[ BHTec ] Hawaii XTX GPU non existent - Continued

[KaRLiToS]

UPDATE

[BHTec] Hawaii XTX GPU non existent - Continued

[BHTec] Hawaii XTX GPU non existent - Continued

Source #2

I don’t know why this rumor persists, but there is absolutely no such product. For a modest sum, anyone could xray the die of a 290X and determine this for themselves.
- our source said.

______

Here’s the irrefutable truth: there is no Hawaii XTX. Not with 48 CUs, nor with higher clocks than Hawaii XT. Let me be very plain: Hawaii XTX does not exist in any way, shape or form. I am 100% certain of this, and no amount of counterfeit PowerPoint slides or Photoshopped die x-rays will ever make it true.

SOURCE

Breaking news: we have just confirmed that Hawaii XT is not the full chip. AMD is working on a new SKU with Hawaii XTX processor!

Hawaii XTX, the first GPU with 3000+ cores?
All those rumors, blurry die shot comparisons and hundreds of comments were correct (and apparently I was not).
Hawaii has more cores than advertised. Hawaii XTX variant was a well-kept secret, until now…

After 8 Pack’s revelations we started digging and asking, until we are able confirm that this information was true.

While the exact specs are unknown, we have a good reason to believe that XTX has 48 Compute Units. That’s one more in each Shader Engine (there are 4). The number of ROPs will likely be the same (64), but TMU count will increase to 192.

 

[Silverforce11]

Crazy stuff, didnt think they could cram that much onto that die.

Either way, now we know for sure whatever AMD's dave says on beyond3d is opposite of reality.

 

[KaRLiToS]

Either way, now we know for sure whatever AMD's dave says on beyond3d is opposite of reality.

True. :sneaky:

Instant credibility loss.

 

[Techhog]

What do you know? Where there's smoke, there really is fire. Or perhaps lava in this case.

 

[Skurge]

GTX480+ level power consumption?

 

[NTMBK]

Hrm. Story announcing "confirmation" on 23rd. It's the 25th now, and no other tech news site is talking about AMD's "confirmation". Still skeptical...

 

[Silverforce11]

GTX480+ level power consumption?

I have a hunch it wont be much different, comparing R290 vs R290X at similar clocks, they draw similar power even with the cut in shaders.

Kinda like the infamous 5830 situation http://www.anandtech.com/show/2947/12

It's a real shame AMD did this move because they got away with selling double harvested dies at premium halo prices... I guess they learnt something from NV. :/

 

[Face2Face]

Looking forward to another enthusiast videocard release, but it's kinda late in the game for this card, don't you think? The extra cores didn't help the R9 290X vs. the R9 290 much, I can't see anything different with this card vs. the R9 290X..

 

[SlickR12345]

Still don't think its true. Like what is their source or who confirmed it to then?

I think this is probably still false.

 

[linaaslt]

We consumers, are just a milking cow, and i don't blame AMD for doing this, why NV should get away with these kind of practices and AMD not. It's all business after all. If you don't need it don't buy it, wait till something better will appear. But it's up to each other.

 

[Erenhardt]

I wonder...
... there were r9-290 that unlocked to r9-290x when flashed with r9-290x bios. What if... It may be possible to unlock those to r9-295 (XTX), because those are not lasered off!

That would boost unlockable r9-290 value considerably! Someone wanting new high-end card should take this into consideration. There is quite some $ to save with a bit of gamble. If it doesn't unlock to 295X, you still have 290 unlocked to 290X for half of what you would have to pay for r9-295X. So tempting

 

[StereoPixel]

1/4 FP64 for Hawaii XTX is confirmed?

 

[Techhog]

Well, the question is then, how reliable has this site been in the past? If they're usually spot-on, it's probably true.

That said,I don't think AMD is really copying Nvidia with the rip-offs. It's obvious that Hawaii was rushed, so a full chip would need a better reference cooler. On top of that, they were having yield issues even with these harvested dies. If they don't change the reference cooler, I'll believe that they pulled a GK110. Otherwise, it doesn't really make sense to intentionally give up the performance crown. Well, unless this was all for hyping the 295X2 as two full chips, but even then the liquid cooling probably could have handled 2 full chips if the radiator were thicker, the temp limit were raised a bit, and they added a third 8-pin connector...

 

[Pinstripe]

Confirmed? Where's the official announcement from AMD then?

 

[tviceman]

GTX480+ level power consumption?

I have a hunch it wont be much different, comparing R290 vs R290X at similar clocks, they draw similar power even with the cut in shaders.

Kinda like the infamous 5830 situation http://www.anandtech.com/show/2947/12

It's a real shame AMD did this move because they got away with selling double harvested dies at premium halo prices... I guess they learnt something from NV. :/

Of course power consumption is going to go up. The 290x consumes more power than the 290 regardless of what mode it is running in. http://www.techpowerup.com/reviews/AMD/R9_290/24.html

The 5830 were crap dies with higher clocks than the 5850. Higher clocks = more watts and crap dies = leakage / more watts.

Enabling that 1 extra CU without bumping clocks will net in < 5% increase in real-world games. AMD will have to up the clocks to at least 1050 to gain any real appreciable difference in performance. The extra CU + higher clocks is going to push power consumption up.

 

[Techhog]

Of course power consumption is going to go up. The 290x consumes more power than the 290 regardless of what mode it is running in. http://www.techpowerup.com/reviews/AMD/R9_290/24.html

The 5830 were crap dies with higher clocks than the 5850. Higher clocks = more watts and crap dies = leakage / more watts.

Enabling that 1 extra CU without bumping clocks will net in < 5% increase in real-world games. AMD will have to up the clocks to at least 1050 to gain any real appreciable difference in performance. The extra CU + higher clocks is going to push power consumption up.

It's really a question of how much. Is GTX 480 a realistic possibility?

EDIT: Actullay, looking again, it's not just a possibility; it's a 100% sure thing. If this exists, it's a 350W card.

 

[MentalIlness]

http://www.tweaktown.com/news/38668...on-r9-295x-the-full-hawaii-xtx-gpu/index.html

It looks like AMD is preparing itself a new , the Hawaii XTX. Right now we have the Hawaii PRO, which is the Radeon R9 290, and the Hawaii XT, which arrives to us in the form of the high-end Radeon R9 290X.

The new Hawaii XTX will arrive in retail form as the Radeon R9 295X, and will feature the full 48 Compute Units - compared to the 40 on the R9 290 and 44 on the R9 290X. Better yet, there will be 3072 cores, this is an increase from the 2816 cores found on the already fast R9 290X. When it comes to TMUs, this will be increased from the 176 on the 290X to 192 on the 295X. The count will stay at 64.

The question now sits with pricing - will AMD be competitive, because these specs will allow the company to compete directly with the GeForce GTX TITAN , as well as the GTX 780 Ti. Because of the increased specs, we should expect the Radeon R9 295X to arrive with custom cooling from AMD's various partners in SAPPHIRE, MSI, and more.

 

[tviceman]

It's really a question of how much. Is GTX 480 a realistic possibility?

EDIT: Actullay, looking again, it's not just a possibility; it's a 100% sure thing. If this exists, it's a 350W card.

AMD is going to put a closed-water loop on the card just like the 295x2. That is my guess.

 

[Techhog]

AMD is going to put a closed-water loop on the card just like the 295x2. That is my guess.

Hm... I'm not too sure on that. I don't think AMD would want to screw over AIBs like that for a single GPU card unless they're targeting a ridiculous pricepoint like $1000. I think they'll just make an open air cooler, or they'll have more than one reference cooler.

 

[tviceman]

Hm... I'm not too sure on that. I don't think AMD would want to screw over AIBs like that for a single GPU card unless they're targeting a ridiculous pricepoint like $1000. I think they'll just make an open air cooler, or they'll have more than one reference cooler.

They won't be screwing over AIB's if the AIB's can sell it. An actual 300 watt card with a blower-style air cooling solution will be fairly expensive and still be inferior to a water cooler in all regards. It makes the most sense, given that core clocks are going to have to increase to realize any appreciable gains.

I could be wrong, but that's just my guess. AMD wasn't afraid to release a mini-nuclear reactor in the 295x2, and their re-released FX9590 comes with a water cooler, so why not put a closed water cooler and charge a small premium for a single-GPU card?

 

[netxzero64]

Oh, here comes the 95c XTX card. *clap clap*

 

[KaRLiToS]

Oh, here comes the 95c XTX card. *clap clap*

Yeah, I can't wait to see their solution to that issue.

Seeing that their new FX chip will be watercooled and the 295x2 was. Maybe they will add a watercooler and will have great overclocking capability.

Anyway, I don't expect much more out of Hawaii.

 

[dn7309]

Oh, here comes the 95c XTX card. *clap clap*

Doubt it, most likely will be release with aftermarket cooler from launch. Just like how the 280x was released.

 

[Silverforce11]

Of course power consumption is going to go up. The 290x consumes more power than the 290 regardless of what mode it is running in. http://www.techpowerup.com/reviews/AMD/R9_290/24.html

The 5830 were crap dies with higher clocks than the 5850. Higher clocks = more watts and crap dies = leakage / more watts.

Enabling that 1 extra CU without bumping clocks will net in < 5% increase in real-world games. AMD will have to up the clocks to at least 1050 to gain any real appreciable difference in performance. The extra CU + higher clocks is going to push power consumption up.

Notice I said at similar clocks. Reference cards have a different clockspeed. When I look at custom card reviews with R290 at 1ghz or 1050mhz, it draws similar power as R290X at those clockspeeds.

 

[HurleyBird]

GTX480+ level power consumption?

With luck it will be more like the GTX 480 to GTX 580 transition.