On Tue, 2007-10-16 at 11:58 +1000, David Gibson wrote:
> On Mon, Oct 15, 2007 at 08:36:13AM -0500, Adam Litke wrote:
> > On Mon, 2007-10-15 at 10:53 +1000, David Gibson wrote:
> [snip]
> > > > +#if defined (__powerpc__) && !defined (__LP64__)
> > > > +static int arch_prepare_segment(struct seg_info *htlb_seg_info)
> > > 
> > > I don't see that there's any point doing this on a per-arch basis.
> > > Although it's only really necessary for powerpc, I suggest we do it
> > > this way always just for simplicity.
> > 
> > We thought of this, but x86 takes a pretty painful performance hit.
> > Steve may have details on just how bad it is.
> 
> Ok, fair enough.  Any idea why there's such a big hit?

Actually, it looks like my initial testing that I mentioned to Adam was
messed up. Here's what I'm getting with my head screwed on straight:

Running stream.c on an x86_64 and an x86 results in a 0.01% improvement
on both.

Running SpecCPU06's omnetpp on x86 results in a 9.6% improvement.

The latter is obviously completely unexpected, so I'm going to double
check that again, as well as try other benchmarks. 

-- 

Steve Fox
IBM Linux Technology Center

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Libhugetlbfs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libhugetlbfs-devel

Reply via email to