• 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.

NVIDIA GeForce GTX 880 and GTX 870 to Launch This Q4

Kaapstad

Kaapstad

Kaapstad

Man of Honour
Joined
21 May 2012
Posts
31,922
Location
Dalek flagship
NVIDIA is planning to launch its next high performance single-GPU graphics cards, the GeForce GTX 880 and GTX 870, no later than Q4-2014, in the neighborhood of October and November, according to a SweClockers report. The two will be based on the brand new "GM204" silicon, which most reports suggest, is based on the existing 28 nm silicon fab process. Delays by NVIDIA's principal foundry partner TSMC to implement its next-generation 20 nm process has reportedly forced the company to design a new breed of "Maxwell" based GPUs on the existing 28 nm process. The architecture's good showing with efficiency on the GeForce GTX 750 series probably gave NVIDIA hope. When 20 nm is finally smooth, it wouldn't surprise us if NVIDIA optically shrinks these chips to the new process, like it did to the G92 (from 65 nm to 55 nm). The GM204 chip is rumored to feature 3,200 CUDA cores, 200 TMUs, 32 ROPs, and a 256-bit wide GDDR5 memory interface. It succeeds the company's current workhorse chip, the GK104.
Click to expand...

http://www.techpowerup.com/202238/nvidia-geforce-gtx-880-and-gtx-870-to-launch-this-q4.html

Looking at the specs above anyone with a GTX 780 or better may as well keep their money in their pocket as I don't think these cards are anything special.
 
JediFragger said:
Let's hope they have some fast-ass memory to help alleviate that narrow bus-width :)
Click to expand...

Tons of extra L2 also helps really much.

Something I wrote on OCN today:

650Ti boost has:
-20% more cuda cores
-60% more TMUs
-50% more ROPs
-67% more memory bandwidth
than the 750Ti.

However the 650Ti boost is only ~9% faster than the 750Ti with the 750Ti consuming much, much less power. And since it clocks so well most of the 750Tis beat the ref 650Ti boost out of the box due to factory OCs.
Click to expand...

Maxwell is capable of overcoming bandwidth limitations.
 
I'm hoping GM204 springs a surprise in the performance (and power!) department :cool:

If the 750Ti is any indication it should be!! AMD best get their shooting boots on, quick sharp! :D
 
JediFragger said:
Let's hope they have some fast-ass memory to help alleviate that narrow bus-width :)
Click to expand...

According to Stulids review of the new 760 from ASUS, the new SK Hynix RAM is insanely fact.

But as I am more use to Afterburner I decided to stick with what I know. Using Afterburner I managed to get the Striker Platinum to a Boost speed of 1333MHz and a quite frankly ridiculous 8000MHz on the RAM! I actually ran out of slider in Afterburner and could not increase the RAM anymore. There where no artifacts or other geometry errors on screen at these settings during testing.
Click to expand...
 
AFAIR that was always the leak rumours, faster than current 780Ti but not flagship and I would think the 256bit bus gives that away.

The flagship card will need be a lot faster than 780Ti but all this was before the 20nm carry on with silicon maker.
 
The GTX650TI Boost 2GB is a salvage part,so is most likely use dies which cannot form functional GTX660 cards. TPU measurements indicates power consumption is close to a GTX660:

http://tpucdn.com/reviews/NVIDIA/GeForce_GTX_650_Ti_Boost/images/power_peak.gif

The L2 cache also takes up a reasonable amount of die space too in the GM107,especially considering the GM107 shares more texture units in the single SMX it uses,when compared to the GK106.

TBH,I don't really expect the GM204 to use massive amounts of cache,since it is going to be less power and cooling constrained. Plus the GTX680 was faster than the GTX580 despite having less memory bandwidth.
 
Last edited:
CAT-THE-FIFTH said:
TBH,I don't really expect the GM204 to use massive amounts of cache,since it is going to be less power and cooling constrained. Plus the GTX680 was faster than the GTX580 despite having less memory bandwidth.
Click to expand...

The GTX680 and 580 both have (without splitting hairs) 192GB/s memory bandwidth. The 680 was faster thanks to an increase in Cuda cores -1536SP @ 1018Mhz (nominal turbo) vs 512SP @ 1344Mhz & better efficiency.
 
_Alatar_ said:
Tons of extra L2 also helps really much.

Something I wrote on OCN today:



Maxwell is capable of overcoming bandwidth limitations.
Click to expand...
It's not so much of "overcoming bandwidth limitations", but more to do with the performance increase from architecture improvement managed to compensate for the low bandwidth (and had it had high memory bandwidth, the performance would had been much higher...even on a card such as 750Ti).

It's kinda let's say memory clock won't go any further beyond a certain frequency, but increasing the performance/frame rate by overclocking the core clock further would make the short-coming of the low memory bandwidth less apparent.
 
You must log in or register to reply here.
Back
Top Bottom