Next gforce card comes out when?

[Ampz]

Anyone know anything about when a newer card (newer then the 7800 gtx) will be coming out? What it will be called etc? Thanks!

[Kanavit]

i think the next gpu will be the G72 core, which features new 90nm process.

[andL64]

sources said the really new core comes, early 06.

there will be some 512Mb versions of the GTX, in a few weeks

[del_fuego]

The next card could be real soon. Check this out - http://www.xtremesystems.org/forums/showthread.php?t=76808

Cya.

[***Deimos***]

Although technically there will likely be another 7800 card soon, with possibly higher GPU clock in addition to faster/more memory (which seems like a given), I think the person who started the thread was wondering about a GF8... wow.. such high number sure sounds weird.

In that case, it will be quite a long time, considering how recently the GF7 was released. Ofcourse there is little benefit to getting a head start when there is no support from the OS side (DX10).. so whenever new Windows Vista gets rolling with new DirectX.. you can count on it, that ATI and nVidia will be there with new cards to take advantage.

[Geforce4ti4200]

I heard both ati and nvidia will roll out new cards in early 2006. ATI will need 32 pipes this time and their x1800s do have 32 pipes, just that current ones have 16 bad pipes

[alexio]

I heard both ati and nvidia will roll out new cards in early 2006. ATI will need 32 pipes this time and their x1800s do have 32 pipes, just that current ones have 16 bad pipes

R520 is 16 pipes as I corrected you in another thread.

Why are you so sure the R520 has 32 pipes? Please don't spread false info and if it isn't BS what you're saying then show us some proof please :rolleyes:

[Unoid]

pehaps hes just estimaing by the transistor count.

the r520 > 7800gtx in transistor count. Most would think a 16pipe card would be much less than a 24pipeline card.

I only see ati comming with a 20-24 pipeline card for the r580.

Hell a 16pipeline card with a at 750mhz/1800mhz would do just fine.

[***Deimos***]

Dont forget they need time to sell of at least a few X1800XT's.. which dont show up until next month. Unless ofcourse ATI believes that nobody wants the X1800XT and market demand is pathetically low.. (www.rage3d.com - ask a big ati fan.. they'll tell you the R580 is the "real" R520.. just as R600 is real R500 and so forth)

Some of nVidia's biggest successes like GF2 and GF4 owe it to the fact they had a long time on the market to establish presence with consumers.

[perkam]

Speculative threads dont belong in the X3D Section I'm afraid. Xtreme 3d Results ONLY.

Thx.

Perkam

[Jupiler]

Topic moved to General Hardware section.

[***Deimos***]

I heard both ati and nvidia will roll out new cards in early 2006. ATI will need 32 pipes this time and their x1800s do have 32 pipes, just that current ones have 16 bad pipes

You are obviously fixated on the big leap in # of transistors.

Lets for a second forget the major issues why you are wrong:

Ultra-threading requires much more dynamic logic that simple scheduler used in R300/R400.
New memory controller implementation requires more transistors
SM3 features require more transistors.. including bigger ALU/FPU to handle 32bits.
Increased number of vertex shaders requires more transistors
AVIVO video processing requires more transistors
Additional ALU/FPU per each VS/PS like 6800->7800 obviously requires more transistors
Larger register file, and more temporary and general registers, and more logic to keep more instructions in flight requires more transistors.
Finally, new AA/AF features, as well as various other optimizations require more transistors.

As far as I know, neither ATI or nVidia ever provide actual count of transistors... the approximate # of gates is multipled by avg #/gate factor. You gotta realize not all transistors are alike. Their size is scaled to mitigate RC delays, larger ones are used for IO, and various other aspects changed for power regulation schemes. X800 supposedly had 160M and X1800 has 320M so I see where your simplified confusion comes from.

Lets look at it from business sense. Die space is expensive. Why would ATI make huge die with 32pp and disable half... Similarly, you could use the analogy of AMD launching 90nm with dual-core chips, with 99% of them with one core disabled.. its just so absolutely absurd. The larger the chip, the lower the yields... and furthermore, you get more chips with smaller die... its why Intel uses a separate smaller die for Celleron.

[Tim]

Well, the Ultra canned rumour is probably bogus...so I'd say the 1st of November date that you see floating around could be the one...

What about specs?

1.26 Ram higher clocks? 110nm tweaked core @ 490-525? 32pipes? (extra pipes I don't think so in 2005, but you can always dream, there has been one or two rumours about 32 pipes...we'll see I suppose)

nVidia told peeps in Amsterdam to expect an Ultra very soon...and I don't think that's gonna be the end of the year :)

:Edit, The first thing I'm gonna do when they release that card, and when I actually have one in my posession, is pop the heatsink off, and check the date on the GPU...now then we know how long it was ready.. :)

http://www.xtremesystems.org/forums/showthread.php?t=72587