http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50505
Bug #: 50505
Summary: gcc-4.6.1 doesn't use "-mtune=k8-sse3" when using
"-march=native"
Classification: Unclassified
Product: gcc
Version: 4.6.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassig...@gcc.gnu.org
ReportedBy: pa...@condmat1.ciencias.uniovi.es
This is not a regression over gcc-4.4, but I don't understand why it doesn't
tune for k8-sse3 even if my processor has sse3 support and it's detected for
-march:
\_ /usr/libexec/gcc/x86_64-pc-linux-gnu/4.5.3/cc1 -quiet - -D_FORTIFY_SOURCE=2
-march=k8-sse3 -msahf --param l1-cache-size=64 --param l1-cache-line-size=64
--param l2-cache-size=512 -mtune=k8 -quiet -dumpbase - -auxbase-strip /dev/null
-o /tmp/ccm0bSsZ.s
If I don't misremember, this is a Venice:
http://en.wikipedia.org/wiki/Athlon_64#Venice_.2890.C2.A0nm_SOI.29
$ cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 47
model name : AMD Athlon(tm) 64 Processor 3200+
stepping : 0
cpu MHz : 2042.838
cache size : 512 KB
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt lm 3dnowext
3dnow rep_good nopl pni lahf_lm
bogomips : 4085.67
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc
Downstream closed bug as INVALID because:
"Probably because you don't "tune" for an instruction set. It doesn't make
sense. -mtune controls things like scheduling behavior, prefetching, branch
costs, etc. These don't change when you tack on SSE3 instructions. -march is
what controls which instructions sets are used. As far as tuning is concerned
k8 and k8-sse3 are the exact same thing, because gcc has one cost model (struct
processor_costs k8_cost in gcc/config/i386.c) and machine model (m_K8 in same)
it uses for all K8 family chips."
-> https://bugs.gentoo.org/show_bug.cgi?id=384189#c5
But, if that is true, gcc man page should be fixed to drop "k8-sse3" from
-mtune options
Thanks a lot for clarifying this :-)
