Hi, my CPU is an AMD Turion(tm) 64 Mobile Technology ML-34 (cpuinfo at the bottom of this page).

Recently I've some problem with the intel fortran compiler (cannot halts on divisions by zero*). The intel suggests that the problem is due to the SSE2 of my CPU. In particular they state «it may be that [CPU] may claim SSE2 support but not actually implement all [the istructions]»

on ubuntu 9.10 the flags of the cpuinfo are
fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt lm 3dnowext 3dnow
note that SSE2 is included!

how can I check if all the SSE2 istructions are implemented???

this is the post @ intel -> intel post (2 pages)

Thank you!

*for the fortran addicted: is the -fpe0 compiler flag

cpuinfo

processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 36
model name : AMD Turion(tm) 64 Mobile Technology ML-34
stepping : 2
cpu MHz : 800.000
cache size : 1024 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt lm 3dnowext 3dnow up pni lahf_lm
bogomips : 1600.25
clflush size : 64
power management: ts fid vid ttp tm stc

This is what the AMD/Intel legal battle was all about. With Intel compilers the Intel compiler will only compile the nice i686 code on Intel chips, for intel chips. On the AMD CPUs it only wants to do i386 code....

Anyway, Intel lost the battle in court and the latest compilers should work properly, go figure. Now you know why Intel had such a performance advantage in some benchmarks, because they are scumbags. If your CPU supports SSE2 it is probably properly implemented, no worries. Just blame Intel.

I read in the linked thread "When the program reads zero from keyboard (0, 0.0, 0.e0 and 0.d0) the output is still Infinity." Are you saying that on divide by Zero the output is infinite? That is the correct answer - division by zero is legal though in general there is an error because it cannot be represented as a number, I suppose zero divided by zero may equal one? but anything else causes an infinite number of zeros.

Not sure how people mention ia32 vs AMD64 is an issue - SIMD instructions are a separate function of the CPU.

I have the similar problem, I have mpeg decoder and use SSE2 instruction set, run my decoder and decode video ok on intel platform, but failed on AMD platform(AMD eng sample). when I used tool CPU-z found that the AMD platform support MMX and SSE(1,2,3,4.1,4.2),why failed, I donot know? could give me help?