PDA

View Full Version : Don't expect much more from the GeForce FX drivers


Pages : [1] 2

AnteP
01-29-03, 06:05 PM
"While Nvidia says the drivers are practically ready to ship--they're reported to be stable, compatible, and feature-complete--drivers with optimized performance are months away."

http://gamespot.com/gamespot/features/pc/nvidiageforcefx/02.html

Sazar
01-29-03, 06:14 PM
many products are like this...

its not a question of can they perform off the bat... we have all seen that the nv30 drivers allow it to compete well with the 9700pro BUT then you have the problem of them seemingly being written specifically for improvements in SOME games rather than all... the drivers are immature and better support will be coming... but I would be highly disappointed if nvidia went ahead and released a card which did not offer decent performance for all games and removed artifacting as it was captured during the benchmarks...

I bought my video card almost as soon as it was released and I did not have a single problem... the drivers were good off the bat...

nvidia will like to duplicate this with their spanking new card...

Smokey
01-29-03, 06:22 PM
Originally posted by AnteP
" feature-complete--drivers with optimized performance are months away."



They will came at the launch of the R350 ;)

AnteP
01-29-03, 06:25 PM
Originally posted by Smokey
They will came at the launch of the R350 ;)

Probably, but not right away. Well I know at least I wouldn't call March/April "moths away".

Filburt
01-29-03, 06:40 PM
Originally posted by Smokey
They will came at the launch of the R350 ;)

It's perfectly likely ATi will do the same thing. Nonetheless, even with "performance" drivers...the NV30 isn't going to be able to compete with the R350.

Chalnoth
01-29-03, 06:49 PM
Originally posted by Smokey
Originally posted by AnteP
" feature-complete--drivers with optimized performance are months away."
They will came at the launch of the R350 ;)
The way you quoted that seems like you may not have understood the context. Here's what it meant:
"While Nvidia says the drivers are practically ready to ship--they're reported to be stable, compatible, and feature-complete,"

"drivers with optimized performance are months away."

Usually it's about 6 months before the drivers are close to fully-optimized.

Steppy
01-29-03, 11:22 PM
Originally posted by Chalnoth
They will came at the launch of the R350 ;)
The way you quoted that seems like you may not have understood the context. Here's what it meant:
"While Nvidia says the drivers are practically ready to ship--they're reported to be stable, compatible, and feature-complete,"

"drivers with optimized performance are months away."

Usually it's about 6 months before the drivers are close to fully-optimized. [/QUOTE] Just in time to go against R400 or R350 with optimized drivers of its own...not an enviable position for NV.

Chalnoth
01-29-03, 11:29 PM
Originally posted by Steppy
Just in time to go against R400 or R350 with optimized drivers of its own...not an enviable position for NV.
The R400 won't have optimized drivers when it is released.

Kruno
01-29-03, 11:53 PM
Originally posted by Chalnoth
The R400 won't have optimized drivers when it is released.

Dude, you are right. OMG!!!! :p
Chalnoth stated the obvious. :D
R400 will be a new architecture would it not? If not completely new there will be changes to it's core that will need to be optimised.


On a serious note, the R300 drivers also don't seem to be optimised much for speed.

Chalnoth
01-30-03, 12:00 AM
Originally posted by K.I.L.E.R
On a serious note, the R300 drivers also don't seem to be optimised much for speed.
No, they don't seem to be well-optimized, though at least the stuttering issues seem to be mostly fixed (though some are still reporting them with some driver revs).

AnteP
01-30-03, 06:38 AM
Originally posted by Chalnoth
No, they don't seem to be well-optimized, though at least the stuttering issues seem to be mostly fixed (though some are still reporting them with some driver revs).

However if the R350 is so damn near identical to the R300 as it seems then it WILL have highly optimized and bug free drivers at launch.

spectre
01-30-03, 06:56 AM
Originally posted by AnteP
However if the R350 is so damn near identical to the R300 as it seems then it WILL have highly optimized and bug free drivers at launch.

True

Kruno
01-30-03, 06:59 AM
Originally posted by AnteP
However if the R350 is so damn near identical to the R300 as it seems then it WILL have highly optimized and bug free drivers at launch.

Nothing is bug free. :)
Even some of nVidia's line of cards have the dang Blood 2 vsync issue. Maybe it is fixed now but I doubt it as I was using the 4x.xx drivers at the time I had my GF3.

Steppy
01-30-03, 08:28 AM
Originally posted by Chalnoth
The R400 won't have optimized drivers when it is released. no it won't and my post didn't say that...it said R400, or R350 with optimized drivers, not R400 or R350 both with optimized drivers. Anyway, I don't think NV is in a good position to depend on optimizing drivers like in the past...the 3 month optimization perf increase would still leave them behind in the perf department(in all likelyhood)which is WAY different than the past, and with NV35 in all likelyhood being a tweaked NV30(going by the typical 12 months from the NVx0 core to the NVx5 core that puts a tweaked NV3x core in the fall timeframe if it's on time), reworking it to FAR exceed NV30 would push it back(it likely will need to be quite a bit more than a tweaked NV30 to compare hardwarewise with R400). So I figure NV has the choice of releasing a card that isn't quite up to the competitons level to get it out on time, or pushing THIS card back to get it to R400 levels. I think NV may have a very hard time getting the performance crown back until NV40, unless ATI screws up. I hope so, because this may force NV to really get back to innovation and make the next gen of card REALLY powerful.

AGP64
01-30-03, 10:21 AM
Originally posted by Chalnoth
No, they don't seem to be well-optimized, though at least the stuttering issues seem to be mostly fixed (though some are still reporting them with some driver revs).

Anand seems to think different:

The interesting thing is that the GeForce FX never overtakes the Radeon 9700 Pro, it consistently falls behind by 4 - 6%. With a higher core clock, the only explanation for the GeForce FX falling behind here would be that the Radeon 9700's raw memory bandwidth advantage helps it out. We debunked this theory by increasing the FX's memory bandwidth by 10% (read: overclocked it to 1100MHz memory) and noted no more than a 0.2% increase in performance, so this test was clearly not saturating the memory bus of the GeForce FX. In theory, the GeForce FX should be outperforming the Radeon 9700 Pro here, but in reality it was not at all.

All evidence points to the GeForce FX's drivers holding it back, but this seems rather atypical of NVIDIA, or does it? Historically, NVIDIA has always provided a driver release 3 - 6 months after the launch of a new GPU that has improved performance anywhere from 10 - 20% on average. With the GeForce FX being a relatively new core, we'd expect to see something very similar from them down the road. The difference this time around is that NVIDIA has a much more powerful competitor in ATI, the Radeon 9700 Pro is no sitting duck. With very mature drivers and solid performance across the board, the GeForce FX vs. ATI's Radeon 9700 is almost reminiscent of the Radeon 8500 vs. NVIDIA's GeForce3 Ti 500, but with the roles reversed.

Source: http://www.anandtech.com/video/showdoc.html?i=1779&p=18

YeuEmMaiMai
01-30-03, 10:41 AM
having the roles reversed is a good thing kinda funny how JC states that when running the standard ARB extensions the R300 lays the smack down on the NV30 and it takes the NV30 code path to compete with R300 in D3

Smokey
01-30-03, 11:12 AM
Originally posted by YeuEmMaiMai
having the roles reversed is a good thing kinda funny how JC states that when running the standard ARB extensions the R300 lays the smack down on the NV30 and it takes the NV30 code path to compete with R300 in D3

At the moment, the NV30 is slightly faster on most scenes in Doom than the
R300, but I can still find some scenes where the R300 pulls a little bit
ahead. The issue is complicated because of the different ways the cards can
choose to run the game.
To make it even more complicated, the exact
precision that ATI uses is in between the floating point precisions offered by
Nvidia, so when Nvidia runs fragment programs, they are at a higher precision
than ATI's, which is some justification for the slower speed. Nvidia assures
me that there is a lot of room for improving the fragment program performance
with improved driver compiler technology.

Where does JC state the R300 lays down the smack

:confused: If you are going to say so and so said this or that, its best to at least quote them ;)

Uttar
01-30-03, 11:22 AM
nVidia seems to say: "drivers are practically ready to ship--they're reported to be stable, compatible, and feature-complete"

I think the "drivers with optimized performance are months away" is just the author's guess.

Anyway, that's just my guess :D


Uttar

jbirney
01-30-03, 11:38 AM
Originally posted by Smokey
Where does JC state the R300 lays down the smack

:confused: If you are going to say so and so said this or that, its best to at least quote them ;)

ok then,


The NV30 runs the ARB2 path MUCH slower than the NV30 path.
Half the speed at the moment. This is unfortunate, because when you do an
exact, apples-to-apples comparison using exactly the same API, the R300 looks
twice as fast, but when you use the vendor-specific paths, the NV30 wins.


Now I will let you guys debate if x2 = smack down :)

Smokey
01-30-03, 12:14 PM
Originally posted by jbirney
ok then,



Now I will let you guys debate if x2 = smack down :)
Sorry I left out the first line to my last quote, which btw is in response to the quote you have used!

The reason for this is that ATI does everything at high precision all the
time, while Nvidia internally supports three different precisions with
different performances. To make it even more complicated, the exact
precision that ATI uses is in between the floating point precisions offered by
Nvidia, so when Nvidia runs fragment programs, they are at a higher precision
than ATI's, which is some justification for the slower speed. Nvidia assures
me that there is a lot of room for improving the fragment program performance
with improved driver compiler technology.

YeuEmMaiMai
01-30-03, 12:22 PM
ok let me set you all straight

1. Nvidia has had since last May to work on drivers for this overgrown leaf blower (It is called using a simulator)

2. We know that they have had working silicon for quite some time well since Cbet (spelling?) right?

3. Their issue has been getting the NV30 to outperform the R300 by jacking up the clock speed to compensate for design weaknesses compared to Ati.

a. FSAA Samples NV30 15.6 billion @ 500Mhz vs 16Billion at 325Mhz
b. Triangle throughput NV30 500Hhz 350MTRi/s vs R300 325Mhz @ 325MTri/s
c. Memory bandwidth NV30 16GB/sec raw vs R300 19.6GB/sec raw
d. DDR-II has higher latency than DDR-I

These are serious limitations and luckily for us that there will be newer cards that will address the shortcommings of each card

Uttar
01-30-03, 12:34 PM
Originally posted by YeuEmMaiMai
These are serious limitations and luckily for us that there will be newer cards that will address the shortcommings of each card

The coming year sure sounds VERY exciting.
It sounds like the NV35 will be based around revolutionary performance, and the R400 around amazing technology ideas.

That's just from combining every single rumor in the world, seeing what makes sense, and speculating on the result. But still, it sounds like there are big chances I'm gonna be right.


Uttar

hithere
01-31-03, 08:18 PM
You know, when R200 was posting inferior quality concerning mip-levels in Quake 3, it was, rightfully in my mind, referred to as a "cheat" to offer competitive performance to the Ti500.

Anyone else get the feeling the FX is using a kind of cheat with its inferior-looking anti-aliasing in the hopes of producing performance in the same category as R300?. I can see that less is going on in the FX's implementation; I'm dreading what it would do if it was doing AA of an IQ commeasurate with R300.

Usually, I wait for a price reduction to buy a new videocard. I was kinda hoping the FX or one of its derivatives would be a good Christmas upgrade for me at the end of the year, with my 9500 @9700 holding me off until then.

Looks like I might have to wait a bit longer.

Steppy
01-31-03, 08:26 PM
Originally posted by hithere
You know, when R200 was posting inferior quality concerning mip-levels in Quake 3, it was, rightfully in my mind, referred to as a "cheat" to offer competitive performance to the Ti500.

Anyone else get the feeling the FX is using a kind of cheat with its inferior-looking anti-aliasing in the hopes of producing performance in the same category as R300?. I can see that less is going on in the FX's implementation; I'm dreading what it would do if it was doing AA of an IQ commeasurate with R300.

Usually, I wait for a price reduction to buy a new videocard. I was kinda hoping the FX or one of its derivatives would be a good Christmas upgrade for me at the end of the year, with my 9500 @9700 holding me off until then.

Looks like I might have to wait a bit longer. I still don't think it was a cheat with the 8500, because the performance was still the same boost after they fixed the IQ. Only time will tell if the performance stays the same once they fix the IQ with the FX...if it stays the same I'll call it a bug...if performance drops it is likely still a bug(that yielded benefits)or a POSSIBLE cheat.

StealthHawk
01-31-03, 08:40 PM
not this old debate again :rolleyes:

even if it IS a cheat in either or both cases, it is so damn obvious, that there is no answer other than IQ will be fixed and still have the same performance. in other words, if either case is/was a cheat it was simply a placeholder for later.

the only company stupid enough to try to get away with lowering quality is SiS. but they're low profile anyway.