CERN Accelerating science

2008

 

  1. At this time the question of how to replace Ian McLaren as sole STT tester and WG coordinator is probably the outstanding item for 2008. The most likely solution looks to be to move the test procedures into the SFT/SPI test mechanism. This will be tackled by Gunter, Anton and the SPI team, with some assistance from Ian. However, this only addresses part of Ian's activities, and the importance of the rest remains to be determined. There are notes and recipes but no one apart from Ian has tried to use them.
  2. Detailed physics and performance measurements on a small number of specially prepared tests will continue to be done by Alberto.
  3. Very large statistic tests of release candidates on the Grid will also continue to be done by Alberto.
  4. It is still planned to set up a dedicated processor, isolated from system volatility to run a small number of relevant benchmarks under controlled conditions.
  5. The speed of modern X86 processors has now made Solaris testing and distributions, whether at CERN or Slac, a real bottleneck. While some testing may still be valuable, systematic testing and releases should be reviewed.
  6. Graphics output testing has been mainly done "by hand" by the Visualization WG. Is a "batch" test of some kind feasible?

Created: 04 February 2008
$Id: 2008.html,v 1.1 2008/02/04 14:16:32 gunter Exp $