[sldev] [VWR] Vectorization on i386?

[Callum Lerwick]

On Mon, 2007-12-03 at 10:40 +0000, Tofu Linden wrote:
> Linux?
> 
> I got this hackily working on i386 Linux but I decided not to spend more
> time on it for these reasons:
> 
> * Only the builds from gcc 3.4 were deployable in the wild due to
> over-zealous processor-specific optimizations in later GCC versions
> (i.e. non-SSE2 CPUs would get SSE2 code elsewhere), making this fragile.

My testing so far has not turned up such a problem. It is and has been
compiling in the SSE stuff for as long as it has been in there,
actually. I just don't understand whats stopping it from being enabled.

I've figured out I can go into the Debug Settings editor and force it
on. Profiling and disassembly with oprofile proves there's SSE code
there, and its being run.

As is, I seem to be getting about 3% better performance according to the
internal perf test. My goal here is to see about bringing that up...

> * When it worked, it benchmarked at between ~4% slower and ~4%
> faster than the pure-C++ versions on a Core2Duo and Celeron-D, so
> it didn't seem worth pushing for on i386.  (Hardware skinners totally
> nullify any gain anyway, so even on machines where this was faster
> in SSE2 than pure-C++, a reasonable GPU can be significantly faster
> still).

The open source r300 drivers don't support this yet.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.secondlife.com/pipermail/sldev/attachments/20071203/01caf88d/attachment.pgp