I suggest you to read KZ2's deferred renderer presentation, they do read z subsamples (see their shadowing solution). I didn't mention DX10.1 just for the sake of it.
I wish I could comment on this
Last edited:
NVIDIA GT200 Rumours & Speculation Thread
- Thread starter Arun
- Start date
- Status
thegrommit
Newcomer
What's the phrase? People in glass houses shouldn't throw stones?
What gets me is the implication from various people (not pointing at Hawkeye) that lower power consumption is equated with lower performance. Intel's Conroe is an example of how this is untrue. I expected better from the denizens of B3D.
jimmyjames123
Regular
I've gotta say that it really irritates me that some people choose not to honor the NDA. It's not just insulting for NVIDIA and AMD who provide the review samples and confidential information in advance, but it's insulting for the reviewers who actually honor the NDA and spend countless hours working up reviews about exciting new products only to have a small number of people leak all the details in advance.
I am still old-fashioned and enjoy reading reviews on new products to actually learn about new and exciting features on the products. Having everything leaked in advance takes some of the fun out of reading (and writing) reviews.
Last edited by a moderator:
IF ? .. and .. letting? it is rather purposeful, wouldn't you say?
- i t would appear to me that Nvidia is releasing info measure by measure to answer speculation and rumors about r700 being a competitor. And it appears they are minimizing a possible quick shrink as being "unimportant" to performance, i would think - to encourage people to buy - right now. I will tell you, they got me. My NewEgg "deferred payment" account is ready! Just when GT280 hits MSRP i am on it [and frankly i am a bit ashamed of myself but still feel like a kid - for the first time since .. since . . . since i was disappointed by nVidia's first 'ultra']. i don't think i am going to be disappointed but i am waiting for the reviews. As usual, AMD is probably going to be smart about pricing their GPUs and i may again be tempted but i doubt it.
and most importantly .. what time do we look forward to your post?
To each his own. i love getting ready to buy by gathering as much info as possible and i am really sick and tired of my 8800GTX or my 2900xt Xfire. i have been looking forward to Tesla, for months! - since i heard it taped out Last December!
jimmyjames123
Regular
Oh come on, you really think NVIDIA is purposely publically leaking information that has already been distributed to reviewers who are under NDA? That wouldn't make any sense. What you are seeing here on the forums is one or two people who choose to ignore NDA and post the material.
IMO, an NDA is more a benefit for reviewers who honor the agreement, as opposed to a benefit for the IHV. Surely by the time that samples have shipped to reviewers, the IHV cannot expect the info to remain secret from the other IHV competitor.
On the forums, it's all too easy to encourage people to post confidential info. Some of it is very interesting info too. But the people who break NDA should realize the potential consequences not just to themselves or to the IHV, but to the reviewers too. It's just not an act that we should be condoning.
IMO, an NDA is more a benefit for reviewers who honor the agreement, as opposed to a benefit for the IHV. Surely by the time that samples have shipped to reviewers, the IHV cannot expect the info to remain secret from the other IHV competitor.
On the forums, it's all too easy to encourage people to post confidential info. Some of it is very interesting info too. But the people who break NDA should realize the potential consequences not just to themselves or to the IHV, but to the reviewers too. It's just not an act that we should be condoning.
Last edited by a moderator:
If your referring to me. Its kinda a "Whats the point" scenerio. it seems people here dont mind leaking crap all the time. NDA"s are a frustrating thing at times to sign because all the people who leak ((and yes most of the people who know about the chip are under NDA or being leaked too by someone under NDA)) screw the honest editors who actually do honor NDA. And I'm just frustrated about it as you are.
It all comes down to the point. "Whats an NDA" Question. And it basically is something that nobody seems to give a damn about respecting. I really wish Nvidia would do something about it. But what can they do about all these people posting anonymously. Editors like you and guys like myself dont have that "anonymous" luxury and I get my ass reemed if I talk too much. With people like CJ leaking a new slide every day. ((Sorry CJ I mean no disrespect to you but it does piss me off just the same.))
Chris
No it has nothing to do with that. People just cant keep themselves from talking these days. I see a new slide leaked every day and I cant help but throw my arms up in the air and go "WTF?". To me the Nvidia NDA's have been a joke lately. ((Not in a sense I dont try and respect them)). But it seems like the real NDA is after editors day when every one with a camera is leaking crap...
Last edited by a moderator:
Berek
Regular
I'm really eager to see the performance gap and overclocking ability of these cards. I just can't justify spending $650 for the top end card, especially when the GTX 260 at a more reasonable $450 looks to be not that much slower.
1GB vs. 896MB of memory is negligible... over 640MB is probably what matters most in the extreme conditions. Shader count is only 1/8 that of the GTX 280, the rest is the same. Only clock speeds are a real unknown factor at this time, and perhaps overclocking will be good here (although with the power this thing requires, heat dissipation, and its DDR3 limits, I'm not holding my breath).
Do we have any solid evaluations on the performance differences between the GTX 280 and 260?
1GB vs. 896MB of memory is negligible... over 640MB is probably what matters most in the extreme conditions. Shader count is only 1/8 that of the GTX 280, the rest is the same. Only clock speeds are a real unknown factor at this time, and perhaps overclocking will be good here (although with the power this thing requires, heat dissipation, and its DDR3 limits, I'm not holding my breath).
Do we have any solid evaluations on the performance differences between the GTX 280 and 260?
Ding ding. This is the correct answer. 3dB increase corresponds to twice the pressure.
The problem for publications is that channel NDAs expired yesterday, so info is bound to leak. Nothing much can be done about it. What can an IHV do if they want a hard launch so must brief AIBs/OEMS->wholesalers/distributors->resellers prior to the launch. They're unlikely to bite the hand that feeds them...
how come my hifi starts at -82db and ends at zero what a stupid measurement system
is that a typo do you mean 7/8 ?
Anarchist4000
Veteran
When dealing with the audio signals you're typically attenuating the signal. So 0dB means you output the thing at full blast. -80dB would mean you're attenuating the signal down to 1/10,000 of the input.
A doubling at 3dB would mean you're measuring values in watts. For all calculations I've had to do it's always 20log10(out/ref). That's good for any voltages, amplitudes, currents, etc.
Quoted for truth.
This is all about corporations controlling media information flow as part of their PR machine. What information is available to whom, when can it be presented, prebaked slides and illustrations that can be neatly used in articles, et cetera.
These NDAs are about the timing of the release of PR material. Nothing else.
If you mind seeing it a few days before the sanctioned date, just don't go to sites where you are likely to find such material.
No offense taken. I understand completely.
But you do have to realize that I did not sign any NDA at the moment and I'm not responsible for the people with NDA's who leak stuff to me. I'm just serving the community and try to satisfy their craving for new info. People want to know stuff, I do my best to get my hands on it... It's just the way how it goes. Hence the leaks on both NV (slides) and AMD side (X5500> score for R700).
I would however like to point out that I do keep myself to NDAs if I would've signed one.
The other point not mentioned here that makes DX10.1 Deferred Shading + MSAA more efficient is the ability to execute a pixel shader per sample. If you don't have that (DX10.0) you are reduced to resolving on-the-fly (which has edge quality problems when optimized light volumes are used and prevent further effects requiring access to the individual post-shaded MSAA samples e.g. HDR tone mapping) or masking each sample output in turn and multipass your light geometry.
http://www.youtube.com/watch?v=K9gwJwCNvT8
Short, poor quality video of the Nvidia Medusa demo? Read from VR-Zone about the demo but never seen or heard anything else about it until now.
Short, poor quality video of the Nvidia Medusa demo? Read from VR-Zone about the demo but never seen or heard anything else about it until now.
i can agree with how you feel and i certainly can't expect anyone with sound morals to purposely release info under NDA; if i were under a NDA, i would respect it - or refuse to sign it.
However, i also know for certain that Nvidia leaks info on purpose; beyond simple Focus Group info - i have been the recipient of such leaks - especially from Nvidia at my old forum - and i believe i understand their reasoning on both sides. i believe Nvidia's very attitude toward NDA itself is changing as is AMD's [a bit tighter]; it appears that it is not a hard and fast rule and NO ONE expects the many people at the Editors Day to keep everything in strictest confidence. They are building excitement for their Tesla line. CJ is not under NDA - simple as that! There are people who leak to him. [thankfully, imo]
Again what is the *purpose* of NDA? i'd say the primary reason is pragmatic - to keep their competitors from knowing what in your hand and from knowing what cards [pun intended] that you hold. However, in THIS case, Nvidia appears over-confident and does not CARE about what AMD has - so the leaks are not as controlled as before. That is my take on it and i have seen strict NDAs and loose ones in the past also.
Last edited by a moderator:
- Status
