Replies to post #41762 on Advanced Micro Devices Inc (AMD)

[DDB]

upc - maybe they used the executable created on the Xeon to do benchmark runs on the A64. This is one possible reason why they got half the score which others got using the same compiler settings on a slower A64. An EM64T optimized, branch intensive code could be very inefficient for the K8.

AT didn't include an option to select the target architecture. I think in this case the compiler uses the one which it is running at.

I'll wait for further reviews to get a better picture.

Just another thing - the main part of the CPU test code of ubench is using transcedentals, logarithms, divisions, power functions, a square root, a comparison and an IF statement. You can see it below. It's really too much dependend on a small set of CPU instructions to give a good base for comparisons. Even SiSoft Sandra is better than this ;)

x=i-j;
y=log(1.0+x);
x=abs(sqrt(y/(2.0+x))*(x*cos(atan(y/(3.0+x)))+y*sin(atan(y/(4.0+x)))));
if ( x > 10.0 )
y=pow(1.0+j/(5.0+x),y/(6.0+x));
else
y=pow(1.0+y/(1+j),x);
x=x*exp(1.0/(1.0+y));
k+=x;