Vancouver, BC, Canada
August 27 & 28 - Co-Located Events, Tutorials, Labs & Lightning Talks
August 29-31 - Conference
Click Here For Information & Registration
Wednesday, August 29 • 11:10am - 11:50am
Introducing Cache Pseudo-Locking to Reduce Memory Access Latency - Reinette Chatre, Intel

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Feedback form is now closed.
Reading from cache is significantly faster than reading from memory. Latency sensitive applications thus greatly benefit from having critical data and/or instructions in the cache.
Cache Pseudo-Locking (building on top of Intel's Cache Allocation Technology (CAT)) is a new feature being enabled in the Linux kernel that pseudo-locks a region of memory into a reserved portion of cache that only serves cache hits. The Cache Pseudo-Locked memory is made accessible to user space where an application can map it into its virtual address space and thus have a region of memory with reduced average read latency.

In this presentation, Reinette Chatre will introduce Cache Pseudo-Locking, demonstrate its benefits, and discuss the interfaces she has been working while enabling this feature in the Linux kernel.


Reinette Chatre

Software Engineer, Intel
Reinette Chatre is a Software Engineer who has worked for Intel's Open Source Technology Center for more than 10 years. Currently she is focused on enabling Intel's Cache Pseudo-Locking in the Linux kernel. Previous Linux kernel contributions include participation in the initial enabling... Read More →

Wednesday August 29, 2018 11:10am - 11:50am PDT
Room 109