Uppsala Architecture Research Team
Cache Pirate: Measuring the Curse of the Shared Cache
The Cache Pirate is a low-overhead method for accurately measuring application performance (CPI) and off-chip bandwidth (GB/s) as a function of available shared cache capacity. The method is implemented on real hardware, with no modifications to the application or operating system. We accomplish this by co-running a Pirate application that "steals" cache space with the Target application. By adjusting how much space the Pirate steals during the Target´s execution, and using hardware performance counters to record the Target´s performance, we can accurately and efficiently capture performance data for the Target application as a function of its available shared cache. At the same time we use performance counters to monitor the Pirate to ensure that it is successfully stealing the desired amount of cache.
Using Pirate Data to predict multicore scaling for cache-bound applications. Poster
The Cache Pirate work received support from the UPMARC
research center.
-
Cache Pirating: Measuring the Curse of the Shared Cache. In Proc. 40th International Conference on Parallel Processing, pp 165-175, IEEE Computer Society, 2011. (DOI).
The Cache Pirate was chosen as the Best Paper of ICPP 2011.