OCUK: 290X "Slightly faster than GTX 780"

[Tempered81]

http://www.techpowerup.com/191768/radeon-r9-290x-clock-speeds-surface-benchmarked.html

 

[VulgarDisplay]

800mhz core? This thing is likely going to be a water cooler's dream.

I also wonder if this is an early ES chip and boost isn't enabled?

 

[Gikaseixas]

Very low clocks for sure. Maybe it will be a overclocker's dream :biggrin:

 

[BoFox]

http://www.techpowerup.com/191768/radeon-r9-290x-clock-speeds-surface-benchmarked.html

Seems like Rage with 4xAA needs to be fixed by the drivers, perhaps..

 

[Saylick]

800mhz core? This thing is likely going to be a water cooler's wet dream.

I also wonder if this is an early ES chip and boost isn't enabled?

FTFY.

Hm, I recall reading of a revived AUSUM switch which ups the clock to the rumored 1020 Mhz. If so, it would explain the whole "ridicule Titan" rumor as well.

 

[Vesku]

That's pretty impressive if it's only running at 800MHz, still somewhat impressive if running at 1000MHz. Given the die size differences.

 

[Borealis7]

I get a gut feeling power consumption is going to be huge on the 290X, more so with an OC as the 79XX series behaves seeing as they are of the same architecture.

 

[3DVagabond]

I get a gut feeling power consumption is going to be huge on the 290X, more so with an OC as the 79XX series behaves seeing as they are of the same architecture.

Power consumption is relative to performance. If a 800MHz 290X can keep up with Titan and use less power, as the benchmarks we've seen show, it's =/> efficient than Titan. If it can run 1200MHz but uses heaps of power doing so, that's fine. It's performance will justify it.

 

[tviceman]

LOL. Oh yeah, that's the same thing.

Typically, there aren't any custom cards on release. To start with all of the chips are used to supply the reference designs. Your making a mountain out of a molehill.

Because FORESEEABLE future and not initially are completely in the same ball park. You're wrong end of discussion.

 

[Cloudfire777]

I can understand that Nvidia locked down Titan since they went allin with the design, vapour chamber, low noise etc.

But the 290X design looks pretty like all AMD reference cards. I cant understand why they would want to lock that down when there is so much better OEM designs than that.

 

[Borealis7]

to teach those insolent OEMs a lesson!
"you make our cards better, eh? let's see how you manage without our flagship."

no, can't think of a reason...

 

[Cloudfire777]

to teach those insolent OEMs a lesson!
"you make our cards better, eh? let's see how you manage without our flagship."

no, can't think of a reason...

What is the difference between forcing OEMs sell your GPU with a locked down design, and OEMs selling your GPU with a ton of different designs?

AMD make money on the chip itself, not the full card.

I think AMD would be better off allowing a broad variety of designs, to lure more people in. OEMs can sell more if they can market "military class design", better power phases, quieter and more efficient cooling" etc etc

 

[3DVagabond]

Because FORESEEABLE future and not initially are completely in the same ball park. You're wrong end of discussion.

There is no discussion with you, just disagreement. We'll see if they won't have custom cards, as you said.

Cheers

 

[BallaTheFeared]

Hopefully they ship a 8GB version, looks like the frame buffer is insufficient for the one game that rules them all.

At 1050 MHz (core) and 4.50 GHz (memory), it's advantage-AMD, looking at these graphs.


But I'm taking it with a grain of salt.

 

[Wag]

Am I regretting buying my 3 Titans now? We'll see... Although I have to say I was quite dissappointed with ATI's Crossfire support last time around, that's partly why I went Nvidia this time.

I am definitely going to pick up that 39" Asus 4k/60Hz monitor (or something comparable) when it hits though. At this point I really can't afford a $3500 4k display.

 

[tviceman]

800mhz core? This thing is likely going to be a water cooler's dream.

I also wonder if this is an early ES chip and boost isn't enabled?

If it's clocked this low with retail chips then it's a power guzzler and/or needs to run this low to keep it cool - which makes AMD's reference-cooler-only lockdown all the more strange.


I'm going to go ahead and think that it's not clocked at 800mhz.

 

[Elfear]

If it's clocked this low with retail chips then it's a power guzzler and/or needs to run this low to keep it cool - which makes AMD's reference-cooler-only lockdown all the more strange.


I'm going to go ahead and think that it's not clocked at 800mhz.

+1

I'm thinking it's closer to 1000Mhz. Titan performance at 800Mhz would be awesome though.

 

[Vesku]

Perhaps the reason for no spec release yet is AMD was deciding to take the power hit and clock it for best review benches or launch at a better point on the power curve but not look as great in review benches.

If so, I'd side with performance over power efficiency in that performance category. But I'm not in charge of AMD's GPU division.

 

[Teizo]

The article was updated to state the clocks were at 1050mhz with no dynamic overclocking feature.

 

[Vesku]

1000-1050 makes more sense, assuming the benches are real. A good showing for a ~430-450mm chip.

 

[blackened23]

The article was updated to state the clocks were at 1050mhz with no dynamic overclocking feature.

I'm guessing the R9-290X has a boost feature similar to what the GHz 7970 did. It probably automatically jumps from 800 to 1050mhz in 3d applications - that is what the 7970GE basically did, swap between 925 and 1050 in 3d applications. Of course that is not nearly as versatile as nvidia's boost - if it's like the GHZ boost, it basically switches between base and boost, with nothing in between (nvidia's boost is more flexible and dynamically changes per TDP and temperatures in 1 bin (13mhz) increments....) That said, the only thing that really matters is what the performance at maximum boost is, and the secondary effects from power consumption. The leaks indicated that power consumption is better than the Titan, hopefully that is indeed the case.

 

[Stuka87]

The article was updated to state the clocks were at 1050mhz with no dynamic overclocking feature.

I am thinking the dynamic clocking is disabled on early samples, which I am going to assume this is.

 

[ShadowOfMyself]

Sounds believable considering the massive memory bandwidth on the R9 290X
I expect completely different results in 1080/1440p though (favoring Titan)

 

[Teizo]

Sounds believable considering the massive memory bandwidth on the R9 290X
I expect completely different results in 1080/1440p though (favoring Titan)

Yes, maybe so. You have to understand they, reasonably so, want the R9 290X shown in the best possible light which is why they used such a high resolution (4k capable).

People were complaining about pre-ordering cards with BF4 with no benchmarks to draw from, so I doubt this leak was anything other than intentional.

 

[SiliconWars]

We're talking 20 GB/s or so higher bandwidth on the 290X (6-7% over Titan assuming 307 GB/s vs 288 GB/s), it's nothing that will make a huge difference at any resolution.