The ABINIT Users Mailing List ( CLOSED )

["Guillaume Dumont" <dumont.guillaume@gmail.com>]

Dear Dr. Bottin,

I succesfully ran your gold case on my machine and was able to get superlinear scaling, e.g. a speedup factor of 40 for 32 cpus. Attached are timmings results for the 1, 4, 6, 8, 16 and 32 processors runs, a graph showing the speedup obtained and a table with the timmings. 

However I was not able to get this kind of scaling with my own case. Here are the main differences between your case and mine:

- You have a lot more bands than I do: 648 vs 80
- Your fft grid is symmetric, mine is not: 108 x 108 x 108 vs 36 x 36 x 512 (my unit cell is not symmetric)
- You have only one k-point: 1 vs 4
- iscf: 7 vs 5

Here are my observations:

- reducing the number of ffts in the z direction did not change the scaling behavior
- reducing the number of k-points (4 -> 1) gave better speedup factors for large number of cpus, 
e.g. pour 32 cpus speedup was 20 with 4 kpts and 27 with one.
- changing iscf to 7 reduced the overall time but did not improve the scaling.
- changing the orientation of the unit cell (and hence de fft grid) did not change the scaling either.

I also noticed that the -O2 flag gives better performance (shorter execution times) than the -O3 flag. All these calculations and those reported in my previous posts were done a 

Altix 3700 with 128 CPUs
CPUs are 1,5 GHz Itanium 2 with a L3 cache of 6 Mb
4 Gb of DDR ram per cpu
communication: NUMAlink 3
we used the intel compiler (version 9.1)  and the mkl library (version 8.1). -O2 flag was used to compiled both.

If you have any suggestions to improve my scaling, let me know.

Regards,

-- 
Guillaume Dumont

=========================
guillaume.dumont.1@umontreal.ca

dumont.guillaume@gmail.com
(514) 341 5298
(514) 343 6111 ext. 13279


This is a message from the MailScanner E-Mail Virus Protection Service
----------------------------------------------------------------------
The original e-mail attachment "bottin_test.zip"
is on the list of unacceptable attachments for this site and has been
replaced by this warning message.

Due to limitations placed on us by the Regulation of Investigatory Powers
Act 2000, we were unable to keep a copy of the original attachment.

At Thu Nov 23 22:00:23 2006 the virus scanner said:
    (bottin_test.zip)

-- 
Postmaster
MAPR
www.mapr.ucl.ac.be

MailScanner thanks transtec Computers for their support