Some Thoughts

Username

Banned
Anyone remember the R300 launch, when the previews didnt have real numbers, just provided benchmarks under ATI controlled settings? This board seemed to embrace that with open arms. I'm shocked ( :LOL: ) at the reaction the last two days!

Anyone remember how the R300 was (and still is) only marginally faster than the Ti4600 except in extreme bandwidth circumstances (e.g. resolutions that most consumers monitors cant even support?) but 30 percent increase over the R300 in those same resolutions is "not fast enough"?

Anyone remember how important it was that the Radeon 8500 had a better pixel pipeline feature set, but going beyond DX9 specs is completely irrelevant (even though its painfully obvious that there are a few high profile developers and software packages lined up to take advantage of the additional features as opposed to other products)?

Anyone remember when launching with very few technical details and vague marketectural information was okay for the R300, but is suddenly a horrendous event that shows the desperation and unscrupulousness of Nvidia?

I could go on and on, but feigning impartiality in the face of the last year's worth of postings is pretty ridiculous. Ripping on other websites, denigrating posters on other boards (or people who's opinions dont agree with yours) is all a pretty sad degeneration.

I knew that regardless of what was announced yesterday, the spin would be negative. The number of posts cautioning people before the announcement that it likely wouldnt be faster was pretty consistent, and there are still posts to that effect on the board. By nearly any metric the NV30 is faster than the R300, across all resolutions.

I wonder if Beyond3D will Benchmark at 2048x with 4x AA (A situation that should clearly favor R300's "bandwidth advantage") and will they have the gall to state that the NV30 is 2.5 times as fast as the R300. We'll find out soon enough I suppose.

My personal opinion on the event is that it was a pretty fantastic success. All the people that actually matter at this point not only got to see and play on real hardware (many for the first time), it was aimed at squarely reassuring Nvidia's partners that the chip was done and they are ready to go. Take care of your content people and your partners first, then worry about the screaming vocal minority web community is a smart and effective strategy.

By the way, for the people that are at Comdex, what did you think of Sweeney's demo? I'm still completely in awe of it, its really what I expected out of the GeForce 3 class. I've been told he is also going to be working not only on soft shadows, but adding umbral and penumbral effects to those shadows as well. I also think Dawn is 1000% more marketable than the werewolf, although she may need a bit more clothing.

I'm also interested what your observations of the reactions of people on the demo stations, its nice to hear not only different perspectives, but peoples impressions of people's different perspectives.

And to make this post not a complete waste to those who are going to be offended by it.. NVIDIA IS D000000000M3DDDD!!!!!!111111 ;)
 
Well, that's quite correct, but people (I) always wants more. I think we still don't know much about the NV30 (and seems to be some obscure things about R300 also :-? ).

Till the new board of Nvidia seems to be more performant (and i don't mind if it's 0.13 or not). Sure, i was hoping for a bit more (in term of feature)
 
Username said:
Anyone remember how the R300 was (and still is) only marginally faster than the Ti4600 except in extreme bandwidth circumstances (e.g. resolutions that most consumers monitors cant even support?) but 30 percent increase over the R300 in those same resolutions is "not fast enough"?

We had pointed out some time ago that in normal resolution rendering, yes, even a Ti4600 was CPU limited, as was the case for R9700. Its not the normal resolution rendering what is really much of a concern at the moment, its the AA performance which its plain to see that R9700 had the upper hand on everything else at the time. Theoretically GeForce FX has the same techniques where AA is concerned (although I'm not sure about fast colour clear on R300), so its going to be interesting to see how fast it is.

I wonder if Beyond3D will Benchmark at 2048x with 4x AA (A situation that should clearly favor R300's "bandwidth advantage") and will they have the gall to state that the NV30 is 2.5 times as fast as the R300. We'll find out soon enough I suppose.

Supply me with a display device capable of 2048 and we will, if no-one does then we won't.

Again, normal resolution rending isn't a concern on current games, only AA and filtering performance is.
 
Also dont mistake bias within forum members for bias by the web site owners/mods/contributors.
 
yes but i dont see Wavey/ John etc flaming nVidia like the post above implied which also implied they would test the NV30 in a way that wouldnt suit it.
 
Randell said:
yes but i dont see Wavey/ John etc flaming nVidia like the post above implied which also implied they would test the NV30 in a way that wouldnt suit it.
I'm not saying that, but i think if you prefer one company to another (or one tech to another) you have some bias to be more appraisal to this company (tech) than the other way (surely less on B3D, or more subtle :D)

And if B3D was not what it is (well documented, in deep reviews, interesting interviews, etc.) i wouldn't be there ;)
 
You mean this messageboard gave a negative outlook on the recent announcement regarding nVidia's latest product? Oh my... I'm shocked.

Pfft..
 
Randell said:
SvP said:
Going beyond DX9 specs would be really impressive, but there are no products that goes, not R300, not NV30.

?

!

After beta1 dx9 was externded to include R300 / NV30 extra features, as well as a lot more that no one knows who will support...
 
Randell said:
SvP said:
going beyond DX9 specs is completely irrelevant

Going beyond DX9 specs would be really impressive, but there are no products that goes, not R300, not NV30.

?

AFAIK DX9 specifies both PS and VS versions 2.0 and 3.0, no? :-?

Anyway, from B3D NV30 preview: As we can see GeForce FX still stays within the Pixel Shader 2.0 specification; however, it does go beyond it in certain circumstances.
 
Some thoughts on your thoughts...

Anyone remember the R300 launch, when the previews didnt have real numbers, just provided benchmarks under ATI controlled settings? This board seemed to embrace that with open arms. I'm shocked ( ) at the reaction the last two days!

There are a few big differences between ATI's launch and nVidia's launch.

1) though ATI did "control" the benchmarks (controlled what could be benchmarked, and on what machines, etc.), ATI did allow people to run benchmarks on hardware. ATI only allowed reporting of "relative" scores, but ATI did not simply give out numbers.

2) ATI's scores absolutely blew away the nearest shipping competition. This is the biggest reason for all the excitement and "acceptance" of 9700 scores. We all knew that even if the scores changed a little, there was no reason to think the end result wouldn't be staggering. Today, the shipping competition is the Radeon 9700, and it is still very uncertain how the NV30 will stack up against it. Unfortunately for nVidia while the NV30 performance improvements over the Ti 4600 are equally as dramatic as the 9700 over the 4 Ti...they are doing it 5-6 months late. People having that "ho-hum...been there done that" attitude is basically the price that nVidia pays for being late.

In short: there was no quesiton that the 9700 would out-class the GeForce4, despite the non "official" benchmarks at 9700 launch. There is considerable question if the NV30 will beat the 9700 at all, and if so, by how much.

3) When 9700 launched, they announced shipping by 30 days. That gives everyone confidence that these things were very close to final, and ATI was merely ramping up production and doing minor tweaks. So we could feel pretty confident over the scores. The launch date of NV30 is still "vague". January - February"ish". This does not instill confidence that even nVidia knows exactly what they will ship, in how much quantity, and at what clock rate.

Anyone remember when launching with very few technical details and vague marketectural information was okay for the R300, but is suddenly a horrendous event that shows the desperation and unscrupulousness of Nvidia?

The main difference here: ATI was VERY tight lipped about the 9700 prior to launch. nVidia had given out specs to certain web sites (like Anand) LAST MARCH. And those sites started pimping it almost immediately. "Looks better than R300 on paper....wait for NV30"....The Cg whitepapers and "48 GB/Sec effective bandwidth" stuff didn't help nVidia in the end...it hurt them. Everyone was sort of expecting some miraculous advancement beyond R-300 because of all of that leaked and pre-launch hype. It turns out we knew pretty much everything of significance about the NV30 already, BEFORE the launch. The exception being the very high clock. In practically every other area where there was speculation: number of texture pipes, different AA methods, bus width, bandwidth savings techniques....the correct speculation all turned out to be the more "conservative" speculation.

In short.... the NV30 launch specs did not live up to the hype.

...it was aimed at squarely reassuring Nvidia's partners that the chip was done and they are ready to go....

Actually, that is not the case if they are waiting for another chip spin to come back. That's the thing. Being able to play on real hardware is of course a significant event...but not quite as confidence inspiring as saying the product will ship in "30 days." NVidia did not commit to that.

Anyone remember how important it was that the Radeon 8500 had a better pixel pipeline feature set, but going beyond DX9 specs is completely irrelevant

You seem to say that there weren't arguments for both sides here on that issue?

In any case, there was a difference: 8500's advancements were officially supported in an official version of DX 8.1. So at least the argument could be made that it might get some support in Driect X games. It doesn't appear (except for some variable caps on available registers?) That NV30's advancements will even be supported to any further extent than R-300s in DirectX 9.

Going beyond PS/VS 2.0 is not completley irrelevant...but unless it is actually PS/VS 3.0 compliant, NV30's advancements don't mean much over the R-300 for DirectX games.

OpenGL is of course a different story.
 
Being able to play on real hardware is of course a significant event...but not quite as confidence inspiring as saying the product will ship in "30 days." NVidia did not commit to that.

A clear illustration of what I mean:

http://www.tomshardware.com/business/02q4/021120/index.html

NVIDIA clearly succeeded in getting people talking about this yet unreleased product. The only people who didn't have a lot to say were clearly the NVIDIA OEMs that we visited with today, who in fact had little information to offer on the GeForceFX. Clearly the obvious question is, "When will cards become available?" and we asked more than one OEM for an answer. The party line was the same: as soon as we have chips from NVIDIA. So, no-one has a more specific date to purchase this monster, and the mystery adds spice to the launch.

I have not seen one OEM Press Release officially announcing their upcoming GeForceFX product (are there any?)...normally we get a slew of those things accompanying the chip launch...
 
SvP said:
Randell said:
SvP said:
going beyond DX9 specs is completely irrelevant

Going beyond DX9 specs would be really impressive, but there are no products that goes, not R300, not NV30.

?

AFAIK DX9 specifies both PS and VS versions 2.0 and 3.0, no? :-?

Anyway, from B3D NV30 preview: As we can see GeForce FX still stays within the Pixel Shader 2.0 specification; however, it does go beyond it in certain circumstances.

I dont know for sure hence teh ? - but thank you for the clarification
 
I believe that there is some confusion in that regard. NV30 goes beyond PS/VS 2.0, but since DX9 has PS/VS support all the way to 3.0 it does not go beyond DX9 specks.
 
DaveBaumann said:
Supply me with a display device capable of 2048 and we will, if no-one does then we won't.

Funny you should mention that, Dave. I ever tell you my Sony Trinitron FD-based KDS monitor does 2048x1536? :LOL:

I should start a poll. Based on the above, who really thinks Dave is going to send John his NV30 review board.

1. Of course he will
2. Don't hold your breath
3. Not even after rigor mortis has set in
 
Back
Top