Status
| We have completed all of the implementation and completed running our simulations
(thanks to Nate for letting us use the ACAL simpool). Our final presentation
had initial numbers from the first batch of runs. However, we found a bug in
the cache pinning implementation in the simulator which was preventing range
pinning from working (thus our L2 numbers were messed up). After fixing the
bug and changing the management of pinned entries to be LRU within the partition
we ran new numbers including several new benchmarks (netperf-no-touch, surge, and
SPECweb99 surge). We have analyzed this new data in our final paper which can be
found in the resources section.
|
|
|