** The Official Nvidia GeForce 'Pascal' Thread - for general gossip and discussions **

[Mr Oblong]

I'm tempted by the 980Ti, but if big Pascal launches H1 2016 then I'm sure it's worth the wait.

Can't help thinking this will be the biggest performance for a while (once all products are out).

 

[shamus21]

it is all looking good for the die shrink, about time we got off the old tech.

 

[Orangey]

There is zero chance of H1 2016... Jen-Hsun himself said H2 for the first cards, almost certainly little boy Pascal or even a 750Ti-equivalent.

 

[Rroff]

On paper it could turn up towards the end of H1 (unexpected issues aside)... but the last 2 times TSMC made such claims about volume production it was infact delayed 6 months :S and already rumours of issues with larger dies.

EDIT: nVidia might be aiming towards larger chips first this time due to industry contracts.

 

[CAT-THE-FIFTH]

EDIT: nVidia might be aiming towards larger chips first this time due to industry contracts.

The GK110/GK210 would have been out for nearly 4 years by then,so it could happen,but then looking at the delay between the Tesla cards and the Titan,it could be a year end launch in 2016 for consumer large die cards,if things work out.

Is there are noise on whether the TSMC 16NM process and how its progressing??

 

[Rroff]

Not seen anything solid - rumours from the normal sources they are having issues with dies >150mm2 on finfets (apparently they are working but rapidly lose all power/thermal efficiency advantages over that size).

EDIT: This is in regard to TSMC 16nm FF and not specific to nVidia and/or a problem nVidia are having though I'd assume the same kind of issues would affect them.

 

[CAT-THE-FIFTH]

Not seen anything solid - rumours from the normal sources they are having issues with dies >150mm2 on finfets (apparently they are working but rapidly lose all power/thermal efficiency advantages over that size).

EDIT: This is in regard to TSMC 16nm FF and not specific to nVidia and/or a problem nVidia are having though I'd assume the same kind of issues would affect them.

I suppose,its still around a year or so until,so there is time.

 

[Orangey]

I wish they (everyone) had chosen some solution other than finfets... they have brung nothing but misery.

 

[Rroff]

I suppose,its still around a year or so until,so there is time.

I assume that is due to electrical leakage, etc. so likely implication being that they won't be able to hit particularly great clock speeds at first but then nVidia always did like a good refresh :S

 

[Nutella33]

Cant wait to see what it can do

Spoiler

 

[Gerard]

^ lol

 

[CAT-THE-FIFTH]

I assume that is due to electrical leakage, etc. so likely implication being that they won't be able to hit particularly great clock speeds at first but then nVidia always did like a good refresh :S

It could be a Kelper type situation when the smaller chip enters mass production and initial large die based cards are trickled out to commercial customers first?

 

[Boomstick777]

Cant wait to see what it can do

Spoiler

Haha brilliant

 

[Gregster]

They nicked my joke

 

[CAT-THE-FIFTH]

Haha brilliant

Its dem eyebrow hairs,LOL. Animated for 50% sterner looks!

 

[Orangey]

Next year instead of the moon landing they will put out a video explaining The Donald's hair.

 

[AthlonXP1800]

TSMC Begins Volume Production of 16nm FinFET Process – Nvidia Pascal GP100 GPU Among the Products in Production

Read more: http://wccftech.com/tsmc-begins-volume-production-16nm-finfet-nvidia-pascal-gp100-gpu/#ixzz3ijE0pi5H

Nvidia should receive GP100 engineering samples soon then consumer samples would be in volume production at end of Q4 2015 before launch in Q1 2016.

 

[Gregster]

Seeing what they have achieved with Maxwell, I am seriously looking forward to Pascal. I am expecting some massive performance increases.

 

[Jono8]

Indeed. I am skipping the 980Ti/Fury as i reckon even small Pascal will make them look slow in comparison.

 

[razorpakk]

If Pascal is such a revolution, Nvidia will hold it back to save some money on future R&D. There is no reason for Nvidia to push things too fast, not if AMD can't catch up.