nVidia has confirmed nv30-125m transistors

[RichL]

Originally posted by Wombat

<I>A selective abridged history:</I>
When the G400 came out, it relied a lot on the CPU to do some of the work that the WARP core couldn't do. Even Matrox didn't know where the ceiling was. Most people had around 600MHz Athlon/P3s at the time, and it wasn't until Greebe got his G400 paired with his unlocked Athlon (which ran at 850) did we see the G400 peak with around a 750-800MHz processor (Quake benchmarks, some apps required more or less).

Actually has anyone discovered the ceiling yet?
My G400max has gone from Duron@1025 to Athlon XP@1575 and my 3dMark2000 score just keeps going up by significant numbers.

Stuff the Parhelia, just gimmie a Cray to power my G400




My last post as a regular MURCer

[[Ch]amsalot]

Guys, here's more leaked specs on NV30, including a neat data flow diagram. The raw performance looks amazing.

http://www.airtrolinc.com/components/nv30.htm

It looks like NVIDIA will release four flavors, the NV-30-4-K, NV-30-2-K, NV-30-4-S, and NV-30-2-S.

I'm not sure if these are official specs-- it mentions "5 micron", which must be a mistake because it's on a .13 micron process. Hell, if it were on a 5 micron process the die would be as big as my desk!

-[Ch]ams

[GT98]

the Link is dead

[Wombat]

Rich, 3DMark is pretty much worthless. It uses things like RAM speed and amount, FSB speed, and processor, to determine your magic little score. Remember when people here managed to skyrocket their 3DMark scores by booting at 100MHz and using SoftFSB to run at 133MHz?

Run your G400 with actual games, things that have frame rates, and you'll see that your G400 plateaued hundreds of MHz ago.

[thop]

congrats RichL!

[RichL]

Actually I saw a significant jump in FPS for Dungeon Siege too. Was bottoming out at 9-10 FPS, now bottoms out at 12-15.

[WyWyWyWy]

Originally posted by [Ch]amsalot
Guys, here's more leaked specs on NV30, including a neat data flow diagram. The raw performance looks amazing.

http://www.airtrolinc.com/components/nv30.htm.

It looks like NVIDIA will release four flavors, the NV-30-4-K, NV-30-2-K, NV-30-4-S, and NV-30-2-S.

I'm not sure if these are official specs-- it mentions "5 micron", which must be a mistake because it's on a .13 micron process. Hell, if it were on a 5 micron process the die would be as big as my desk!

-[Ch]ams

LOL nice one mate! haha...
The most interesting bit is... it can be panel mounted!

[superfly]

Wombat....


The fact is that you can compare any card released in the last 18 moths(GF3,radeon 8500,etc...) run them at resolutions no higher than 1024*768 and you'll be hard pressed to tell any difference between those cards and the fastest card currently on the market in current games.....Even when the fastest cpu is used in either setup,so to me,that spells out either a cpu or I/O limitation....


At The current pace of evolution that video cards are going thru,which easily outpaces cpu speed improvements by far,we'll essentially be playing games at 1600*1200 32 bit with 0 percieved performance hit, in perhaps as little as 18 months from now,when cards will most likely be at least 3 times faster,since even at those resolutions,they'll still be either I/O or cpu bound,even with the fastest cpu available by then....

[BuddMan]

Originally posted by RichL
Actually I saw a significant jump in FPS for Dungeon Siege too. Was bottoming out at 9-10 FPS, now bottoms out at 12-15.

Well, Dungeon Siege is known to heavily rely on the CPU. Do you get any increases in any other games?

[Tempest]

For those that didn't already guess this one, this is from Beyond3D:

DrT: NVDA's earnings conference call is ongoing right now. The CEO just said NV30 productions shipments wouldn't occur until Jan 03.
---
duncan36: Actually he said it wouldnt ship in Q4 and would ship sometime in the Jan Quarter. A very important distinction, and a sly piece of spin by Nvidia.

[Wombat]

The fact is that you can compare any card released in the last 18 moths(GF3,radeon 8500,etc...) run them at resolutions no higher than 1024*768 and you'll be hard pressed to tell any difference between those cards and the fastest card currently on the market in current games.....Even when the fastest cpu is used in either setup,so to me,that spells out either a cpu or I/O limitation....

Well, what can I say, except that you're wrong? There's no correlation or causation there.

Besides that, how can you say that a "faster" video card requires more CPU power? Designing video cards that <B>require</B> CPU time as a crutch is by no means a faster card. A "faster" video card design will stand on its own.

[WyWyWyWy]

Originally posted by superfly
Wombat....


The fact is that you can compare any card released in the last 18 moths(GF3,radeon 8500,etc...) run them at resolutions no higher than 1024*768 and you'll be hard pressed to tell any difference between those cards and the fastest card currently on the market in current games.....Even when the fastest cpu is used in either setup,so to me,that spells out either a cpu or I/O limitation....


At The current pace of evolution that video cards are going thru,which easily outpaces cpu speed improvements by far,we'll essentially be playing games at 1600*1200 32 bit with 0 percieved performance hit, in perhaps as little as 18 months from now,when cards will most likely be at least 3 times faster,since even at those resolutions,they'll still be either I/O or cpu bound,even with the fastest cpu available by then....

I agree with you bu that's not entirely true.
3D scene (a scene in a game, including environment, characters, effects, etc) would only get more complex as entry-level PC advances and programmers/designers get more experience.

So I think, atm it is not CPU not catching up, it is the games. (but well they have to take entry-level PCs in consideration so they are not to blame.)

[superfly]

Originally posted by Wombat
Well, what can I say, except that you're wrong? There's no correlation or causation there.

Besides that, how can you say that a "faster" video card requires more CPU power? Designing video cards that <B>require</B> CPU time as a crutch is by no means a faster card. A "faster" video card design will stand on its own.

Well...For one,if you're using the fastest card and gradually increase the resolution all the way to 1600*1200 and don't notice any frame rate drops all all,you're obviously not video card limited as far as pxel/texel fillrate and bandwith,that's for sure....Which is what will happen soon enough...

Even more so considering that the card needs to dish out 4x the fillrate at 1600*1200 32 bit compared to say more modest resolutions,such as 1024*768 32 bit,in order to maintain a given fps value(say 60 fps in both cases)...

[superfly]

Originally posted by WyWyWyWy


I agree with you bu that's not entirely true.
3D scene (a scene in a game, including environment, characters, effects, etc) would only get more complex as entry-level PC advances and programmers/designers get more experience.

So I think, atm it is not CPU not catching up, it is the games. (but well they have to take entry-level PCs in consideration so they are not to blame.)

I agree with that completely,with the caveat that the graphics effects that developers should try to use mostly be video card dependant,as to stress all that extra fill rate/memory bandwith reserve,and not mostly I/O OR cpu bound effects..

[[Ch]amsalot]

The link to the secret NV30 specs didn't work because there was a period "." at the end. Here it is again:

http://www.airtrolinc.com/components/nv30.htm

-[Ch]ams