• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

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

Founders

Apologies if this has been suggested already, it’s a massive thread. (core binning not speed binning) Ignore now if I seem to repeat a previous post. If that’s the case I agree with them.

I think/hope the founder’s versions have the best silicon in regards to cores and that some are strategically disabled currently for power/heat locations and to be enabled later.

If team green are wanting to be a “more” end to end supplier we need a reason to buy direct from them and not third parties. So keep your options open, artificially limit core count for now to protect tesla and cover team red but later offer free upgrade to founders and a charged upgrade to others. With a disclaimer as others may have cheaped out on the implementation of the card.

Gives us few years to get used to change in the landscape.
As for an HBM version of the card. Think Micron5X for now and remember micron/Intel that worked on Xpoint.

Yes blue and green (sky and land) Intel/Nvidia all the way coming to AI/PC/Car/Console’s for the future.
 
Spectator said:
Apologies if this has been suggested already, it’s a massive thread. (core binning not speed binning) Ignore now if I seem to repeat a previous post. If that’s the case I agree with them.

I think/hope the founder’s versions have the best silicon in regards to cores and that some are strategically disabled currently for power/heat locations and to be enabled later.

If team green are wanting to be a “more” end to end supplier we need a reason to buy direct from them and not third parties. So keep your options open, artificially limit core count for now to protect tesla and cover team red but later offer free upgrade to founders and a charged upgrade to others. With a disclaimer as others may have cheaped out on the implementation of the card.

Gives us few years to get used to change in the landscape.
As for an HBM version of the card. Think Micron5X for now and remember micron/Intel that worked on Xpoint.

Yes blue and green (sky and land) Intel/Nvidia all the way coming to AI/PC/Car/Console’s for the future.
Click to expand...


Their not cherry picked just reference. Tbh I'm glad their doing it going reference DOES offer some benefits. Generally no coil whine for one & thus time round with vapour chamber? Possibly 67c at load with OC a decent cooler!!
 
Mauller said:
Yep, you can only do dual sli with pascal. And from preliminary information it appears to lack Async, but that is just from comparisons of DX11 to DX12 in ashes of the singularity.
Click to expand...

Do we know if duo is pascal limitation or just 1080/1070 or even just 1080 since i presume the 1070 uses the older style sli links still?
 
drunkenmaster said:
“The rumour has a context, it's not just the April date, it's "we think it's launching in April because 'stuff' ". If the stuff is wrong the rumour is wrong and the article is bogus, that doesn't mean April is wrong which is why I repeatedly explained that.”

““Yet here you are again attacking me for claims I wasn't making as you attacked me then for claims I wasn't making. It's hilarious how you wilfully misinterpret what I'm on this.”
Click to expand...
NVidia made it clear at the graphics shows in early 2016 that Pascal was due Q2 which is where the April number came from. As I said before some websites just used the start of Q2 April instead of the Q2 range. I tried to explain that to you and your response was “I literally provided proof of where these rumours came from and you're saying I'm wrong.” Well as it turned out you was wrong.

Another point you was wrong about was saying some of NVidia statements are about sampling. As I said ages ago. Not sampling, but availability Q2.

Don’t forgot you also pointed to some wrong news at coinspeaker.com Then when I said that news was wrong and “I would rather go with the news direction from NVidia over some 3rd party news site I have never heard of before.” You had a go at me and tried to make out the false news was direct from NVidia website even though it wasn’t.
 
There it is...

NVIDIA-GeForce-GTX-1080-3DMark-Performance-V2.png
 
You may or you man not like 3DMark benchmarks, but I think they are very good indicators of what GPUs are capable of. While those values barely ever translate into gaming performance, they serve as a base for architecture comparisons.

This time to simplify the chart we are not using overclocked and multi-gpu results. I think this way it’s more readable, especially since we added 5 different benchmarks into one chart.

All cards in this comparison are using stock settings. The variance for GTX 1080 clocks is between 1860 to 1886 MHz, probably caused by new GPU Boost 3.0 technology.

We are leaving the interpretation of this chart to you.
Click to expand...

If that chart is true it's exactly what I'd expect, die shrink mid range is about 20-30% faster than previous gen full fat.

Just like 580 to 680 when this event last happened.
 
straxusii said:
If that chart is true it's exactly what I'd expect, die shrink mid range is about 20-30% faster than previous gen full fat.

Just like 580 to 680 when this event last happened.
Click to expand...

Yeah, i think this card is going to be amazing for people on 960 downwards (like me)

Anything higher i dont think its going to be the performance jump people think
 
21% over Titan, 24% over Ti, 27% over FuryX
41% over 980 hardly 2xtimes
my guess 1070 will be trading blows with furyX rather than being better than a TitanX
 
Last edited:
AlamoX said:
21% over Titan, 24% over Ti, 27% over FuryX
41% over 980 hardly 2xtimes
my guess 1070 will be trading blows with furyX rather than being better than a TitanX
Click to expand...

More math fail, it's
70% faster than the 980!
32% faster than the 980ti
49% faster than a Fury.
 
You must log in or register to reply here.
Back
Top Bottom