Bookmark and Share


Nvidia Corp.'s next-generation Kepler family of graphics cards will be able to dynamically speed up themselves when it is most needed, according to unofficial information. The details are scarce, but it is clear that the technology will have a strong influence on the market.

For many years both Advanced Micro Devices and Nvidia Corp. pointed to overclocking as a way to substantially boost performance of their graphics processing units (GPUs). Overclocking naturally limits potential lifespan of graphics cards and while "true" enthusiasts hardly use them for more than a year, average end-users are cautious about overclocking. In a bid to offer a bit higher performance for everyone, Nvidia's code-named Kepler graphics solutions, which are supposed to become available in March or April will offer dynamic overclocking technology akin to Intel Corp.'s Turbo Boost and AMD's Turbo Core. web-site reports that Nvidia GeForce "Kepler" graphics cards will support Dynamic Clock Adjustment technology, which will overclock graphics boards automatically in cases when the speed is needed by 5% - 7%. While such speed boost is hardly truly significant, provided that it will be fully automatic, it will not only have influence on benchmark results, but will also boost user experience in some cases.

Numerous partners of Nvidia try to earn additional margins by squeezing maximum juices out of Nvidia GeForce-based graphics cards and sell factory-overclocked models. Potentially, DCA technology may either help them to further improve Kepler graphics cards or make it impossible to make pre-overclocked boards.

Nvidia did not comment on the news-story.

Tags: Nvidia, Kepler, Geforce, Tesla, Quadro, 28nm


Comments currently: 11
Discussion started: 03/08/12 02:51:20 PM
Latest comment: 03/11/12 06:53:12 AM
Expand all threads | Collapse all threads


If it has the ability to sense the framerate is dropping bellow a certain playable point and then boosted the speeds it would be effective. On the other hand, if this is similar to Intel's Turboboost it will only boost the speeds when the load is minimal, or when you first start the game / or start a graphically intense area. Unintelligently boosting the clock when the load is light, or at the begining of a hard area will be compleatly pointless, but make benchmarks look better.

I will have to wait and see how it senses when to boost, but i have a strong feeling it wont be at the "right" times.
0 0 [Posted by: cashkennedy  | Date: 03/08/12 02:51:20 PM]
- collapse thread

Well technically speaking Intel's TurboBoost at least guarantees 1 full bin even when all the cores + HT are active. So if it's anything like Intel's Turbo, that's free 5-10% automatically. Rumor has it the GPU core will scale from 705mhz all the way to 950mhz when the GPU is at 100% load. There will also be the introduction of a 300MHz low-power state.

The full-fledged flagship, GK110, won't be out until the summer, however.
0 0 [Posted by: BestJinjo  | Date: 03/08/12 03:36:07 PM]

I know intel turbo boost only overclocks will overclock when not all cores are active

on a gpu all cores are active all the time

or is there something I didn't understand
0 1 [Posted by: madooo12  | Date: 03/08/12 10:04:19 PM]
- collapse thread

"I know intel turbo boost only overclocks will overclock when not all cores are active"

100% not true.

Intel's Turbo Boost works even if all cores are 100% loaded. The only difference is how much of a boost the CPU gets. This depends on how many cores are loaded.

With 1155 Intel's i5/i7 CPUs, you always get at least 1 bin even when all the cores are loaded.

I.e., 2500k 3.3ghz --> 3.4ghz.

1 bin is guaranteed at minimum, and up to 4 bins with 1 core (in this example up to 3.7ghz).
0 0 [Posted by: BestJinjo  | Date: 03/09/12 06:24:08 AM]

Are these yet another way to keep overburning chip at moderately low consumption. Should we happily embrace yet another Fermination series of GPUs. Just this time envy doesn't market it as the hottest and fastest card on Earth but vaguely describe same feature they share with their older sisters (GF100/GF110 based cards) as *dynamic overclocking feature* aka. When you need to heat up your place you could dynamically overclock your fancy new glittery envydio card.
0 1 [Posted by: OmegaHuman  | Date: 03/08/12 10:48:40 PM]
- collapse thread

You are not understanding what TDP is.

Intel's CPUs are rated at 95W TDP. It doesn't mean they operate at 95W at stock speeds. In fact, we know they don't - look at any review. That's why Intel is able to give you "free" TurboBoost and still stay under 95W TDP even with Turbo applied when all 4 cores are loaded. At stock speeds, the power consumption never approaches 95W for the CPU cores alone. That's the entire TDP allocated for both the GPU and CPU cores. This gives Intel room to stay within TDP parameters with Turbo.

You can rate a card at 250-300W TDP. That just helps OEMs and AIBs to figure out what PSU and cooling they need. For example, HD7970 is rated at 250W TDP but only consumes about 190-200W of actual power at stock speeds.

In this case, NV automatically overclocks the card to its maximum rated TDP limits. This is no different than buying an HD7970 and manually overclocking it to its 250W TDP limit. The difference is, NVidia is doing that for you automatically. This is similar to how AMD and Intel give you "free" overclocking within TDP limits on modern CPUs. All it means is less of a performance gain from manual overclocking since some of the overclocking was already given away by Intel/AMD through TurboBoost.

Also, if you have a problem with 250-300W high-end GPUs, then they just aren't for you. HD7870 line is perfectly suited for someone who wants a 175W GPU.

Most enthusiasts who spend $500+ on a GPU want maximum performance. They could care less if it's a 250-300W part as long as it lays waste to previous high-end.
0 1 [Posted by: BestJinjo  | Date: 03/09/12 06:33:44 AM]

Dynamic overclocking on GPU ? Since modern day GPU already throttles between low and max clocks, then what is this "Dynamic overclocking"? For CPU Turbo clocks works and its effective when not all cores are using. But while gaming GPU is always pushed to the max. Its better to say simply Higher Stock Clocks isn't ?
Nvidia's marketing trick.................
0 1 [Posted by: tks  | Date: 03/09/12 04:53:58 AM]
- collapse thread

Wow, there is so much misunderstanding here.

So AMD's and Intel's CPU TurboBoosts are also marketing?

This is exactly the same. Read my post above. It explains it. TDP rating and real world power consumption are not the same. Intel and AMD give you extra CPU boost to work within the TDP limits. This is how the industry has operated in CPU space since at least November 2008 when Core i7 920 launched.

Instead of embracing it on GPUs, people are calling this a marketing trick. Fanboism at its finest.
0 2 [Posted by: BestJinjo  | Date: 03/09/12 06:41:35 AM]
CPU max out within few minutes while GPU max out while you play and that could be hours or whole day, what will happen? Your GPU will brow considering the limitation of GPU cooling compared to CPU cooling.
0 0 [Posted by: pogsnet  | Date: 03/11/12 06:53:12 AM]

This is going to give Nvidia an unfair advantage in relation to benchmarks. I hope there's a way to turn it off so that costumers can truly see how they compare to AMD's lineup.

...If the TDP is 225W, what do you guys think will happen when more performance is needed and the card is already consuming 225W?
0 0 [Posted by: rwwot  | Date: 03/09/12 07:18:43 AM]

This does not make sense. GPU cant clock different cores independently. Besides if games will load 100% (almost) all the time then the GPU will blow. Unlike in CPU most calculation done at loading only then the GPU kicks in.
0 0 [Posted by: pogsnet  | Date: 03/09/12 04:13:01 PM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture