Gabe Newell on Graphics companies

Re: Off on a slight tangent....

ZoinKs! said:
Well, this isn't what we're looking for, but I found it interesting.

Gary McTaggart said:
5600 runs HL2 at 5-10fps with ps_2_0 in common maps, 5600 runs HL2 at
30-40fps with ps_1_1 in the same map. This is running at 1024x768. The
ps_2_0 shader may be a little bit more complex in this case, but not
much. I could push down the resolution to get a decent framerate, but
why if it runs great as ps_1_1 hardware?
(from directxdev mailing list)

At least it gives us an idea of what to expect with lower end hardware. I shudder to think what those poor souls with fx5200 "dx 9" cards are gonna experience...
No worries for me, though. :D
Nice find, really interesting.
 
Seemed like a perfectly sensible reply from Gabe with my coloured commentry from a source who admits to his bias. The reply is all in past tense when referring to the DX9 issues, and anyone with a grain of common sense knew that the R3xx was better than the nv3x at DX9.

When it comes to R4xx Vs. nv4x I suspect we'll see tiny differences. HL2 will be a bad bench for PS2 because nVidia are going to pour a lot of effort into this game.
 
AlphaWolf said:
PowerK said:
One thing for sure, IMO. Performance difference between NV40, and R420 in Half-Life2 will be A LOT smaller than their performance difference in DOOM3.

With or without AA?
Sorry, I missed your post. It's actually a very good point.
IMO, general AA performance on ATI is better compared to nVIDIA. (thanks to its higher fillrate). *Plus* there is a Temporal AA with new Catalyst. 8) It should be very interesting when final build of DOOM3 ships, and lots of benchmarks become available on the net.
 
I would love to see nVIDIA owning ATI in HL2 while ATI owns nVIDIA in Doom 3. It would be quite surprising if that happened.
 
I'd rather seem them be within 5 fps in both games, so the pour souls who bought the NV cards don't whine.
 
Back
Top