LVC21F-205 Performance Optimization of ceph all-flash Based on aarch64

Session Abstract

Level: Advanced Nowadays with the reduction of the price and the increase of the capability for the flash devices, the all-flash solution in the distributed storage area has attracted a lot of interest. In this presentation, we will talk about the optimization of Ceph all-flash solution on Arm64, we will elaborate some details as below: Avoid cross-NUMA data access CPU partition optimization thread resource scheduling Boost performance with 64K PageSize Software prefetching reduces cache miss. Issues encountered during PMEM enabling

Session Speakers

Wade wu

huawei software engineer (huawei)

Five years of distributed storage development experience

Level: Advanced Nowadays with the reduction of the price and the increase of the capability for the flash devices, the all-flash solution in the distributed storage area has attracted a lot of interest. In this presentation, we will talk about the optimization of Ceph all-flash solution on Arm64, we will elaborate some details as below: Avoid cross-NUMA data access CPU partition optimization thread resource scheduling Boost performance with 64K PageSize Software prefetching reduces cache miss. Issues encountered during PMEM enabling

comments powered by Disqus

Other Posts

Sign up. Receive Updates. Stay informed.

Sign up to our mailing list to receive updates on the latest Linaro Connect news!