-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 I adjusted the algorithm from the Weighted Round Robin Queue and resolved the SSD performance issue. Since it is different, I've renamed it so that it doesn't cause confusion later.
My tests are all showing a performance improvement of 3-17%. The enqueue and dequeue latencies are the same or just slightly better than Prioritized Queue, but the distribution of OPs is much more fair especially in the contention situations (more enqueue than dequeue). The new queue is always right on the expected distribution in all cases, even with highly skewed ops (higher priority ops have low costs/size and low priority ops have high costs/size). I could probably get a little more performance out of it by using intrusive containers, but since I'm getting the same latency with better overall performance gains, I wanted to get this in and tested. I can create another PR later if intrusive containers provide more performance. The PR is at https://github.com/ceph/ceph/pull/6964 I've closed the previous RP (https://github.com/ceph/ceph/pull/6781) as this one supersedes it. Any feedback is appreciated. Thanks, - ---------------- Robert LeBlanc PGP Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2 FA62 B9F1 -----BEGIN PGP SIGNATURE----- Version: Mailvelope v1.3.2 Comment: https://www.mailvelope.com wsFcBAEBCAAQBQJWdDMuCRDmVDuy+mK58QAAet0P/iRkGILB96Ryy6HPYZny ODzzn1ld/LkwK/JmVsWti48tOZbnG91WhxQIAQMeha5GeRAcv/Ie2OBZe7Y4 Jo4tCaXbPGwZ9sk1kxjwTm4SYHKwCYf0vuyOVIT2+On0BrXUPfDFPFBHpUwY nDUkyZTi6YCo9W7qgST7AQJHI9wAeb0SlClAsBN3vd8JOoZfuWuA6+hNP6q5 tI6RueXXy8LgbiIiNmm5MvAJQg/7I+rWs2JZyyyu0BaBSm3GPY15LPpIWa7F zg0+ax8pb5J0Ug4fSwzG7iKKNhO/TSC32bC2kRWPThQCmVQ6mUmrBWW311vK ZpFRAPy6mXCVKysFqZFxav5BFagkQZO470Vjej4riYHxRa4QdGOkjZODAb6+ 2GN+wtQH76dVsxm4mOGi4sUFJ5QLjk+nTDILIS5uh3x7nJ94UbqAetzgQA7m bAVQocMb5B0JZb6vHjg4TkwQ2pVgzBWxYQX7Lum3hm3DDVhm4BPbU0juCZaO o2XD6KgM40mxJmsjmyL/siCUV8wqZDamoHMAyljIr0hfkvBR+AB9a6B2+1iE DIIBX/blM7lSQRHIOrBj/FRBeplGbGP2cBXS0v4+N3l5jGFgk4Aisf1KPr22 qaGbxv6pteMJhbpeBf4bOTeU8spwHmpAxmLINYECUk6ySoYJNva9T+IYsBiQ EoPN =bQwT -----END PGP SIGNATURE----- -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
