[Gmp-commit] /var/hg/gmp-6.2: Avoid the x18 register since it is reserved on ...
mercurial at gmplib.org
mercurial at gmplib.org
Sun Nov 29 22:08:00 UTC 2020
details: /var/hg/gmp-6.2/rev/f4ff6ff711ed
changeset: 18133:f4ff6ff711ed
user: Torbjorn Granlund <tg at gmplib.org>
date: Sun Nov 29 23:07:23 2020 +0100
description:
Avoid the x18 register since it is reserved on Darwin.
diffstat:
mpn/arm64/aors_n.asm | 12 ++++++------
mpn/arm64/aorsmul_1.asm | 13 +++++++++----
mpn/arm64/aorsorrlshC_n.asm | 12 ++++++------
mpn/arm64/cnd_aors_n.asm | 12 ++++++------
mpn/arm64/logops_n.asm | 12 ++++++------
mpn/arm64/lshift.asm | 10 +++++-----
mpn/arm64/lshiftc.asm | 10 +++++-----
mpn/arm64/mul_1.asm | 14 +++++++-------
mpn/arm64/rsh1aors_n.asm | 20 ++++++++++----------
mpn/arm64/rshift.asm | 10 +++++-----
mpn/arm64/sqr_diag_addlsh1.asm | 10 +++++-----
11 files changed, 70 insertions(+), 65 deletions(-)
diffs (truncated from 512 to 300 lines):
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/aors_n.asm
--- a/mpn/arm64/aors_n.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/aors_n.asm Sun Nov 29 23:07:23 2020 +0100
@@ -68,7 +68,7 @@
EPILOGUE()
PROLOGUE(func_n)
CLRCY
-L(ent): lsr x18, n, #2
+L(ent): lsr x17, n, #2
tbz n, #0, L(bx0)
L(bx1): ldr x7, [up]
@@ -77,7 +77,7 @@
str x13, [rp],#8
tbnz n, #1, L(b11)
-L(b01): cbz x18, L(ret)
+L(b01): cbz x17, L(ret)
ldp x4, x5, [up,#8]
ldp x8, x9, [vp,#8]
sub up, up, #8
@@ -88,7 +88,7 @@
ldp x10, x11, [vp,#8]
add up, up, #8
add vp, vp, #8
- cbz x18, L(end)
+ cbz x17, L(end)
b L(top)
L(bx0): tbnz n, #1, L(b10)
@@ -101,7 +101,7 @@
L(b10): ldp x6, x7, [up]
ldp x10, x11, [vp]
- cbz x18, L(end)
+ cbz x17, L(end)
ALIGN(16)
L(top): ldp x4, x5, [up,#16]
@@ -114,8 +114,8 @@
ADDSUBC x12, x4, x8
ADDSUBC x13, x5, x9
stp x12, x13, [rp],#16
- sub x18, x18, #1
- cbnz x18, L(top)
+ sub x17, x17, #1
+ cbnz x17, L(top)
L(end): ADDSUBC x12, x6, x10
ADDSUBC x13, x7, x11
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/aorsmul_1.asm
--- a/mpn/arm64/aorsmul_1.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/aorsmul_1.asm Sun Nov 29 23:07:23 2020 +0100
@@ -32,10 +32,15 @@
include(`../config.m4')
-C cycles/limb
-C Cortex-A53 9.3-9.8
-C Cortex-A57 7.0
-C X-Gene 5.0
+C addmul_1 submul_1
+C cycles/limb cycles/limb
+C Cortex-A53 9.3-9.8 9.3-9.8
+C Cortex-A55 9.0-9.5 9.3-9.8
+C Cortex-A57 7 7
+C Cortex-A72
+C Cortex-A73 6 6
+C X-Gene 5 5
+C Apple M1 1.75 1.75
C NOTES
C * It is possible to keep the carry chain alive between the addition blocks
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/aorsorrlshC_n.asm
--- a/mpn/arm64/aorsorrlshC_n.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/aorsorrlshC_n.asm Sun Nov 29 23:07:23 2020 +0100
@@ -65,14 +65,14 @@
ASM_START()
PROLOGUE(func_n)
- lsr x18, n, #2
+ lsr x6, n, #2
tbz n, #0, L(bx0)
L(bx1): ldr x5, [up]
tbnz n, #1, L(b11)
L(b01): ldr x11, [vp]
- cbz x18, L(1)
+ cbz x6, L(1)
ldp x8, x9, [vp,#8]
lsl x13, x11, #LSH
ADDSUB( x15, x13, x5)
@@ -94,7 +94,7 @@
ADDSUB( x17, x13, x5)
str x17, [rp],#8
sub up, up, #8
- cbz x18, L(end)
+ cbz x6, L(end)
b L(top)
L(bx0): tbnz n, #1, L(b10)
@@ -107,7 +107,7 @@
L(b10): CLRRCY( x9)
ldp x10, x11, [vp]
sub up, up, #16
- cbz x18, L(end)
+ cbz x6, L(end)
ALIGN(16)
L(top): ldp x4, x5, [up,#16]
@@ -124,8 +124,8 @@
ADDSUBC(x16, x12, x4)
ADDSUBC(x17, x13, x5)
stp x16, x17, [rp],#16
- sub x18, x18, #1
- cbnz x18, L(top)
+ sub x6, x6, #1
+ cbnz x6, L(top)
L(end): ldp x4, x5, [up,#16]
extr x12, x10, x9, #RSH
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/cnd_aors_n.asm
--- a/mpn/arm64/cnd_aors_n.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/cnd_aors_n.asm Sun Nov 29 23:07:23 2020 +0100
@@ -65,7 +65,7 @@
CLRCY
- lsr x18, n, #2
+ lsr x17, n, #2
tbz n, #0, L(bx0)
L(bx1): ldr x13, [vp]
@@ -75,7 +75,7 @@
str x9, [rp]
tbnz n, #1, L(b11)
-L(b01): cbz x18, L(rt)
+L(b01): cbz x17, L(rt)
ldp x12, x13, [vp,#8]
ldp x10, x11, [up,#8]
sub up, up, #8
@@ -86,7 +86,7 @@
L(b11): ldp x12, x13, [vp,#8]!
ldp x10, x11, [up,#8]!
sub rp, rp, #8
- cbz x18, L(end)
+ cbz x17, L(end)
b L(top)
L(bx0): ldp x12, x13, [vp]
@@ -99,7 +99,7 @@
b L(mid)
L(b10): sub rp, rp, #16
- cbz x18, L(end)
+ cbz x17, L(end)
ALIGN(16)
L(top): bic x6, x12, cnd
@@ -116,8 +116,8 @@
ADDSUBC x9, x11, x7
ldp x10, x11, [up,#32]!
stp x8, x9, [rp,#32]!
- sub x18, x18, #1
- cbnz x18, L(top)
+ sub x17, x17, #1
+ cbnz x17, L(top)
L(end): bic x6, x12, cnd
bic x7, x13, cnd
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/logops_n.asm
--- a/mpn/arm64/logops_n.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/logops_n.asm Sun Nov 29 23:07:23 2020 +0100
@@ -78,7 +78,7 @@
ASM_START()
PROLOGUE(func)
- lsr x18, n, #2
+ lsr x17, n, #2
tbz n, #0, L(bx0)
L(bx1): ldr x7, [up]
@@ -88,7 +88,7 @@
str x15, [rp],#8
tbnz n, #1, L(b11)
-L(b01): cbz x18, L(ret)
+L(b01): cbz x17, L(ret)
ldp x4, x5, [up,#8]
ldp x8, x9, [vp,#8]
sub up, up, #8
@@ -99,7 +99,7 @@
ldp x10, x11, [vp,#8]
add up, up, #8
add vp, vp, #8
- cbz x18, L(end)
+ cbz x17, L(end)
b L(top)
L(bx0): tbnz n, #1, L(b10)
@@ -110,7 +110,7 @@
L(b10): ldp x6, x7, [up]
ldp x10, x11, [vp]
- cbz x18, L(end)
+ cbz x17, L(end)
ALIGN(16)
L(top): ldp x4, x5, [up,#16]
@@ -127,8 +127,8 @@
POSTOP( x12)
POSTOP( x13)
stp x12, x13, [rp],#16
- sub x18, x18, #1
- cbnz x18, L(top)
+ sub x17, x17, #1
+ cbnz x17, L(top)
L(end): LOGOP( x12, x6, x10)
LOGOP( x13, x7, x11)
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/lshift.asm
--- a/mpn/arm64/lshift.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/lshift.asm Sun Nov 29 23:07:23 2020 +0100
@@ -61,7 +61,7 @@
add rp, rp_arg, n, lsl #3
add up, up, n, lsl #3
sub tnc, xzr, cnt
- lsr x18, n, #2
+ lsr x17, n, #2
tbz n, #0, L(bx0)
L(bx1): ldr x4, [up,#-8]
@@ -69,7 +69,7 @@
L(b01): NSHIFT x0, x4, tnc
PSHIFT x2, x4, cnt
- cbnz x18, L(gt1)
+ cbnz x17, L(gt1)
str x2, [rp,#-8]
ret
L(gt1): ldp x4, x5, [up,#-24]
@@ -89,7 +89,7 @@
PSHIFT x13, x5, cnt
NSHIFT x10, x4, tnc
PSHIFT x2, x4, cnt
- cbnz x18, L(gt2)
+ cbnz x17, L(gt2)
orr x10, x10, x13
stp x2, x10, [rp,#-16]
ret
@@ -123,11 +123,11 @@
orr x11, x12, x2
stp x10, x11, [rp,#-32]!
PSHIFT x2, x4, cnt
-L(lo0): sub x18, x18, #1
+L(lo0): sub x17, x17, #1
L(lo3): NSHIFT x10, x6, tnc
PSHIFT x13, x7, cnt
NSHIFT x12, x7, tnc
- cbnz x18, L(top)
+ cbnz x17, L(top)
L(end): orr x10, x10, x13
orr x11, x12, x2
diff -r 63bce6cacb48 -r f4ff6ff711ed mpn/arm64/lshiftc.asm
--- a/mpn/arm64/lshiftc.asm Sun Nov 15 10:25:36 2020 +0100
+++ b/mpn/arm64/lshiftc.asm Sun Nov 29 23:07:23 2020 +0100
@@ -61,7 +61,7 @@
add rp, rp_arg, n, lsl #3
add up, up, n, lsl #3
sub tnc, xzr, cnt
- lsr x18, n, #2
+ lsr x17, n, #2
tbz n, #0, L(bx0)
L(bx1): ldr x4, [up,#-8]
@@ -69,7 +69,7 @@
L(b01): NSHIFT x0, x4, tnc
PSHIFT x2, x4, cnt
- cbnz x18, L(gt1)
+ cbnz x17, L(gt1)
mvn x2, x2
str x2, [rp,#-8]
ret
@@ -90,7 +90,7 @@
PSHIFT x13, x5, cnt
NSHIFT x10, x4, tnc
PSHIFT x2, x4, cnt
- cbnz x18, L(gt2)
+ cbnz x17, L(gt2)
eon x10, x10, x13
mvn x2, x2
stp x2, x10, [rp,#-16]
@@ -125,11 +125,11 @@
eon x11, x12, x2
stp x10, x11, [rp,#-32]!
PSHIFT x2, x4, cnt
More information about the gmp-commit
mailing list