PDA

View Full Version : Anand's Half-Life 2 article not good news for GeForce FX


Pages : [1] 2 3 4 5 6 7 8 9 10 11

Martrox
09-11-03, 11:57 PM
http://www.anandtech.com/video/showdoc.html?i=1863

Hank Lloyd
09-12-03, 12:00 AM
It sure looks like a NV3x is but a GF4 with some functionality bolted on. Of course, I'm not being serious, but you wouldn't know it looking at these pathetic numbers.

I'm really impressed with the 9800's ability to run > 60 FPS at 12x10. Hats off to the engineers at ATI.

GlowStick
09-12-03, 12:00 AM
I just read it and laughed.

Valve had very strict requirements about the test systems they let us use; the systems were only allowed to use publicly available drivers and thus we used NVIDIA's Detonator 45.23s and ATI's Catalyst 3.7s, both publicly available from the respective websites.

The Dell PCs we used were configured with Pentium 4 3.0C processors on 875P based motherboards with 1GB of memory. We were running Windows XP without any special modifications to the OS or other changes to the system.

So this allowed us to lean nothing more than what Valve released.

Seems like they are hella scared of useing the det 50's.

Hank Lloyd
09-12-03, 12:05 AM
Umm. Yeah, they're scared alright. They're scared that nVidia is going to pull the same shi* that they've been so good at doing over the course of 2003.

It's called taking a stand and preserving integrity, something that nVidia has clearly lacked for the past 12 months.

PreservedSwine
09-12-03, 12:05 AM
Originally posted by GlowStick
I just read it and laughed.



So this allowed us to lean nothing more than what Valve released.

Seems like they are hella scared of useing the det 50's.

Well, since they don't render fog, it kinds presents a problem, dontcha think?:retard:

GlowStick
09-12-03, 12:07 AM
Originally posted by Hank Lloyd
Umm. Yeah, they're scared alright. They're scared that nVidia is going to pull the same shi* that they've been so good at doing over the course of 2003.

It's called taking a stand and preserving integrity, something that nVidia has clearly lacked for the past 12 months. Defiantly not, the preliminar benches of the 50's show decreaed preformance where there were known cheats.

Its called, they are afraid of the results and are forceing their results or no one can play it.

While the Doom3 benches were allowed whatever hardware and drivers they wanted.

GlowStick
09-12-03, 12:09 AM
Originally posted by PreservedSwine
Well, since they don't render fog, it kinds presents a problem, dontcha think?:retard: Hm, on one map, can you identify this map that was used for the demo?

Joe DeFuria
09-12-03, 12:14 AM
Originally posted by GlowStick
Defiantly not, the preliminar benches of the 50's show decreaed preformance where there were known cheats.

Its called, they are afraid of the results and are forceing their results or no one can play it.

While the Doom3 benches were allowed whatever hardware and drivers they wanted.

You're kidding me, right?

Doom3 benchmarking computers were shipped under lock and key, under nVidia's authority.

The only one afraid here is nvidia. All they have to do is publicallyrelease their "best driver ever det 50s", and then they can be used for the benchmarks.

Why don't they release them to the public? What is NVIDIA afraid we'll find?

GlowStick
09-12-03, 12:18 AM
Originally posted by Joe DeFuria
You're kidding me, right?

Doom3 benchmarking computers were shipped under lock and key, under nVidia's authority.

The only one afraid here is nvidia. All they have to do is publicallyrelease their "best driver ever det 50s", and then they can be used for the benchmarks.

Why don't they release them to the public? What is NVIDIA afraid we'll find? Well, that would be nice, but the reviews could use what hardware they wanted, and what video cards they wanted. However the hard drive with Doom3 could not be kept or copyed.

again, the funny thing is, reviewers were allowed to use any driver there was.

What could they be afraid of? How bout some unimplmented extention or crap, then you have fan boys running around calling it a cheat.

Edge
09-12-03, 12:24 AM
Here, I'll just copy/paste what I said in the other thread:

Hmm, the Anandtech article proves pretty much what we already know. It seems in mixed mode the FX5900 is just a bit below the 9600pro most of the time. Oh, and it appears that even the ATI cards get a big performence drop from DX9 stuff, in many cases the TI4600 under DX8 is pretty close to the 9700 in DX9. Looks like both ATI's and Nvidia's cards have performence hits in DX9...just that Nvidia's is a hell of a lot higher.

Oh, anyone else noticing some odd discrepencies in that Anandtech article? I don't doubt that their numbers are right, but for example under the last benchmark they mention "Interestingly enough, the Radeon 9600 Pro comes out ahead here by a slight margin over the Radeon 9700 Pro", but on the FPS chart the 9700 is ahead by almost 6 FPS. I think they have some of their graphics mixed up. Was someone really tired when they wrote that article or something?

ChrisW
09-12-03, 12:33 AM
Originally posted by GlowStick
Defiantly not, the preliminar benches of the 50's show decreaed preformance where there were known cheats.

Its called, they are afraid of the results and are forceing their results or no one can play it.

While the Doom3 benches were allowed whatever hardware and drivers they wanted.
LOL! They didn't want their game benchmarked using hacked drivers that were developed for one purpose...to artifically inflate the score and mislead the public. If nVidia wants these drivers used in the benchmark then they should release them publicly. Valve does not want their game benchmarked using drivers nVidia has no intention of releasing to the public or face their scrutiny. You might have noticed that nVidia sent these "special" drivers to the reviewers. Why do you think they would do that? Besides, if you are going to allow them to use their beta drivers then ATI should also be allowed to use their own beta drivers.

And the Doom3 benchmarks were made only with an nVidia employee present to oversee the tests.

PreservedSwine
09-12-03, 12:34 AM
Anyone else notice the R9600PRO number are within 1/10th of an fps at 1024 x 768 AND 1280 x1024 ??:wtf:

Evan Lieb
09-12-03, 12:34 AM
Benchmark graphs have all been fixed. Thanks for being patient guys. :)

GlowStick
09-12-03, 12:42 AM
That dose make everything right,

the 9600pro looks pritty intresting right now.

Edge
09-12-03, 12:51 AM
Originally posted by PreservedSwine
Anyone else notice the R9600PRO number are within 1/10th of an fps at 1024 x 768 AND 1280 x1024 ??:wtf:

Now that you mention it, that is pretty damn strange. Wonder if it would be faster then a 9800 if you bumped it up to 1600x1200 :D

Seriously though, that's a very odd issue. It's obviously not CPU limited because the 9700/9800 was able to have a higher framerate, but while the 9700/9800 were losing 10-15 FPS from the resolution change, the 9600 was rock solid. Very fishy...and I'm stumped as to what the reason is.

The only thing I can this of is it's a driver bug, since in the techdemo_5 benchmark the 9600 is at about 40% slower then the 9800 at 1024x764 resolution, yet in the c17_02 demo it's within 6% of it at 1280x1024 res. Well, I guess if nothing else it gives people an incentive to crank up the res on their 9600 card:)

volt
09-12-03, 12:52 AM
Looks like HL2 IS delayed :p But a benchmark to be released on the 30th.

With the game almost done and a benchmarkable demo due out on September 30th, it isn't a surprise that we were given the opportunity to benchmark the demos shown off by Valve at E3 this year.

Clay
09-12-03, 12:52 AM
I took a screen shot from Valve's Source engine movie released yesterday. It looks impressive. This video is the highest quality one I've seen so far.

http://www.nvnews.net/images/news/200309/hl2_b.jpg

FredGSanford
09-12-03, 12:54 AM
Valve was quite insistent that we only used publicly available drivers on publicly available hardware

Thats pretty funny... HL2 Bench itself is not publicly available... Or did I miss it? I don't really care if they optimized the hell out of the Det 50's or not as i would still like to see how they worked out with the HL2 Bench.

Ether way I'm sending back my 5900U for the 9800 Pro... Only question now is, 128mb or 256mb? ;)

volt
09-12-03, 01:02 AM
Nice Max! However my Steam won't update lol.

Also that cardboard looking thing on ground looks weird :p

Clay
09-12-03, 01:06 AM
Originally posted by volt
Nice Max! However my Steam won't update lol.

Also that cardboard looking thing on ground looks weird :p
Thanks! I didn't get it via Steam...lemme see where was that link...oh yeah here it is: http://www.fileplanet.com/download.aspx?f=130227

And yeah, that cardboard thing does look weird. :D

Clay
09-12-03, 01:13 AM
I just took another look at the movie...those are actually pieces of wood that have something like a 1" thickness to them. The dark area to the right of them is the shadow that they cast. The angle I took the screenshot from doesn't do them justice. From other angles they look fine, especially when you can see the thickness to them.

One other thing I noticed is that when you leave the initial "alcove" area and cross over to the other building's rooftop....there appears to be some funky banding with the lighting...not overly noticeable but it's there.

Regardless of that though, the movie is pretty impressive for just being an engine demo flick. :)

volt
09-12-03, 01:19 AM
Thanks for the feedback. I was hoping it only looked weird on the screenshot :)

I would have thought they'll release the movie through steam first. I'll wait :D

FredGSanford
09-12-03, 01:22 AM
You could always grab the video from http://www.3dgamers.com without having to wait on those long and slow lines. Came down at 256kbs for me from 3dgamers.

Hellbinder
09-12-03, 01:24 AM
Here is the Key point from Anand Article in my opinion.

The sacrifices you encounter by running either the mixed mode path or the DX8 path are obviously visual. The 5900 Ultra, running in mixed mode, will exhibit some banding effects as a result of a loss in precision (FP16 vs. FP32) but still looks good - just not as good as the full DX9 code path. There is a noticeable difference between this mixed mode and the dx82 mode, as well as the straight DX8 path - for example, you'll notice that shader effects on the water aren't as impressive as they are in the native DX9 path.

Are the visual tradeoffs perceptive? Yes. The native DX9 path clearly looks better than anything else, especially the DX8.0/8.1 modes.

There is a Difference In IQ. One can only wonder what even Bigger trade offs Nvidia will try to get away with in the Det 50's

There are about 5 or so other major points in there as well.

volt
09-12-03, 01:25 AM
Originally posted by FredGSanford
You could always grab the video from http://www.3dgamers.com without having to wait on those long and slow lines. Came down at 256kbs for me from 3dgamers.


Ahh it's there :dance: