• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

GeForce GTX 1180, 1170 and 1160 coming in August. Prices inside

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.
hm, you just said a few hours ago that the 120w chip will be equal to a 1080 Ti, and that the GV106 will be the 1160.

You're all over the place now.

Predictions are fun, but maybe take a breath before so that you don't change twice a day.
 
crisium said:
hm, you just said a few hours ago that the 120w chip will be equal to a 1080 Ti, and that the GV106 will be the 1160.

You're all over the place now.

Predictions are fun, but maybe take a breath before so that you don't change twice a day.
Click to expand...

Please quote the message I stated the gtx1160 will be as fast as a gtx1080ti.

I thought I said gtx1080, I will correct it. I cant find it.
Thanks
 
Last edited:
happy medium said:
$749 ,180 watt will be the gtx1190, and will be 40% faster than a gtx1080ti.
$600 ,150 watt will be the gtx1180 and will be 20 % faster than a gtx1080ti.
$500 ,120 watt will be the gtx1170 and will be as fast as a gtx1080ti.

$350 , 100 watt GV106, gtx1160, will equal a gtx1080.





All three will be GV104 chips.
And later a GV106, gtx1160 chip.
Thats my guess, give or take 5%.
Click to expand...

happy medium said:
A gtx1060 is 120 watts correct?
12nm will give a 40% power reduction. The next card performing in the 1060 bracket should not use a 6 pin connector.

Wasn't a gtx1060 as fast as a gtx980? And it was a gp106 chip.

I believe the next 1160 will be a cut down GV104 chip. The gtx1160 will be good competition for a Vega 64, should be a gtx1070ti equal.

The next GV106 chip (1150ti) will be as fast as a gtx1060 or faster.
Click to expand...

happy medium said:
Please quote the message I stated the gtx1160 will be as fast as a gtx1080ti.

I thought I said gtx1080, I will correct it. I cant find it.
Thanks
Click to expand...
you misread his post. there's a comma in there that separates the clauses.

you did say GV106 [will be] the 1160, then later said you expect the 1160 to be a cut down GV104.


also, considering that the maxwell --> pascal move was a 2 node die shrink, whereas this seems to be a half node, expecting the same jump in performance (i.e. 980 ~ 1060) is probably a bit optimistic. though nvidia could just up the transistor budget and die size, these chips are plenty small.
 
Last edited:
ElFenix said:
you misread his post. there's a comma in there that separates the clauses.

you did say GV106 [will be] the 1160, then later said you expect the 1160 to be a cut down GV104.


also, considering that the maxwell --> pascal move was a 2 node die shrink, whereas this seems to be a half node, expecting the same jump in performance (i.e. 980 ~ 1060) is probably a bit optimistic. though nvidia could just up the transistor budget and die size, these chips are plenty small.
Click to expand...
Thank you, I see now....
I expect a large frequency jump, a small IPC increase, and slightly lower tdp per performance tier.
 
also, if you are expecting 40% power reduction for similar performance, you'd be hitting 150 watts for 1080Ti level (assuming the 1080Ti has adequate cooling), as it's a 250 watt card. maybe you did the fractional math backward?
 
ElFenix said:
also, if you are expecting 40% power reduction for similar performance, you'd be hitting 150 watts for 1080Ti level (assuming the 1080Ti has adequate cooling), as it's a 250 watt card. maybe you did the fractional math backward?
Click to expand...
12nm is a LPP , I expect a 40% reduction in power for the same performance as Pascal.
What they do with that 40%, I dont know exactly, but I'm sure performance will increase plenty.

The gtx1180 should be 150 watts and be faster than a gtx1080ti, unless they have to clock it through the roof,and lose efficiency.
 
happy medium said:
12nm is a LPP , I expect a 40% reduction in power for the same performance as Pascal.
What they do with that 40%, I dont know exactly, but I'm sure performance will increase plenty.

The gtx1180 should be 150 watts and be faster than a gtx1080ti, unless they have to clock it through the roof,and lose efficiency.
Click to expand...

So you're expecting a 40% power reduction AND a 20% performance improvement from a half node shrink?
 
Elfear said:
So you're expecting a 40% power reduction AND a 20% performance improvement from a half node shrink?
Click to expand...
yep
Check the gtx780 vs the gtx980 that were on the SAME 28NM node.
It's definitely possible.
67719.png
 
Last edited:
One thing I am wondering about is what is the performance of the 1170 is in comparison to the 970? A worthy upgrade?
 
PeterScott said:
FWIW: Latest Pricing rumor from WCCFTECH:
https://wccftech.com/nvidias-next-g...cifications-pricing-and-nomenclature-details/
  • The 120W NVIDIA next-gen Turing GPU will be priced around $499 MSRP.
  • The 150W NVIDIA next-gen Turing GPU will be priced around $599 MSRP.
  • The 180W NVIDIA next-gen Turing GPU will be priced around $699-749 MSRP.
Click to expand...

I think WCCF got some rumors about pricing, but probably invented everything else- like power consumption numbers, memory bus, etc. It would be interesting if someone could show how exactly could such cards be built and make sense. For example: if the memory bus is 352 bit, does that mean GDDR5(X)? Because if GDDR6 is used with such bus- then that is way too much bandwidth for a card that uses 180W. Because, unless it uses 7nm (which it will not)- GPU core will have ~135W after memory power usage, which is like 33% reduction in power consumption 14nm->12nm and it still has to be running higher clocks to be faster than 1080Ti.

Instead, what I'd expect are basically GP102 remade at 12nm, at 256-bit GDDR6 for similar bandwidth, higher clocks, and similar ~250W power use for GTX1180 and GTX1170. 10%-25%+ performance, and cheaper to make.
And maybe a second chip, at half the die, like most Gx106 are, with 6GB of slower GDDR6, and speeds similar to GTX1070/1070TI at ~150W.
 
Last edited:
TSMC 12FFN is only like 3% smaller or something close to that. It's not a real shrink. So I would expect it to be close to GP102 in terms of size, depending on the core count. Slightly less cores and two less memory controllers (albeit GDDR6) but more ROPs and TMUs.

If you remember back there were rumors that nVidia was close to launching new GPUs but then mining took off. I don't know how GDDR6 availability would have played into that but I suppose it's possible that the memory controller supports both GDDR5X and GDDR6.
 
No, I'm pretty sure Nvidia won't price the 1160 at $500 because prices never jump so much in one generation. You cannot shock the customers with prices, it has to be gradually done. I believe 1160 will start at no more than $349. 1170 however should be $500.
 
mohit9206 said:
No, I'm pretty sure Nvidia won't price the 1160 at $500 because prices never jump so much in one generation. You cannot shock the customers with prices, it has to be gradually done. I believe 1160 will start at no more than $349. 1170 however should be $500.
Click to expand...
Unfortunately for Nvidia I'm one those folks who thinks spending more then $250 to $300 for video card is simply a really silly thing to do.
 
mohit9206 said:
No, I'm pretty sure Nvidia won't price the 1160 at $500 because prices never jump so much in one generation. You cannot shock the customers with prices, it has to be gradually done. I believe 1160 will start at no more than $349. 1170 however should be $500.
Click to expand...
They already did that.But they renamed cards.They renamed GTX660TI and call it GTX680 and sell it for 500USD instead of 250USD.
Same with TITAN they basicaly renamed what supposed to by GTX680 and call it TITAN and price it at 1000USD instead of 500USD.
 
Head1985 said:
They already did that.But they renamed cards.They renamed GTX660TI and call it GTX680 and sell it for 500USD instead of 250USD.
Same with TITAN they basicaly renamed what supposed to by GTX680 and call it TITAN and price it at 1000USD instead of 500USD.
Click to expand...
BS. There was never a x60 card with a 256mb bit bus. Assuming they called the 680 a 660ti... that would have put AMDs graphics division out of business. AMDs flagship at the time was the 7970. So imagine that $250 660ti on par or beating the 7970. 🙄
 
Last edited:
amenx said:
BS. There was never a x60 card with a 256mb bit bus. Assuming they called the 680 a 660ti... that would have put AMDs graphics division out of business. AMDs flagship at the time was the 7970. So imagine that $250 660ti on par or beating the 7970. 🙄
Click to expand...
GTX 560 and the similar 460 had 256 bit. The 260 even had a 448 bit bus.
 
Head1985 said:
https://www.techpowerup.com/gpudb/287/geforce-gtx-560
https://www.techpowerup.com/gpudb/273/geforce-gtx-560-ti

They renamed GTX660TI to GTX680 because 7970 was underclocked so they saw opportunity how double prices and get away with it.Thats why they double margins in kepler era.
GTX680 was first x80 card with midrange DIE.To that day x80 card was always fastest single card from Nvidia with biggest DIE.
Click to expand...
Yea AMD did the same thing.instead of releasing the full fat fury x chip they gave us the 290x/390x mid range chip and then charged us $650 for the real high end Fury X.
They renamed the 290x the Fury X because they saw the gtx980 was underclocked so they saw a way to double prices too,



Sounds ridiculous ha?
And its way way off topic.
Make a new thread about it.( for the ninth time.)
 
Last edited:
maddie said:
GTX 560 and the similar 460 had 256 bit. The 260 even had a 448 bit bus.
Click to expand...
My bad, missed the way older cards. Speaking of kepler and beyond. Point still holds, you cant have a midrange card = competitions flagship performance at half the price. Unless you want to be accused of trying to put them out of business 😀.
 
VirtualLarry said:
How do these cards look for mining? Better / worse than a $200 RX 570 4GB (Newegg has a ShellShocker for one for $210 today, AR, so I figure that's the price they're heading to soon)?
Click to expand...

Nvidia is probably going to take the lead in every segment, including mining. This is when you realize with absolutely no doubt the competition is falling completely flat on its face.

If they use same width but with GDDR6, x60 should be able to achieve 35-40MH/s. Of course, I think the price will be higher in proportion to mining/gaming performance improvements.
 
Status
Not open for further replies.

TRENDING THREADS

Back
Top