[HardOCP] Asus DC II 290X max overclock versus GTX 780ti max overclocking review:

[3DVagabond]

The Hawaii is designed solely for gaming (and is worse than the Tahiti for compute) while a large portion of the GK110 die is designed for super computing and CUDA development. Those portions are disabled on the Geforce parts, so it isn't an apples to apples comparison.

Hawaii, on the other hand, is a completely gaming-centric part. There will not be a Hawaii variant designed for super computing, it's a gaming and only gaming card. That is why GK110 is a larger die.

But who cares, really? Who goes and buys a video card based on die size. LOL. That's pretty funny. I suspect consumers only care about the end result and metrics - performance, software, and features. Nobody buys a video card based solely on die size, most consumers don't know or care.

Hawaii is not a game centric part. The Radeon versions have crippled DP. The full Firepro version will have 1/2 DP.

We've also come to learn that AMD changed the double-precision rate from 1/4 to 1/8 on the R9 290X, yielding a maximum .7 TFLOPS. The FirePro version of this configuration will support full-speed (1/2 rate) DP compute, giving professional users an incentive to spring for Hawaii's professional implementation.

Certainly didn't scale back design wise from Tahiti.

Hawaii also employs eight revamped Asynchronous Compute Engines, responsible for scheduling real-time and background task to the CUs. Each ACE manages up to eight queues, totaling 64, and has access to L2 cache and shared memory. In contrast, Tahiti had two ACEs. The Kabini and Temash APUs we wrote about earlier this year come armed with four. Why is Hawaii so dramatically different? Some evidence exists to suggest that Hawaii’s asynchronous compute approach is heavily influenced by the PlayStation 4’s design, though AMD won't confirm this itself. Apparently, Sony’s engineers are looking forward to lots of compute-heavy effects in next-gen games, and dedicating more resources to arbitrating between compute and graphics allows for efficiencies that weren’t possible before.

The one metric you left out that's directly effected by size that consumers do care about is price.

 

[Elfear]

To be fair some of those tests are suspect, the R9 290X's don't like heat and may throttle down if you run them longer than just for benchmarks. The counterargument of course being that you can simply crank up the fan speed, but that can be loud depending on the card, case, cooling, ambient temps, etc. and how much voltage you are adding to achieve an overclock. So I have difficulty taking the 15-23% number at face value. ~15% I think is more accurate. But as someone said, this is splitting hairs, they are both very quick cards.

All of the 290X's in the reviews I linked were aftermarket cards so I'm assuming that throttling was under control. I know that ComputerBase warms their cards up before benching, not sure about the others.

Your own shootout had the 780 beating a water cooled 290, and let's not forget your Lightning was hamstrung.

1692 @ 1440 in Heaven?

I got 1638 @ 1254 in a cheese run.

3270 @ 1440 in Valley?

I got 3307 @ 1330...

I'm not trying to bag, but come on your 1300MHz Elpida Lightning beat your 1265MHz 290 and considering 1.4v I imagine only one of those was a 24/7 clock.

I mean what do you want me to say? Benchoff?

Take another look at the benchmarks. At 1600p the 1150Mhz 290 kept up very well with the 1340Mhz 780 with the exception of a couple games. Both air-cooled.

And you're going to assume the Lightning was gimped because the scores were lower than yours but not suspect the 290 that didn't scale much at all going from 1150-->1265Mhz in some games?

In all honestly, I will believe the results of the professional reviews over my own benchmarks. I did my best to make multiple runs and eliminate as much variability as I could but I'm no professional reviewer.

 

[BallaTheFeared]

In all honestly, I will believe the results of the professional reviews over my own benchmarks. I did my best to make multiple runs and eliminate as much variability as I could but I'm no professional reviewer.

I don't have a R290 to compare your results with, I do however have a 780 and can see some obvious conflicts. I have no reason to believe you deliberately botched the 780 results, nor do I feel your 4.7GHz Haswell was found wanting.. What I am saying is if it's not those, then what? Vram seems to make sense? Especially as you go to highlight 1600, which is more sensitive to bandwidth than 1080 is.

I'm fine with hinging my idea of performance on professional reviews, so long as they make sense. I don't see why you're so hung up on old reviews with sketchy results, we have newer ones of overclocked cards to consider as well.

 

[Nomanor]

Seriously man...? How in the heck can you compare a card that is for all intents $100-150 dollars cheaper. Not apples to apples.

It may be $100 cheaper, but it's also of lower quality.

- slower
- very bad temps
- very bad cooling
- very bad noise

I'd rather overpay $100 and get a freaking stellar product that 780Ti is.

 

[tulx]

It may be $100 cheaper, but it's also of lower quality.

- very bad temps
- very bad cooling
- very bad noise

I'd rather overpay $100 and get a freaking stellar product that 780Ti is.

I don't think anyone's trying to convince you to buy a reference R9 290/x anymore. The custom cooled cards are mostly quieter than a 780Ti and very well cooled.

 

[3DVagabond]

I don't think anyone's trying to convince you to buy a reference R9 290/x anymore. The custom cooled cards are mostly quieter than a 780Ti and very well cooled.

Yes, quieter and cooler, so that won't be important anymore. Average pricing really needs to come down, though.

 

[HurleyBird]

Yes, quieter and cooler, so that won't be important anymore. Average pricing really needs to come down, though.

I gotta agree... custom 290X would be a much better buy at MRSP, but with the current inflated prices 780Ti looks like the better deal.

 

[3DVagabond]

I gotta agree... custom 290X would be a much better buy at MRSP, but with the current inflated prices 780Ti looks like the better deal.

I don't think the 780 ti has ever been a good deal. The 780 though is looking a lot better until 290 (non X) gets back down to near $399.

 

[wand3r3r]

Given the discussion, I meant OC 290x.

Ti is only 10% faster than 780 c4c.

This review:

http://fcenter.ru/online/hardarticles

From the previous thread puts a R290x @ 1130 2% faster than the 780 @ 1162 at 1080 and 4% at 1600.

A 780 at the Ti clocks in the [H] review would have been just as quick as the R290x, and 780 after months now has proven to clock higher on reference, aftermarket, and soon most likely elite aftermarket because instead of 1300 780s it will be 1400 as you pointed out.

Not true, according to the largest base of OC's that exists. Keep dropping cherry picked and possibly unstable OC benchmark runs to promote your brand superiority though. I think this review is actually fairly close to the actual average given the very close results to the average in hundreds or even thousands of OCs in hwbot.

Show me a compilation of the 20 reviews with a 1300 much less 1400 clocks you go on about on the 780. It's as simple as that, you can't. Why you ignore a few golden 290/x's and pretend that golden 780's are normal is interesting.

The core is the same, it's physical limitations seem pretty similar to the 780 ti in the end.

GTX 780
1171/2052MHz
http://hwbot.org/hardware/videocard/geforce_gtx_780/

 

[wand3r3r]

The crazy thing here is that R9 290X OC'd has 25% higher load power consumption than GTX 780 Ti OC'd. And even though 2 out of the 3 games tested were AMD-sponsored games, the GTX 780 Ti OC still comes out on top in performance.

Maxwell will take efficiency to yet another level beyond this.

Uh, so the 390x will remain "the same" and not "take efficiency to yet another level"? Kind of a naive statement at this point in time.

You also clearly missed the overclockers review in this thread, check out the 780 ti consumption.

 

[ams23]

Uh, so the 390x will remain "the same" and not "take efficiency to yet another level"? Kind of a naive statement at this point in time.

LOL, naive? Maxwell is expected to be here much sooner than whatever island GPU you refer to. And yes, I expect Maxwell to be a significant improvement in power efficiency vs. the competition.

 

[Skurge]

LOL, naive? Maxwell is expected to be here much sooner than whatever island GPU you refer to. And yes, I expect Maxwell to be a significant improvement in power efficiency vs. the competition.

Based on?

 

[Grooveriding]

Based on?

Unicorns and pixie dust.

 

[wand3r3r]

LOL, naive? Maxwell is expected to be here much sooner than whatever island GPU you refer to. And yes, I expect Maxwell to be a significant improvement in power efficiency vs. the competition.

If it's not just a guess to promote company x and downplay company y, please provide a source. The statement implied you have knowledge about it, which clearly you just made up to promote a brand.

This thread is full of attempts to grasp upon all of a single reviews flaws that they can find. I actually find it representative of the avg oc vs avg oc, ironically very near hwbot's statistics. The cherry picking, goal-post shifting, and whining is pretty lame.

 

[el etro]

This fcenter.ru reviews is BS!

 

[wand3r3r]

What I'd really like to see is the 780 vs. 780 ti vs. 290x vs. 290 face off with the average oc statistics.

1187/2151MHz
http://hwbot.org/hardware/videocard/geforce_gtx_780_ti/

1171/2052MHz
http://hwbot.org/hardware/videocard/geforce_gtx_780/

1153/2091MHz
http://hwbot.org/hardware/videocard/radeon_r9_290x/

1150/1759MHz
http://hwbot.org/hardware/videocard/radeon_r9_290/

People here like to claim 1300+ is normal for the 780/ti, yet pretend like the 290/x don't overclock over the average too. Regardless, since everything is down to a silicon lottery I'd like to see these exact clocks to see what the average user may be able to get.

While I'm very interested in max OC's, I only take them in the interest of data, not to support the brand allegiance.

Again, this review was fairly close to the avg OC and I'd like to see the 780 and 290 tossed into the mix. [h], do your thing!

 

[ams23]

@ wand3r3r: Use some common sense. Maxwell rumors have started increasing in frequency. GTC 2014 is in March 2014. Maxwell will be the first NVIDIA architecture to really leverage the efficiency gains made with Kepler.M .

 

[ams23]

duplicate post

 

[wand3r3r]

@ wand3r3r: Use some common sense. Maxwell rumors have started increasing in frequency. GTC 2014 is in March 2014. Maxwell will be the first NVIDIA architecture to really leverage the efficiency gains made with Kepler.M .

Don't make up statements without data, especially when only to promote/demote company xyz (or put something to indicate you are simply guessing).

In the end, if they are still on 28nm, it's still to be seen what the benefits could be, clearly they have a better chance on a node drop vs. on the same node. I won't be surprised to see efficiency gains, however they could pull a hawaii on you and go for performance instead. :biggrin:

 

[Jaydip]

Yeah Maxwell should be very efficient because we have never seen such a large jump in performance in the same node before, 680->780Ti is indeed very impresssive.One thing bothers me though, maxwell will exist as a mobile gpu first so hopefully NV doesn't completely disregard the pure performance crowd.On a side note I hope they do justice to Maxwell, he is considered one of the best physicists for a reason

 

[Silverforce11]

What I'd really like to see is the 780 vs. 780 ti vs. 290x vs. 290 face off with the average oc statistics.

1187/2151MHz
http://hwbot.org/hardware/videocard/geforce_gtx_780_ti/

1171/2052MHz
http://hwbot.org/hardware/videocard/geforce_gtx_780/

1153/2091MHz
http://hwbot.org/hardware/videocard/radeon_r9_290x/

1150/1759MHz
http://hwbot.org/hardware/videocard/radeon_r9_290/

People here like to claim 1300+ is normal for the 780/ti, yet pretend like the 290/x don't overclock over the average too. Regardless, since everything is down to a silicon lottery I'd like to see these exact clocks to see what the average user may be able to get.

While I'm very interested in max OC's, I only take them in the interest of data, not to support the brand allegiance.

Again, this review was fairly close to the avg OC and I'd like to see the 780 and 290 tossed into the mix. [h], do your thing!

From all the reviews Ive read of late for 780ti and R290/X, it seems like it has about a 50mhz extra OC potential AVERAGE across multiple reviewers (reviewers and not forum users who go modded bios or suicide bench runs). Around 1.2ghz boost, whereas R290/X seem to average around 1.15ghz.

So hwbot core clocks seem spot on, but its vram clock is wacked. I havent seen a single R290/X that got above 1.8ghz vram OC, including forum OCers.

 

[ams23]

Don't make up statements without data, especially when only to promote/demote company xyz (or put something to indicate you are simply guessing).

Everything you just quoted in my most recent post (about increasing freq. of rumors, GTC 2014 dates, and Kepler.M) is fact. Like it or not, it is what it is.

 

[Silverforce11]

Yeah Maxwell should be very efficient because we have never seen such a large jump in performance in the same node before, 680->780Ti is indeed very impresssive.One thing bothers me though, maxwell will exist as a mobile gpu first so hopefully NV doesn't completely disregard the pure performance crowd.On a side note I hope they do justice to Maxwell, he is considered one of the best physicists for a reason

Not sure if serious, or you forgot a minor detail.. whats the die size of gk104 vs gk110 again?

 

[Jaydip]

Not sure if serious, or you forgot a minor detail.. whats the die size of gk104 vs gk110 again?

Unimportant metric to you and me except NV.

 

[Silverforce11]

Unimportant metric to you and me except NV.

Except it pertains to your wildly optimistic outlook.. you believe NV is going to get such a large % boost in die size again on the same node?

Lets see how much bigger they can inflate from GK110 with its >500mm2 die shall we..

I mean be free to be optimistic, but at least find another metric that better suits your outlook. A large jump in performance on the same node because of gk104 vs gk110 isnt a good measuring stick when there is such a massive die size difference.