ati 9500

[kmolazz]

quote from [H] on comdex :

" There is also news that there are significant performance increases ( up to 25% ) on the new Radeon 9500 with new optimized drivers. Sweet. "



that much of an increase? any thoughts?

 

[KimB]

Of course. Current benchmarks show the 9700 really having troubles with slower CPUs.

That means poor driver optimization. We had better see performance gains like that before long on the 9x00 series!

 

[lwells]

Well, I would be interested to know how much of the boost is drivers and how much comes from having finalised hardware, with the right amount of ram running at the right speed, etc, etc, rather than the crippled cards that ATi sent out for p/review.

On a side note: I guess this would also tie in with conspiracy theories that ATi will 'do an nvidia' and release a new set of killer drivers when the NV30 is *actually* released

LW.

 

[Joe DeFuria]

Of course. Current benchmarks show the 9700 really having troubles with slower CPUs.

I have not seen that...link?

I have seen one or two benchmarks in CPU limited situations (where driver quality tends to come into play) where the 9700 is at, or only slightly behind the so called "gold standard" GeForce drivers....

The interesting thing to see is if the new drivers increase CPU limited scores, or fillrate / bandwithd limited scores.

 

[Althornin]

Of course. Current benchmarks show the 9700 really having troubles with slower CPUs.

That means poor driver optimization. We had better see performance gains like that before long on the 9x00 series!

 

[MuFu]

quote from [H] on comdex :

" There is also news that there are significant performance increases ( up to 25% ) on the new Radeon 9500 with new optimized drivers. Sweet. "



that much of an increase? any thoughts?

Doesn't that make sense? I'd imagine it's taken a while to optimise for the halved rendering arrary - preliminary benchmarks were probably done with drivers closely based on those for the 8x1 cards. It is understandable that optimisation for the cut-down 9500 architecture might yield significant performance gains.

If they are referring to the 9500 Pro as well, then maybe the optimisation extends to making best use of the halved memory bus also.

OpenGL guy?

MuFu.

 

[jb]

That means poor driver optimization. We had better see performance gains like that before long on the 9x00 series!

Yea but with any frist gen of new hardware (DX9 FP pipes, ect) I would not expect the very first few sets of drivers to be very optimized.

 

[BoddoZerg]

ATi and nVidia seem to be pulling a complete role reversal from GeForce3 vs. Radeon8500 to Radeon9700 vs. GeForceFX. Wouldn't be surprised if ATi pulls up a new Catalyst so that Radeon9700Pro can slightly edge out an immature-drivers GeForceFX.

If GeForce3's driver maturation process is any indication for GFFX, the FX may not be operating at full speed until 6 months worth of driver optimizations have passed.

By that time it may be time for R400...

 

[Dreadnought]

Perhaps the delays on the GFFX may help the driver team. Even with crippled/underclocked hardware this long leadin to retail may give them extended time to tweak the drivers.

 

[gkar1]

Of course. Current benchmarks show the 9700 really having troubles with slower CPUs.

That means poor driver optimization. We had better see performance gains like that before long on the 9x00 series!

What do you expect from him. Give him a break.
Wait wait now my turn
Sry, I just had to pick that bone.

Timelines have been reversed it seems with regards to each company now. Now the other vendor is the one gearing up to rain on the other's parade by releasing performance increasing drivers. A testament of how fast things can change and turn around in this industry. I don't intend to blink anytime soon, do you?