[IPOL discuss] about performance comparisons in IPOL articles

[Nicolas Limare]

Dear all,

When you compare software performances in an IPOL article, please be
precise. Saying that "program A is twice as fast as program B" is not
verifiable. I think that, to be meaningful, one must at least provide
this information:
* exact version of the codes
* exact compiler version and compilation options
* harwdare model, at least the CPU (model and speed) and RAM (size and
  speed)
* if and where some parallelism is involved
* operating system type, model and version
* version of important libraries used by the code if relevant
* input data, parameters
* how the performance is measured

If possible, the codes should be submitted with the paper as
supplementary non-reviewed material. When some of the above
information is not available, it should be explained.

All the best,

PS: see also David H. Bailey, "Twelve ways to fool the masses when
    giving performance results on parallel computers," Supercomputing
    Review, Aug 1991, pg. 54-55.
    http://www.davidhbailey.com/dhbpapers/twelve-ways.pdf

-- 
Nicolas LIMARE - CMLA - ENS Cachan         http://limare.perso.math.cnrs.fr/
IPOL journal                                             http://www.ipol.im/
-> image processing, reproducible research, open science
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <https://tools.ipol.im/mailman/archive/discuss/attachments/20130612/747d0fb8/attachment.pgp>