Project

General

Profile

Actions

Bug #2183

open

Time statistics for RCV is incorrectly calculated in local launches

Added by Evgeny Novikov almost 13 years ago. Updated over 11 years ago.

Status:
Open
Priority:
High
Assignee:
Category:
Infrastructure
Start date:
01/11/2012
Due date:
% Done:

0%

Estimated time:
Detected in build:
svn
Platform:
Published in build:

Description

During understanding #2178 issue, I noticed that time for RCV (and thus for DSCV) isn't consistent. For instance:

<rcv verifier="blast">
  <time name="SMT_SOLVER">3380</time>
  <time name="ALL">670</time>
  <time name="PBLAST.OPT">570</time>
  ...
</rcv>
<dscv>
  <status>OK</status>
  <time name="ALL">670</time>
  <desc></desc>
</dscv>

How can the total execution time for RCV and DSCV can be less then their component execution times?
Also you can observe the problem with time calculation at the page with statistics on reference launches: for cluster launches SMT_SOLVER time isn't calculated at all, while for local launches it doesn't summed to a total time of RCV.


Related issues 2 (2 open0 closed)

Related to Linux Driver Verification - Bug #2178: Time statistics for RI is incorrectly calculated in local launches OpenVitaly Mordan01/10/2012

Actions
Related to Linux Driver Verification - Bug #2793: Time statistics for BCE is incorrectly calculated in local launchesNewVitaly Mordan04/27/2012

Actions
Actions #1

Updated by Evgeny Novikov almost 13 years ago

  • Assignee deleted (Vadim Mutilin)
Actions #2

Updated by Alexey Khoroshilov over 11 years ago

We should consider CPU Accounting Controller (https://www.kernel.org/doc/Documentation/cgroups/cpuacct.txt) to solve this issue.

Actions #3

Updated by Evgeny Novikov over 11 years ago

  • Assignee set to Vitaly Mordan

The most important issue among other related with time/memory!

Actions

Also available in: Atom PDF