Linux 3.15 Kernel I/O Scheduler Benchmarks

[phoronix]

Phoronix: Linux 3.15 Kernel I/O Scheduler Benchmarks

After writing yesterday about the BFQ I/O scheduler update with its hopeful intentions of landing within the mainline Linux kernel, some readers wrote in about updated I/O scheduler results... Here they are...

http://www.phoronix.com/vr.php?view=MTcwMjk

[AnAkIn]

Unfortunately none of these benchmarks show application start-up time. They were measured in the BFQ test results: http://www.algogroup.unimo.it/people...ed/results.php

[gamerk2]

It comes down to application performance versus application latency. You'd expect BFQ/CFQ to do better in the latency department, but worse on performance, which bears out in the results.

[Rallos Zek]

Unfortunately none of these benchmarks show application start-up time. They were measured in the BFQ test results: http://www.algogroup.unimo.it/people...ed/results.php

Take those BFQ benchmarks with a grain of salt. A picture of Elvis drinking beers with Bigfoot is more believable than those made up graphs masquerading as benchmarks.