Re: ZFS committed to the FreeBSD base.

From: Ulrich Spoerlein <uspoerlein_at_gmail.com>
Date: Fri, 13 Apr 2007 20:44:25 +0200
Rick C. Petty wrote:
> On Thu, Apr 12, 2007 at 01:51:59PM -0500, Craig Boston wrote:
> > For something this low level my opinion is it's better to stay with
> > compile time options.  After all, in the above example, cmpxchg8 is a
> > single machine instruction.  How much overhead does it add to retrieve a
> > variable from memory and check it, then jump to the correct place?
> > Enough that it outweighs the benefit of using that instruction in the
> > first place?
> 
> [...] 
> The problem is that ZFS would be compiled (by default) to work for many
> platforms, and thus a majority of systems wouldn't get the nice
> optimization.

Disclaimer: I have no clue what cmpxchg8 actually does, but ...

We are talking about optimizing a filesystem by speeding up the
necessary CPU computations. Now, whenever the CPU waits for I/O (which
the ZFS threads will do plenty of times) it has literally thousands of
cycles to burn.

I don't see how this could possibly make ZFS any faster if it does not
avoid I/O operations entirely.

Ulrich Spoerlein
-- 
"The trouble with the dictionary is you have to know how the word is
spelled before you can look it up to see how it is spelled."
-- Will Cuppy
Received on Fri Apr 13 2007 - 16:55:08 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:08 UTC