On Wed, Nov 11, 2020 at 06:22:53PM -0600, Segher Boessenkool wrote:
> I'm running an all-arch comparison with this patch, just to see what it
> does, but [...]
Results: C0 is trunk, C1 with patch:
C0 C1
alpha 6422312 99.971%
arc 3783838 100.000%
arm 10168277 100.000%
arm64 20077721 0
armhf 14886534 100.000%
c6x 2509915 100.000%
csky 0 0
h8300 1229802 100.000%
i386 12040952 0
ia64 18555229 100.000%
m68k 3868729 100.000%
microblaze 5885763 100.000%
mips 9158101 100.000%
mips64 7402870 100.001%
nds32 4833031 100.000%
nios2 3917080 100.000%
openrisc 4571561 100.000%
parisc 7725308 100.000%
parisc64 0 0
powerpc 11004119 100.000%
powerpc64 22618492 100.000%
powerpc64le 19609678 100.000%
riscv32 1639840 100.000%
riscv64 7658668 0
s390 15345481 0
sh 0 0
shnommu 1694176 100.000%
sparc 4744809 100.000%
sparc64 7205254 100.000%
x86_64 19870124 0
xtensa 2658455 100.002%
0 means it did not build... So some targets newly ICE (x86, riscv, z).
It surprisingly only helps alpha a bit, and all other changes are in the
wrong direction (but very slightly).
Segher
