GMP 4.3.1 misdetects intel X7460
dmharvey at cims.nyu.edu
Sun Jul 26 13:04:40 CEST 2009
On Jul 26, 2009, at 12:12 PM, Torbjorn Granlund wrote:
> Any idea of what type of processor that is? There are, to my
> three different "core" microarchitecture variants.
> Gen 1, 65nm "Core 2" with Conroe, Merom, etc, with slow integer
> Gen 2, 45nm, "Core 2" with Penryn, with faster integer division
> Gen 3, 45 nm, "Core iN", Nehalem with Bloomfield,Lynnfield with many
> pipeline differences (better pipelining, sometimes longer latencies)
> Do you know to which of these "model 29" belongs?
No, I have no idea. This is sage.math.washington.edu. All I know is
from /proc/cpuinfo. It's a 24-core machine. I could ask William Stein
for more info if that helps, he probably knows more about it.
> If not, please run and time this program, it will tell the latter two
> .globl main
> mov $2666666667, %ecx
> 1: mul
> dec %ecx
> jnz 1b
I changed the "mul" to "mul %rbx", and got 6.94s.
More information about the gmp-bugs