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

[Orangey]

According to wccf it has had every type of RAM under the sun and been coming out every month of the year.

 

[Jono8]

It seemed leaker was wrong, 1080 do not have GDDR5X.

http://wccftech.com/nvidia-pascal-gp104-gpu-pictured-leaked/

1080 got K4G80325FB-HC25 memory chips on PCB.

I checked Samsung pdf below confirmed K4G80325FB-HC25 is GDDR5, not GDDR5X.

on.

That is not what the article says. The board pictured is the 1070 ( which is GP104-200-A1 apparantely) with normal gddr5 and K4G80325FB-HC25 chips . The leaker has said that the 1080 ( not pictured) has GDDR5x though.

Or maybe they have got it wrong and if a 1080Ti IS coming out on GP104 this summer perhaps the Gddr5x will be reserved for that and it will release a bit later than the 1070/1080. That would at least be something to seperate the Ti from the rest and make it more worth the inevitable premium price.

 

[D.P.]

It seemed leaker was wrong, 1080 do not have GDDR5X.

http://wccftech.com/nvidia-pascal-gp104-gpu-pictured-leaked/

1080 got K4G80325FB-HC25 memory chips on PCB.

I checked Samsung pdf below confirmed K4G80325FB-HC25 is GDDR5, not GDDR5X.

http://www.samsung.com/semiconductor/global/file/insight/2015/08/PSG2014_2H_FINAL-1.pdf



Compute Preemption look like Nvidia answer to AMD Async Compute.



Interesting thing about GP104 qualification sample with 1614A1 on it was manufactured on week 14 which was 2 weeks ago in early April. It was the final test sample before full volume production.

The image of the GPU on a PCB was of the GP104-200 part for 1070 which is supposed to be GDDR5. We don't have a photo of the 1080p mounted to a PCB to infer anything else.

 

[Gregster]

all I want to know, is when I can buy one

Check my sig, I'm running igp for heavens sake!

Heaven struggles on a IGPU

 

[drunkenmaster]

The image of the GPU on a PCB was of the GP104-200 part for 1070 which is supposed to be GDDR5. We don't have a photo of the 1080p mounted to a PCB to infer anything else.

If you're going to read wccfcrap and take notice of what they are saying, they quite clearly also stat 104-150 for the 1070 and 104-200 for the 1080.

 

[Harlequin]

NVIDIA GeForce GTX 980 GP104-200-A1 Reference and AIB Early June

 

[CAT-THE-FIFTH]

Compute Preemption look like Nvidia answer to AMD Async Compute.



Interesting thing about GP104 qualification sample with 1614A1 on it was manufactured on week 14 which was 2 weeks ago in early April. It was the final test sample before full volume production.

Nvidia upto now did draw called based pre-emption. AMD from GCN1.2 onwards has done fine grained based pre-emption which is hardware based.

According to wccf it has had every type of RAM under the sun and been coming out every month of the year.

 

[Seanspeed]

How would anybody know which is x70 or x80 or whatever based on those pics?

People are clearly just making assumptions. Unsafe ones.

 

[tyler_jrb]

the die shrink alone don't do much. Maybe some clock speed increase. It mostly counts because you can squeeze more transistors to a given place.

And what do those extra transistors do? You can fit more cores in the same die area.

They dont just apply a 40% die shrink so they can make the die 40% smaller than before. They apply it so they can fit 40% more transistors in the same/similar space. Hense more gpu power.

 

[CAT-THE-FIFTH]

And what do those extra transistors do? You can fit more cores in the same die area.

They dont just apply a 40% die shrink so they can make the die 40% smaller than before. They apply it so they can fit 40% more transistors in the same/similar space. Hense more gpu power.

Da shaders have got extra bling,innit?

 

[D.P.]

If you're going to read wccfcrap and take notice of what they are saying, they quite clearly also stat 104-150 for the 1070 and 104-200 for the 1080.

I wasn't, I was going by VideoCardz:

This GP104-200 variant is supposedly planned for GeForce GTX 1070

A Picture supposedly of the 1070 shows it uses GDDR5 memory, as rumored.

 

[DarrenM343]

Are we there yet?
Come on, I want to pick up some gaming soon, get yer new parts out

 

[Jono8]

Are we there yet?
Come on, I want to pick up some gaming soon, get yer new parts out

Probably got at least another month to go. Something is definitely on the horizon though. New leaks, pics and rumours seem to be coming out every day now.

I think the touted early June release is very likely.

 

[AthlonXP1800]

I wasn't, I was going by VideoCardz:




A Picture supposedly of the 1070 shows it uses GDDR5 memory, as rumored.

I don't know where the rumours about 1070 are GP104-200 came from but the guy who leaked that picture is somebody called Serandur from overclock.net forum. I checked Serandur posts but he never said 1070 will use GP104-200 SKU.

http://www.guru3d.com/news-story/new-pascal-gp104-die-photos-surface.html

Nvidia planned three Pascal SKUs but 1070 will use GP104-150 chip as Videocardz said over a week ago. It came from reliable HardwareBattle source.

http://videocardz.com/59009/rumor-nvidia-plans-three-gp104-skus-in-june

 

[kthula666]

Ever wonder how they problem solve chips.

 

[AthlonXP1800]

Here is first ever picture of either GTX 1080 Ti, 1080X or could be 1090 GP104-400 full fat chip with GDDR5X leaked that will replace 980 Ti.



http://wccftech.com/nvidia-pascal-gp104-gpu-gddr5x-leak/

These GDDR5X chips on GP104-400 PCB is odd, the part number 6GA77 Z9TXT was not listed on Micron site.

 

[shamus21]

GDDR5 / 256bit Bus and small die don’t look very promising for a mid range Pascal Gpu`s.

 

[Harlequin]

the first production micron GDDR5X is MT58K256M32JA-100 - which is stamped on it.

 

[muziqaz]

citation needed (and not info about the Tesla P100 as for all we know, that is completely unrelated to the geforce GP104 cards.)

Citation is common sense. Unless nvidia decided to throw money at 2 different pascal designs, I don't see info of Tesla pascal being unrelated to consumer parts.

citation needed

Also, few think the 1080 will be a "massive leap" but i think it is sensible to assume it will at least be a decent bit quicker (say 20% or so)

Nvidia are not going to release a flagship card over a year after the old one, with a whole new generation and name, a whole new arhchitecture and die shrink, for the same price with no increase in performance. It is commercial suicide, especially with AMD releasing new cards as well.

AMD looks like is releasing mid range 390 replacement without touching fury x. At least it seems so. I don't see why it would be commercial suicide for nvidia to do similar thing to replace high volume parts with new ones containing smaller fooprint in order to start generating some revenues.

Also, I am very curious of AMD Press conference the other day when they sounded so sure Polaris will be extremely competitive, and that they are in much better competitive position that they haven't been in years. Somehow that doesn't compute, unless they are just sweet talking the share holders

 

[Rroff]

Also, I am very curious of AMD Press conference the other day when they sounded so sure Polaris will be extremely competitive, and that they are in much better competitive position that they haven't been in years. Somehow that doesn't compute, unless they are just sweet talking the share holders

They always say stuff like that, always.