This is only tangential, but what I still haven't heard is whether DX10 and Vista support is such a discontinuity that we may be looking at parallel product lines from one or both IHVs in transition longer than you would normally expect, and if so what kind of impact that has --do you end up doing an R590 or G7xn after Vista ships, for instance?
Next NV High-end
- Thread starter Geo
- Start date
geo said:
Not sure if i buy that, vista is targeting gaming and graphics far more then any other windows based OS, i'd think MS wants to have launch titles with DX10 to show it off right away, as opposed to a DX9C thing where we still dont really have any titles using much of the actual large graphic effects all these months later. We already know companies have the DX10 API (Crytek) so i can only assume both Nvidia and ATI are well under way with full DX10 support on their pure breed nextgen cards. And as such im sure both companies would discontinue any cards previous to DX10 phasing them out as fast as possible. I mean once vista ships, the people that are really going to be sucking it up initially i think will be enthusiasts and gamers. Obviously major company computer purchases and major OEM sales dont come primarily from things like DirectX offerings. So whos going to pay 300-400+ for a high end card of previous gen. It would be like mentioning the high end 6800 line getting a ressurection after the 7800 launch, i dont see that flying just as i dont see an R590 or G7X coming out right after an R600 or G80..
Last edited by a moderator:
Joe DeFuria said:
I dont think Nvidia wants to delay for long, personally i think they've already been delaying the whole time. Look at the core name change for example. NV47 initially to simply act as a refresh to the NV40 turned into a full fledged flagship core with quite a work over, further more the NV47 turned into the G70, and i believe that was done because they simply would of run out of numbers. I have a conspiracy theory once the actual full DX10 core launches they may revert back to the NV line of core titles. Thats my conspiracy though. Im waiting to see if there really will be a core named the G80 and if it really will be the NV50 lineage which will surprise me a little. Again, my thought, is that any further cores signed with the "GXX" codename will be cores we see between the NV40 line and the true NV50.
nVidia has, in the past, launched refresh parts as if they were entirely new products. I wouldn't read anything into the launch of the GeForce 7800 GTX, as it's clearly nothing more than a refresh part. I would similarly read nothing into the naming change.
Expect product generations to continue to lengthen over the next few years, with more refreshes per generation, with each one called "new" by the marketting people.
Expect product generations to continue to lengthen over the next few years, with more refreshes per generation, with each one called "new" by the marketting people.
{Sniping}Waste
Regular
I see something that is being missed. The X1800XT engineering sample are able to hit core speeds around 700mhz and up. In the past, ATI had a XT PE verson for reviews but not this time. I think ATI is waiting to see what NV will release to compeat with the X1800XT 512 meg and release a X1800XT PE 512 meg to trump what NV puts out with a core around 700mhz and ram around 780/1560DDR.
{Sniping}Waste said:
I could see an XTPE part happening if Nvidia releases a significantly higher clocked GTX with 512 mb such that it dominates benchmarks. I think ATI will be fine with a competetive part until r580, I doubt they feel a need to be dominant.
AlphaWolf said:
Agreed, and I'd bet that the 512MB G70 will indeed be higher clocked just because NVDA does feel the need to be dominant (and who can blame them considering the margins they've been raking in on the high end since NV40?). Except this time I think ATI will be more prepared to respond, assuming those yields are much improved as has been reported unofficially.
{Sniping}Waste said:
I'd say we first should see adequate X1800XTs on shelves before we can even talk about higher frequencies. They're having availability problems with 625MHz, I don't want to even think how rare they'd be at 700MHz right now.
Both IHVs would be better off to concentrate on followup products (R580/G7x-whatever) than higher clocked versions of existing GPUs. I can't exclude the possibility of G70's with bigger framebuffers, yet when it comes to frequencies there are already more than a few vendors selling at higher clockspeeds.
Yes, but the tradeoff was that ATI was offering slightly higher performance with less features (not to mention very late availability relative to NV40), which wasn't enough to dethrone the 6800GT/Ultra from their darling position at the high end. This time it's slightly better performance along with more features. I'm not sure Jen-Hsun will stand for that. 
D
Deleted member 2197
Guest
Last edited by a moderator:
Slightly OT, but according to Firingsquad, nVIDIA have launched the GeForce 6600 DDR2 (8ps/3vs 350MHz Core/400MHz Mem, 128-bit 256MB RAM) to compete with ATI's RADEON X1300 PRO, which as Dave suggested further upstream in relation to the 6800 GS SKU, could either be an indicator of an imminent product launch (7600, 7200, etc) as nVIDIA clears NV4x inventory or a variable delay in new products as existing products are consolidated.
http://patft.uspto.gov/netacgi/nph-...&co1=AND&d=ptxt&s1=nvidia&OS=nvidia&RS=nvidia
Version posted this link in the console forum.
It's dated for today. Is this not G72, Nvidia dual-GPU part?
I apologize if this discussion is more about G80 at this point, but I'm curious how load balancing would work if at all beyond what already happens with SLI except in a much more efficient manner.
Version posted this link in the console forum.
It's dated for today. Is this not G72, Nvidia dual-GPU part?
I apologize if this discussion is more about G80 at this point, but I'm curious how load balancing would work if at all beyond what already happens with SLI except in a much more efficient manner.
kemosabe said:Yes, but the tradeoff was that ATI was offering slightly higher performance with less features (not to mention very late availability relative to NV40), which wasn't enough to dethrone the 6800GT/Ultra from their darling position at the high end. This time it's slightly better performance along with more features. I'm not sure Jen-Hsun will stand for that.
One shouldn't forget that ATI has still a large delay and lacklustering availability - for the time being - for the ultra high and mainstream part of the market on it's back.
Judging from the according reactions lately from NV's PR department, I can't sense that they feel particularly threatened right now. Is this whole "Ultra" expection mostly because there's a model missing theoretically with such a name or because of Fuad's senseless ramblings at times?
scificube said:
Errm, that looks like shared memory for SLI to me --anybody disagree?
Where's the rest of my committee?
We're still accepting new memberships, btw. http://www.beyond3d.com/forum/showpost.php?p=597621&postcount=228
Last edited by a moderator:
scificube said:
Seems like this would simply do away with one of the major shortcomings of today's dual-GPU solutions - framebuffer wastage. Might have some performance impact as well.
Actually it does seem quite focused on performance improvements as well. I didn't know that off-card memory transfers were still necessary in the case of dual GPU's on a single board.
Last edited by a moderator:
