NV30 Conference
- Thread starter alexsok
- Start date
http://www.nvmax.com/cgi-bin/community.pl?num=1028668195
blech, those are already turning out to be some of the lamest questions possible. I think the only thing thats really on my mind right now is "is the thing taped out or not? if so, could we be on the final rev this time?"
blech, those are already turning out to be some of the lamest questions possible. I think the only thing thats really on my mind right now is "is the thing taped out or not? if so, could we be on the final rev this time?"
Mulciber said:http://www.nvmax.com/cgi-bin/community.pl?num=1028668195
blech, those are already turning out to be some of the lamest questions possible. I think the only thing thats really on my mind right now is "is the thing taped out or not? if so, could we be on the final rev this time?"
If you think that's lame, take a look here:
http://www.ve3d.com/comments.taf?postID=29913
Just goes to prove that the ONLY source that can be trusted with good questions is Beyond3D!
Nappe1 said:Mulciber said:... I think the only thing thats really on my mind right now is "is the thing taped out or not? if so, could we be on the final rev this time?"
that sounds like Bitboys Q&A...
LOL, at this point, thats what we've been reduced to
Seriously, what else really matters about the chip for the next 6 months, if it hasn't produced initial silicon yet
...and IMO, situation at the nVidia might pretty same as Bitboys had on the time when they were firing up Glaze3D development...
(let's see If I am right this time... we'll see if the NV30 has 256Bit bus, though they said they don't need a one in March - April timeline...)
Yep, a kind of creepy situation; because everyone else has been gone to the 256Bit bus, they might have got first revisions from factory and found out that 128Bit bus just isn't enough. This causes them to pull the chip back to the R&D to update memory controller to handle wider bus.
for company like nVidia I doubt it causing more than 3-5 months delay if the MCU has been designed so that it's easily upgradable. But if they are trasforming things to the UMC at the same time, it could be devastating and pretty expensive move.
IMO we are living VERY interesting times right now. I think there's enough ppl to remember what happened to "3dfx's unreleased monster chips" when they decided change the fab in the middle of the design process...
and yes, (after this long off topic thingy) I am asking same question too: Has NV30 gone thru tape out?
(let's see If I am right this time... we'll see if the NV30 has 256Bit bus, though they said they don't need a one in March - April timeline...)
Yep, a kind of creepy situation; because everyone else has been gone to the 256Bit bus, they might have got first revisions from factory and found out that 128Bit bus just isn't enough. This causes them to pull the chip back to the R&D to update memory controller to handle wider bus.
for company like nVidia I doubt it causing more than 3-5 months delay if the MCU has been designed so that it's easily upgradable. But if they are trasforming things to the UMC at the same time, it could be devastating and pretty expensive move.
IMO we are living VERY interesting times right now. I think there's enough ppl to remember what happened to "3dfx's unreleased monster chips" when they decided change the fab in the middle of the design process...
and yes, (after this long off topic thingy) I am asking same question too: Has NV30 gone thru tape out?
Captain Chickenpants
Regular
alexsok said:Mulciber said:http://www.nvmax.com/cgi-bin/community.pl?num=1028668195
blech, those are already turning out to be some of the lamest questions possible. I think the only thing thats really on my mind right now is "is the thing taped out or not? if so, could we be on the final rev this time?"
If you think that's lame, take a look here:
http://www.ve3d.com/comments.taf?postID=29913
Just goes to prove that the ONLY source that can be trusted with good questions is Beyond3D!
ROFL!
[edit] Just to clarify I am laughing at the posts on that board, not you. [/edit]
Qroach said:
I doubt it...
We all know pretty much everything about the CineFX architecture, while we pretty much know nothing for sure about NV30.
Now, if the information would be about NV30, he would surely not be allowed to talk about it, so the fact that Nvidia does allow him to talk about it only confirms (probably) that the stuff he is able to reveal is all the stuff that is mentioned on Nvidia's whitepapers.
Although, I'd be happy if I was wrong!
Nothing too interesting. Many stuff are similar to the CineFX paper available on NVIDIA's site. They show in what cases the CineFX architecture beats the R300 and also said that one of the official specs of the R300 may not be correct. Don't ask me which one ATM though
Nothing on speific products thiugh - it was about architecture. They won't tell me clock speeds or stuff like that. I have asked for such (and other stuff) in a follow-up email (the conf call itself was a near disaster... see my last post in the "Questions about NV30" Hot Threads link on VE's front page)... don't know if all my questions will be answered.
Nothing on speific products thiugh - it was about architecture. They won't tell me clock speeds or stuff like that. I have asked for such (and other stuff) in a follow-up email (the conf call itself was a near disaster... see my last post in the "Questions about NV30" Hot Threads link on VE's front page)... don't know if all my questions will be answered.
Reverend said:Nothing too interesting. Many stuff are similar to the CineFX paper available on NVIDIA's site. They show in what cases the CineFX architecture beats the R300 and also said that one of the official specs of the R300 may not be correct. Don't ask me which one ATM though
Nothing on speific products thiugh - it was about architecture. They won't tell me clock speeds or stuff like that. I have asked for such (and other stuff) in a follow-up email (the conf call itself was a near disaster... see my last post in the "Questions about NV30" Hot Threads link on VE's front page)... don't know if all my questions will be answered.
For anyone who doesn't want to search through VE, I'll post Reverend's rather funny conversation here:
There was no Q&A. It was a briefing involving two NVIDIA guys. The session itself was a near-disaster.
Just before the call came for me, about midnight here in Malaysia, my 3-year-old son woke up crying. Both me and the woman thought it was a case of a bad nightmare (as he's had these before). Then the call came in. So I thought my wife would be able to calm our son down and I decided to take the call. Throughout the call, my son kept on crying and kept asking for me. Knowing that there was a bit of effort involved in arranging the conference call (my main NVIDIA contact, the one that arranged this, is in Texas, the other guy is in California, both were involved in the call), I decided to just try and let the conference take its course. One hour later, when the conference appeared to have reached its end, I had to tell the two guys that I have to end the conference rather abruptly coz my son was still crying. Anyway, to cut it all short, I didn't get to ask any questions during the call. That will have to be through email, which I've already sent. As for what was wrong with my son (which I'm sure everyone reading this will be more concerned about, instead of not getting any NV30 info here
Also, the conference was about "architecture" as opposed to "specific products". I misunderstood. The entire briefing was about the architecture that will be present/supported by NV30 and NV3x. I did press about a couple of things. One was about Cg and the word was essentially Cg = DX9 HLSL, absolutely identical. The other was when I asked how they can achieve "fast performance while delivering the pretty pictures as the architecture supports them" (fast performance were key words by the NVIDIA guy throughout the conference). The answer was "Unfortunately, I can't reveal in this conference how we can achieve that."
Reverend - the answers u receive from Nvidia (assuming u receive them), will you post the answers together with the questions here on the board? Or they will be strickly under NDA?
Matt Burris
Newcomer
Trawler said:
FWIW, in the long 3 year history that 3DGPU been around, NVIDIA has always thrown these briefings/conferences. It's their way of keeping the community informed, and also a way to get some questions in, if you have any. This was the case when Cg came out, when the nForce came out, when the GeForce2 GTS came out, and so on. So it's not damage control, but just a NVIDIA tradition for as long as I've known it.
Hellbinder
Banned
I still dont understand why the NDA stuff at this time... They are touting Cg, unofficially claiming the Nv30 is the same Cinfx engine etc... Launching in october (supposedly) with a real product.... only a month after the retail release of the R300... I just dont get the secrecy.... at this stage of the game.
And whats with that comment about one of the R300's specs being incorrect??? Do they have one?? I dont think so. I am going to try not to get upset, but it is already starting to sound like the same old Nvidia misinformation habbit ala Kyro doc.
Hellbinder[CE said:
If you do get upset, please do it elsewhere.