ZFS performance bottlenecks: CPU or RAM or anything else?

[Freddie Cash]

On Tue, May 17, 2016 at 2:11 PM, Steven Hartland <steven at multiplay.co.uk>
wrote:

> Raidz is limited essential limited to a single drive performance
> per dev for read and write while mirror is single drive performance for
> write its number of drives for read. Don't forget mirror is not limited to
> two it can be three, four or more; so if you need more read throughput you
> can add drives to the mirror.
>
> To increase raidz performance you need to add more vdevs. While this
> doesn't have to be double i.e. the same vdev config as the first it
> generally a good idea.
>
> Don't forget that while it rebalances write performance of a multi vdev
> raidz will be limited to the added vdev.
>

​Everybody is missing the point of the OP.

They're not asking for ways to improve the performance of a raidz-based
pool; they're asking why they get different performance metrics from the
exact same pool when they change the CPU and RAM.

And, more importantly, why a Core-i3-based system shows better performance
than a Core-i7-based system.​  Is there something inherent to the way ZFS
works that favours one setup over another (lower CPU core counts running at
higher speeds is better/worse than higher CPU core counts running at lower
speeds; more RAM channels is better/worse; things like that).


-- 
Freddie Cash
fjwcash at gmail.com