ストレージI/O高速化への取り組み organization ideas Linux BLK-MQ/SCSI-MQ - Interconnect Your Future - InfiniBand and RoCE
Interconnect Your Future - InfiniBand and RoCE A technical blog about high speed interconnect / 本ブログの内容は個人的見解であり 所属組織及び企業の意見を代弁するものではありません / # The contents expressed here are only my personal opinions.
Merged for the Linux 3.13 kernel was the multi-queue block layer allows for better SSD performance with reduced latency and by balancing I/O workload across multiple CPU cores and supporting multiple hardware queues. With the upcoming Linux 3.16 kernel , the "blk-mq" code is expected to be feature complete and deliver great performance.
Building upon the major blk-mq work for the multi-queue block layer, the SCSI multi-queue code is now in good shape according to its developers, is delivering very promising performance results, and should be merged into the Linux 3.17 kernel cycle.
In organization ideas terms of the performance potential of this new code to better take advantage of multiple CPU cores and designed to work with (eventually) multiple hardware queues too, "The usage of blk-mq dramatically decreases CPU usage under all workloads going down from 100% CPU usage that the old setup can hit easily to usually less than 20% for maxing out storage subsystems with 512byte reads and writes, and it allows to easily archive millions of IOPS."
月別アーカイブ
Hatena Blog / Hatena::Diary
Interconnect Your Future - InfiniBand and RoCE A technical blog about high speed interconnect / 本ブログの内容は個人的見解であり 所属組織及び企業の意見を代弁するものではありません / # The contents expressed here are only my personal opinions.
Merged for the Linux 3.13 kernel was the multi-queue block layer allows for better SSD performance with reduced latency and by balancing I/O workload across multiple CPU cores and supporting multiple hardware queues. With the upcoming Linux 3.16 kernel , the "blk-mq" code is expected to be feature complete and deliver great performance.
Building upon the major blk-mq work for the multi-queue block layer, the SCSI multi-queue code is now in good shape according to its developers, is delivering very promising performance results, and should be merged into the Linux 3.17 kernel cycle.
In organization ideas terms of the performance potential of this new code to better take advantage of multiple CPU cores and designed to work with (eventually) multiple hardware queues too, "The usage of blk-mq dramatically decreases CPU usage under all workloads going down from 100% CPU usage that the old setup can hit easily to usually less than 20% for maxing out storage subsystems with 512byte reads and writes, and it allows to easily archive millions of IOPS."
月別アーカイブ
Hatena Blog / Hatena::Diary
No comments:
Post a Comment