Date: | Wed, 03 Jan 2007 19:32:14 -0500 |
---|---|
From: | Shougata Ghosh <shougata@xxxxxxxxxxxxx> |
Subject: | [Gems-users] cpu-switch-time |
HiHow does the simulation time vary with cpu-switch-time? I'm running the tm-deque (transactional microbenchmark) program with MESI_SMP_LogTm_directory (Ruby only) and have set cpu-switch-time to 1. It seems to be running extremely slow! Earlier, I had mistakenly run it with cpu-switch-time set to 1000 (default) and it ran much faster. Any ideas on this? I expected cpu-switch-time to affect the correctness of transactional programs but not quite sure why it is slowing down simulation time. Thanks |
[← Prev in Thread] | Current Thread | [Next in Thread→] |
---|---|---|
|
Previous by Date: | Re: [Gems-users] Profiling the physical address for each L2 miss, Mike Marty |
---|---|
Next by Date: | Re: [Gems-users] cpu-switch-time, hongxia sun |
Previous by Thread: | Re: [Gems-users] confused by a comment in the MOESI LogTM code, 郭锐 |
Next by Thread: | Re: [Gems-users] cpu-switch-time, hongxia sun |
Indexes: | [Date] [Thread] |