Re: Benchmarking ROOT 2.21/07 + PGCC (2)

Rene Brun (Rene.Brun@cern.ch)
Fri, 12 Mar 1999 10:59:49 +0000

Hi Jacek,
I am pleased to see that you reacted!!
The idea of the ROOTMARK is (in principle) to always give the
same number for all ROOT benchmarks on the same machine.
I have normalized the ROOTMARK to be 27 ROOTMARKS on my old HP
for all benchmarks.
In the case of "stress", as I mentionned, you can run it in compiled
and interpreted mode. The two runs give 27 ROOTMARKS on the old HP.
You should get the same ROOTMARK on your machine.
However, you will see some discrepancies between these numbers
depending your configuration (memory,disk,cpu) because the ratio
between compiled code is not necessary the same with your compiler
and the HP CC compiler.

It is my intention to start collecting the results of these
benchmarks as soon as we have released 2.21 sometime next week.
I will post the results on the web.

Rene Brun

Jacek M. Holeczek wrote:
>
> Nobody commented my results, but I'm still curious :
> ----------------------------------- compiled "stress"
> stress : Total I/O = 425.6 Mbytes, I = 296.4, O = 129.3
> stress : Compr I/O = 407.4 Mbytes, I = 286.3, O = 121.0
> stress : Real Time = 98.70 seconds Cpu Time = 83.92 seconds
> * ROOTMARKS = 160.4 * Root 2.21/07 990308/1327
> ----------------------------------- interpreted "stress"
> stress : Total I/O = 425.6 Mbytes, I = 296.4, O = 129.3
> stress : Compr I/O = 407.4 Mbytes, I = 286.3, O = 121.0
> stress : Real Time = 185.76 seconds Cpu Time = 167.58 seconds
> * ROOTMARKS = 152.3 * Root 2.21/07 990308/1327
> -----------------------------------
> Note : both the Real Time and the Cpu Time of the interpreted "stress" are
> about twice of the compiled "stress". But ROOTMARKS are the same ?
> What's going on ?
> Do you also observe this ( Rene, which result did you give in your
> benchmark - compiled or interpreted ? Could you, please provide me with
> BOTH results from your PII/400 ? ) ?
> Is there somewhere a "table" which compares different machines ( I mean
> something like it was done for root 0.8 ) ?
> Thanks in advance,
> Jacek.