[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: linpack benchmark

On Sun, Nov 4, 2012 at 7:29 PM, Stan Hoeppner <stan@hardwarefreak.com> wrote:
> That is incorrect.  HPL exercises only the CPU/memory subsystem (mostly
> CPU).  It doesn't test interconnect performance.  The program "solves a
> dense system of linear equations" and that's it.  It is a "show off"
> test which demonstrates maximal FP throughput of the processor.
> http://icl.cs.utk.edu/hpcc/
> --
> Stan

Hi Stam,

Thanks for your time, and the link to the documentation. Now I see
there is 7 programs (tests) and I'm only interested in one, namely,
HPL, or the use of calculus (in this case a linear system, but anyone
would do the job) to estimate the FLOPS of a system.

I see it is such a rocket to kill a fly, but I think I might be able
to understand and adapt the results, as long as I get stable and
consistent results.

I'm finishing here. Any other doubt I'll ask in a more specialized
list, as you pointed out.


Dr Beco
A.I. researcher

--> . <--

"Look again at that dot. That's here. That's home. That's us. On it
everyone you love, everyone you know..." (Carl Sagan, 1934-1996)

Reply to: