 1 * SIZE
	}
	;;
	{ .mmi
	(p18) STFD	[YST1] = f111, 5 * SIZE
	(p18) STFD	[YST2] = f123, 5 * SIZE
	}
	{ .mmi
	(p14) LDFD	f35  = [AO1], 1 * SIZE
	(p14) LDFD	f83  = [YLD1], 1 * SIZE
	}
	;;
	(p15) LDFD	f36  = [AO1], 1 * SIZE
	(p15) LDFD	f84  = [YLD1], 1 * SIZE
	;;
	(p15) LDFD	f37  = [AO1], 1 * SIZE
	(p15) LDFD	f85  = [YLD1], 1 * SIZE
	;;
	(p14) FMA	f80 = f8,  f32, f80
	(p14) FMA	f81 = f9,  f32, f81
	(p14) FMA	f82 = f8,  f34, f82
	(p14) FMA	f83 = f9,  f34, f83
	(p15) FMA	f84 = f8,  f36, f84
	(p15) FMA	f85 = f9,  f36, f85
	;;
	(p14) ADD3	f80 = f9,  f33, f80
	(p14) ADD4	f81 = f8,  f33, f81
	(p14) ADD3	f82 = f9,  f35, f82
	(p14) ADD4	f83 = f8,  f35, f83
	(p15) ADD3	f84 = f9,  f37, f84
	(p15) ADD4	f85 = f8,  f37, f85
	;;
	(p14) STFD	[YST1] = f80, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f81, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f82, 1 * SIZE
	;;
	(p14) STFD	[YST1] = f83, 1 * SIZE
	;;
	(p15) STFD	[YST1] = f84, 1 * SIZE
	;;
	(p15) STFD	[YST1] = f85, 1 * SIZE
	;;

.L990:
	{ .mmi
	mov	YST1 = Y
	mov	YST2 = Y
	mov	pr.rot= 0
	}
	{ .mib
	mov	YLD1 = YY
	shr	J = M, 2
	(p10) br.cond.dptk .L999
	}
	;;
	{ .mmi
	cmp.eq	p6, p0 = r0, J
	adds	INCY = - SIZE, INCY
	mov	ar.ec = 4
	}
	{ .mmi
	cmp.eq	p16, p0 = r0, r0
	adds	J = -1, J
	tbit.nz	p13, p0 = M, 1
	}
	;;
	{ .mib
	nop	__LINE__
	mov	ar.lc = J
	(p6) br.cond.dpnt .L995
	}
	;;
.L992:
	{ .mfi
	(p19)	STFD	[YST2] = f35, 1 * SIZE
	(p18)	FADD	f34 = f34, f66
	}
	{ .mmi
	(p16)	LDFD	f64 = [YLD1], 1 * SIZE
	(p16)	LDFD	f32 = [YST1], 1 * SIZE
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f39
	(p18)	FADD	f38 = f38, f70
	(p19)	add YST2 = YST2, INCY
	}
	{ .mmi
	(p16)	LDFD	f68 = [YLD1], 1 * SIZE
	(p16)	LDFD	f36 = [YST1], INCY
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f43, 1 * SIZE
	(p18)	FADD	f42 = f42, f74
	}
	{ .mmi
	(p16)	LDFD	f72 = [YLD1], 1 * SIZE
	(p16)	LDFD	f40 = [YST1], 1 * SIZE
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f47
	(p18)	FADD	f50 = f50, f82
	(p19)	add YST2 = YST2, INCY
	}
	{ .mmi
	(p16)	LDFD	f76 = [YLD1], 1 * SIZE
	(p16)	LDFD	f44 = [YST1], INCY
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f51, 1 * SIZE
	(p18)	FADD	f54 = f54, f86
	}
	{ .mmi
	(p16)	LDFD	f80 = [YLD1], 1 * SIZE
	(p16)	LDFD	f48 = [YST1], 1 * SIZE
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f55
	(p18)	FADD	f58 = f58, f90
	(p19)	add YST2 = YST2, INCY
	}
	{ .mmi
	(p16)	LDFD	f84 = [YLD1], 1 * SIZE
	(p16)	LDFD	f52 = [YST1], INCY
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f59, 1 * SIZE
	(p18)	FADD	f46 = f46, f78
	}
	{ .mmi
	(p16)	LDFD	f88 = [YLD1], 1 * SIZE
	(p16)	LDFD	f56 = [YST1], 1 * SIZE
	}
	;;
	{ .mfi
	(p19)	STFD	[YST2] = f63
	(p18)	FADD	f62 = f62, f94
	(p19)	add YST2 = YST2, INCY
	}
	{ .mmb
	(p16)	LDFD	f92 = [YLD1], 1 * SIZE
	(p16)	LDFD	f60 = [YST1], INCY
	br.ctop.sptk.few .L992
	}
	;;

.L995:
	(p13)	LDFD	f32 = [YST1], 1 * SIZE
	(p13)	LDFD	f40 = [YLD1], 1 * SIZE
	tbit.nz	p14, p0 = M, 0
	;;
	(p13)	LDFD	f33 = [YST1], INCY
	(p13)	LDFD	f41 = [YLD1], 1 * SIZE
	;;
	(p13)	LDFD	f34 = [YST1], 1 * SIZE
	(p13)	LDFD	f42 = [YLD1], 1 * SIZE
	;;
	(p13)	LDFD	f35 = [YST1], INCY
	(p13)	LDFD	f43 = [YLD1], 1 * SIZE
	;;
	(p14)	LDFD	f36 = [YST1], 1 * SIZE
	(p14)	LDFD	f44 = [YLD1], 1 * SIZE
	;;
	(p14)	LDFD	f37 = [YST1], INCY
	(p14)	LDFD	f45 = [YLD1], 1 * SIZE
	;;
	(p13)	FADD	f32 = f32, f40
	(p13)	FADD	f33 = f33, f41
	(p13)	FADD	f34 = f34, f42
	(p13)	FADD	f35 = f35, f43
	(p14)	FADD	f36 = f36, f44
	(p14)	FADD	f37 = f37, f45
	;;
	(p13)	STFD	[YST2] = f32, 1 * SIZE
	;;
	(p13)	STFD	[YST2] = f33
	(p13)	add YST2 = YST2, INCY
	;;
	(p13)	STFD	[YST2] = f34, 1 * SIZE
	;;
	(p13)	STFD	[YST2] = f35
	(p13)	add YST2 = YST2, INCY
	;;
	(p14)	STFD	[YST2] = f36, 1 * SIZE
	;;
	(p14)	STFD	[YST2] = f37
	;;

.L999:
	mov	r8 = r0
	adds	r9 = 1 * 16, SP
	;;
	ldf.fill  f16 = [SP], 32
	ldf.fill  f17 = [r9], 32
	mov	 ar.lc = ARLC
	;;	
	ldf.fill  f18 = [SP], 32
	ldf.fill  f19 = [r9], 32
	mov pr    = PR, -1
	;;	
	ldf.fill  f20 = [SP], 32
	ldf.fill  f21 = [r9], 32
	mov	ar.pfs = ARPFS
	;;	
	ldf.fill  f22 = [SP], 32
	ldf.fill  f23 = [r9]
	br.ret.sptk.many b0
	;;
	EPILOGUE