Re: CPU results on Freescale i.MX

["Jonathan A. Kollasch" <jakllsch%kollasch.net@localhost>]

On Mon, Jan 24, 2011 at 06:43:52AM -0500, Christopher Berardi wrote:
> I wrote a simple program to 'stress-test' a cpu. Basically it is a program 
> that
> computes a couple of sqrt roots a billion or so times, which I have used as
> a (crude?) benchtest. 

Maybe use whetstone?  Or one of the various other benchmark packages
that give a good overall number.

> The results are...
> 
> Genesi Efika MX (Smarttop) with Freescale i.MX 515 (ARMv7 processor rev5 
> (v7l))
> running Ubuntu 10.10
>     real 37m15.173s
>     user 36m37.130s
>     sys   0m09.610s
> 
> 
> ...and for comparison...
> 
> Lenovo T61 with Intel T9300 (Intel Core2 Duo @ 2.5GHz)
> running Ubuntu 10.04
>     real 15m43.094s
>     user 15m41.411s
>     sys   0m00.056s
>
> Homebuilt with AMD (AMD Athalon dual-core @ 2.5GHz)
> running NetBSD 5.1
>     real 182.95s (3m02.95s)
>     user 182.95s (3m02.95s)
>     sys    0.00s

This is probably NetBSD/amd64, and is thus using the SSE unit
instead of the x87 unit.  I get similarly slow results when
I'm on a x86_32 port compared to a x86_64 box of similar GHz.

Also, a Debian GNU/Linux/amd64 box fared competitively to a
NetBSD/amd64 box for me on nearly identical hardware.

Also, does the i.MX 515 have a FPU, and is Ubuntu using it?
Did you compile with optimization there?  That makes a huge
difference on RISC boxes.

        Jonathan Kollasch