NVIDIA: Beyond G80...

[INKster]

Because NVIO is where the SLI data transfer is performed.

Jawed

And you need two of them ?
Where's BR03, then ?

Surely you're not expecting it to have two PCI-Express x16 connectors too, are you ?

 

[Jawed]

NVIO has, as far as I remember, always been described as supporting the SLI connection (amongst other things) because that functionality isn't within G80 itself.

Jawed

 

[INKster]

NVIO has, as far as I remember, always been described as supporting the SLI connection (amongst other things) because that functionality isn't within G80 itself.

Jawed

How can you explain then, that it exists in G84/G86 since, obviously, they sport an SLI edge connector and do not have a NVIO chip ?
What has been generally accepted is that NVIO includes RAMDAC, TMDS and HDCP functionality because it was assumed that moving those to an external chip was beneficial to yields, due to G80's huge 90nm die, but no mention of SLI also being included was ever made.
At least not officially.

 

[Geo]

It was always my theory, and I think Demirug hinted as much as well, that NVIO would make it possible to need only one of those for a GX2 instead of duplicating the logic in both G80s.

 

[Jawed]

[...]but no mention of SLI also being included was ever made.

I'm off to bed. There's a big pile of stuff that isn't "official"...

http://www.beyond3d.com/content/reviews/1/3

You can also see dual SLI connectors that feed in to NVIO (pictured below), NVIDIA's brand new I/O processor for handling signal input and output to their next generation of GPUs. NVIO marshalls and is responsible for all data that enters the GPU over an interface that isn't PCI Express, and everything that outputs the chip that isn't going back to the host. In short, it's not only responsible for SLI, but the dual-link DVI outputs (HDCP-protected), all analogue output (component HDTV, VGA, etc) and input from external video sources.

Jawed

 

[INKster]

It was always my theory, and I think Demirug hinted as much as well, that NVIO would make it possible to need only one of those for a GX2 instead of duplicating the logic in both G80s.

Hence post #319...

 

[Geo]

Hence post #319...

My apologies for agreeing with you. If I ever feel the urge again, I will supress it at once.

 

[INKster]

G...98 ?
What happened to G88 ?

http://forums.vr-zone.com/showthread.php?t=138859

 

[neliz]

I was wondering the same thing. is nVidia just playing the number game? I doubt there would be any G90 (NV55) derivatives this early. thang.. the complete 50 product range hasn't even come out yet.

ven implying it's G90 derivative would have so many consequences, DX10.1 support etc...

 

[trinibwoy]

Yeah saw that too. G80 is supposedly late which would imply they were shooting for the refresh soon but most likely it's no big deal.

 

[Techno+]

new info http://www.fudzilla.com/index.php?option=com_content&task=view&id=295&Itemid=1

 

[Tim Murray]

"DX10.1 and SM4.1" are the new DX9.1.

 

[Geo]

"DX10.1 and SM4.1" are the new DX9.1.

Hey, didn't I say that somewhere already!?

 

[neliz]

"DX10.1 and SM4.1" are the new DX9.1.

I'd rather have DX4 running SM0.1!

 

[Blazkowicz]

yea, why was DX4 not released? presumably it was optimised for some bitboyz chip but there was a giant conspiracy involving NV, ATI, S3, powerVR etc. to make them fail.

 

[Mariner]

yea, why was DX4 not released? presumably it was optimised for some bitboyz chip but there was a giant conspiracy involving NV, ATI, S3, powerVR etc. to make them fail.

IIRC, at the time it was mentioned that it was because the number 4 is considered unlucky in Japan. Something to do with the word for 4 also sounding like the word 'death'.

 

[neliz]

yea, why was DX4 not released? presumably it was optimised for some bitboyz chip but there was a giant conspiracy involving NV, ATI, S3, powerVR etc. to make them fail.

After DirectX 3, we had planned a DirectX 4 for December 1996 that would allow access to some special features that Cirrus Logic was going to put into laptop video chips (I think, its been 9? years). When the chips got delayed, we opted not to ship DirectX 4 as it had us in a huge rush (3 months between 3 & 4) for no reason. We had also told the game developer community about Direct X 5 that was targeting summer of 1997, and so we decided to simply skip DirectX 4 rather than confuse people. DirectX 5 shipped on July 16, 1997 - and to this day, people ponder about what happened to DirectX 4. So much for avoiding confusion.

http://craig.theeislers.com/2006/02/directx_then_and_now_part_1.php

 

[neliz]

IIRC, at the time it was mentioned that it was because the number 4 is considered unlucky in Japan. Something to do with the word for 4 also sounding like the word 'death'.

That's Taiwan/China, but even the Pentium IV sold well there..

 

[INKster]

Back on topic.

NV's jumping directly from 90/80nm to 55nm ??
This means we'll probably have 55nm half-node GPU's in 2007/early 2008, which i find very surprising to say the least (considering the 80nm delays from both IHV's).

http://www.digitimes.com/bits_chips/a20070328PB202.html

 

[AlNom]

Back on topic.

NV's jumping directly from 90/80nm to 55nm ??
This means we'll probably have 55nm half-node GPU's in 2007/early 2008, which i find very surprising to say the least (considering the 80nm delays from both IHV's).

http://www.digitimes.com/bits_chips/a20070328PB202.html

Perhaps it's for the mobile sector that also uses eDRAM ?