W1zzardOh btw, one of these is my card. Benching right now :D

Can't give any performance results before official launch.

The suspense is killing me! Hahaha

Looking forward to your comprehensive review sir :toast:

HumanSmokeThe full die specs are probably legit - just remains to be seen whether it is fully enabled out of the gate. I estimated that GM 200 was basically GM 204 + GM 206 resource-wise. I think Jen-Hsun stated 8 billion transistors (probably rounded to a whole number).
1.5 x performance might be on the cards for 4K, but any testing at 25x16/1440 and 19x10 should come in lower as the card becomes dependent upon CPU and memory subsystem.

I'm hoping they are not harvesting chips for a future fully enabled Titan, don't take me wrong, I love my Titans, but if I knew back then Titan Black would be released a few months after I would've saved my money for those instead.

Not making that mistake again :)

15th WarlockI'm hoping they are not harvesting chips for a future fully enabled Titan, don't take me wrong, I love my Titans, but if I knew back then Titan Black would be released a few months after I would've saved my money for those instead.
Not making that mistake again :)

Even if the Titan X is fully enabled, it might pay to hold off for a few months if history is any indicator. The Titan branding has been reference only, and without direct competition there is little pressure for Nvidia to allow AIB's free rein. If you're willing to sacrifice half the installed vRAM, I suspect the AIB shackles will come off with the GTX-numbered version when the 390X gets closer to reality. Custom VRM, dual 8-pin power, higher power limits (removing 6GB of vRAM should allow that ~30-50W to be diverted to the GPU at a minimum), and higher clocks, all in a cheaper package. If the 390X delivers on the promise, maybe Nvidia allow AIBs more leeway with voltage/power limits. although I suspect Nvidia value their well cultivated OEM markets too much to step outside the PCI-SIG guidelines for any reference card.

^ Very good advice.

However I'll use all of those GB for non gaming purpose so I guess I'll guinea pig.

Looks like a dual announcement at GTC judging by this dual Quadro M6000 (and $7K RR-X UHD video processing card) setup

Maybe Nvidia are ditching the SLI bridge like AMD did with CrossfireX starting with the 290 series

Caring1Maybe Nvidia are ditching the SLI bridge like AMD did with CrossfireX starting with the 290 series

Many compute tasks such as those utilised by Quadro don't leverage SLI.
AFAIK, NVLink- Nvidia's sort-of analogue of XDMA on steroids is at the moment compute orientated. If it can be integrated into the PCI-E 4.0 spec then it will arrive for consumers - if not, then it will probably remain a motherboard vendor chip addition option like PLX's bridge chips, or any other third-party controller.

great to play all your console ports!

Easy Rhinogreat to play all your console ports!

:mad:

If you want to play above 1080p at more than 30 FPS like you console gamers it's good but yeah, this card is probably going to be 2.5 times the price of a console. Pretty steep unless you need it for work too. :)

64K:mad:

If you want to play above 1080p at more than 30 FPS like you console gamers it's good but yeah, this card is probably going to be 2.5 times the price of a console. Pretty steep unless you need it for work too. :)

or 4k gamer with little extra $

arbiteror 4k gamer with little extra $

4K gaming for textures created for play at 1080p ? I know this is a little off topic but today's games are not meant to be viewed at 4K. Sure, you can play them at 4K but at that point you aren't getting better visuals. Higher resolution does not automatically equal better graphics.