Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [tweaktown] AMD Radeon R9 Fury X2 should see 12 TFLOPS of compute performance
New Posts  All Forums:Forum Nav:

[tweaktown] AMD Radeon R9 Fury X2 should see 12 TFLOPS of compute performance - Page 6

post #51 of 80
Quote:
Originally Posted by Gilles3000 View Post

Remember the Titan Z? Yeah...

Nvidia has made it plenty clear that they can and will price stuff however they want. People will buy it anyway.

If it delivers a clear performance advantage over the Fury X2, people will definitely buy it...myself included.
New Gamer
(12 items)
 
Beater
(10 items)
 
 
CPUMotherboardGraphicsGraphics
Core i7-6950X ASUS Rampage 10 Edition  NVidia Titan X Pascal NVidia Titan X Pascal 
RAMHard DriveHard DriveHard Drive
8x16GB Corsair Dominator Platinum DDR4-2800MHz Intel P3700 2TB Samsung SM961 1TB 2x Samsung PM1725 SSD 6.4TB (11.6TB Windows Sof... 
CoolingOSPowerCase
Corsair H110 2x140mm AIO water cooler Windows 7 Ultimate Corsair AX1500i Corsair Vengeance C70 
CPUMotherboardGraphicsRAM
Xeon E5-1680 V3 Asrock X99E-itx/ac EVGA GeForce Titan X Superclocked 64GB (2x32GB) Crucial DDR4-2400 
Hard DriveOptical DriveCoolingOS
Samsung 950 Pro 512GB Panasonic Slim Blu-ray Asetek 92mm AIO (w/Delta PWM fans push/pull) Windows 10 Professional 
PowerCase
Corsair SF600 NCase M1 
CPUMotherboardGraphicsGraphics
2x Xeon E5-2687W V2 Asus Z9PE-D8 GeForce GTX Titan X (Maxwell) GeForce GTX Titan X (Maxwell) 
RAMHard DriveOptical DriveCooling
256GB Samsung PC3-14900R LRDIMMs Intel P3700 2.0TB (NVMe boot) ASUS Blu-ray  2x Intel Retail LGA2011-3 HSF 
OSPowerCase
Windows Server 2012 Standard R2 Corsair AX1500i 1500W Chenbro SR10769 
  hide details  
Reply
New Gamer
(12 items)
 
Beater
(10 items)
 
 
CPUMotherboardGraphicsGraphics
Core i7-6950X ASUS Rampage 10 Edition  NVidia Titan X Pascal NVidia Titan X Pascal 
RAMHard DriveHard DriveHard Drive
8x16GB Corsair Dominator Platinum DDR4-2800MHz Intel P3700 2TB Samsung SM961 1TB 2x Samsung PM1725 SSD 6.4TB (11.6TB Windows Sof... 
CoolingOSPowerCase
Corsair H110 2x140mm AIO water cooler Windows 7 Ultimate Corsair AX1500i Corsair Vengeance C70 
CPUMotherboardGraphicsRAM
Xeon E5-1680 V3 Asrock X99E-itx/ac EVGA GeForce Titan X Superclocked 64GB (2x32GB) Crucial DDR4-2400 
Hard DriveOptical DriveCoolingOS
Samsung 950 Pro 512GB Panasonic Slim Blu-ray Asetek 92mm AIO (w/Delta PWM fans push/pull) Windows 10 Professional 
PowerCase
Corsair SF600 NCase M1 
CPUMotherboardGraphicsGraphics
2x Xeon E5-2687W V2 Asus Z9PE-D8 GeForce GTX Titan X (Maxwell) GeForce GTX Titan X (Maxwell) 
RAMHard DriveOptical DriveCooling
256GB Samsung PC3-14900R LRDIMMs Intel P3700 2.0TB (NVMe boot) ASUS Blu-ray  2x Intel Retail LGA2011-3 HSF 
OSPowerCase
Windows Server 2012 Standard R2 Corsair AX1500i 1500W Chenbro SR10769 
  hide details  
Reply
post #52 of 80
Quote:
Originally Posted by lutjens View Post

If it delivers a clear performance advantage over the Fury X2, people will definitely buy it...myself included.

But not sure if sales figure will be worth developing a 2nd Titan Z.
post #53 of 80
Quote:
Originally Posted by Gilles3000 View Post

Remember the Titan Z? Yeah...

Nvidia has made it plenty clear that they can and will price stuff however they want. People will buy it anyway.

Titan z was a mega failure because it was a poorly marketed card and could not justify its asking price. It sold poorly and had to have a price slash to get the ball rolling.

Titan Z was a weird card. Titan Z under it's original purpose isn't a bad deal. It's a compute heavy card that offers more double precision than any card on the market that was fairly priced for pro cards

However, no one seriously took the titan moniker as a semi professional card seriously as it didn't come with the drivers that pro cards received. As a result, it was an over price card when used for gaming purposes put gamers could cough up 1k for a card. At 3k, there was no takers and it was a failure. Titan Z are generally rarely seen or owned scanning peoples signatures.

Any Titan xz at 3k is going to fail because it doesn't even have the professional cards dp going for it. At this point in time, max Nvidia can charge is 2k but a better price would be 1500 which is still pricey.

If AMD only releases this as a 12tf card. I wouldn't even bother countering with a double card again. If they want to steal AMD's thunder just launch a highly clocked gtx titan x with water cooling and 12gigs of 2ghz memory ddr5.

Cards like the water force or seahawk gtx 980 ti are in the same ballpark as a 295x2. Add in another 100mhz, use a full maxwell chip and 2ghz memory and should have a card within the same performance as a fury gemini. If NV charged 1200 dollars for such a card, it would likely do pretty well.
Edited by tajoh111 - 2/2/16 at 6:18pm
post #54 of 80
Quote:
Originally Posted by tajoh111 View Post

Titan z was a mega failure because it was a poorly marketed card and could not justify its asking price. It sold poorly and had to have a price slash to get the ball rolling.

Titan Z was a weird card. Titan Z under it's original purpose isn't a bad deal. It's a compute heavy card that offers more double precision than any card on the market that was fairly priced for pro cards

However, no one seriously took the titan moniker as a semi professional card seriously as it didn't come with the drivers that pro cards received. As a result, it was an over price card when used for gaming purposes put gamers could cough up 1k for a card. At 3k, there was no takers and it was a failure. Titan Z are generally rarely seen or owned scanning peoples signatures.

Any Titan xz at 3k is going to fail because it doesn't even have the professional cards dp going for it. At this point in time, max Nvidia can charge is 2k but a better price would be 1500 which is still pricey.

If AMD only releases this as a 12tf card. I wouldn't even bother countering with a double card again. If they want to steal AMD's thunder just launch a highly clocked gtx titan x with water cooling and 12gigs of 2ghz memory ddr5.

Cards like the water force or seahawk gtx 980 ti are in the same ballpark as a 295x2. Add in another 100mhz, use a full maxwell chip and 2ghz memory and should have a card within the same performance as a fury gemini. If NV charged 1200 dollars for such a card, it would likely do pretty well.

Talking about stealing AMD's thunder, I don't think Nvidia will bother cause its simply not worth the cost.
post #55 of 80
Quote:
Originally Posted by lutjens View Post

If it delivers a clear performance advantage over the Fury X2, people will definitely buy it...myself included.

Even if it's dual GM204? I believe that was what was rumored (along with dual GM200, either cut or full). Although I'm not sure dual GM204 would beat dual Fiji XT given that TPU put the Nano at 6% better @ WQHD and 11% better @ 4K UHD. But, GM204 uses less power so they could clock it a bit higher to make them perform equally (unless of course the Fury X2 is clocked lower, then I could see GM204 @ stock clocks matching it). Anyway, with dual GM204, I can see them match a Fury X2 in performance, power consumption, and price. Perhaps even beat them on price. It will be interesting to see the route they go, if they even pursue the option of creating a dual-GPU card at this stage.
post #56 of 80
A dual GM204 card would be utterly pointless when the 980 Ti exists. Assuming it uses the full GM204, it would probably end up around 25% faster than 980 Ti, but overclock worse with has less vram, and comes with all the pitfalls of a multi-GPU setup, likely for a higher price than the 980 Ti as well. If such a card was released, it would pretty much be DoA because it serves no purpose whatsoever.
post #57 of 80
Quote:
Originally Posted by magnek View Post

A dual GM204 card would be utterly pointless when the 980 Ti exists. Assuming it uses the full GM204, it would probably end up around 25% faster than 980 Ti, but overclock worse with has less vram, and comes with all the pitfalls of a multi-GPU setup, likely for a higher price than the 980 Ti as well. If such a card was released, it would pretty much be DoA because it serves no purpose whatsoever.
Yet would still sell by the truckload.
Lil' Roy Taylor
(11 items)
 
  
Reply
Lil' Roy Taylor
(11 items)
 
  
Reply
post #58 of 80
I doubt it. Even the diehard fanboys avoided Titan Z. Granted, they're not going to price a dual GM204 card at $3000, but even if it comes in at say $999, you'd really have to be very oblivious to facts to pick up such a turd. Then again...
post #59 of 80
Quote:
Originally Posted by magnek View Post

A dual GM204 card would be utterly pointless when the 980 Ti exists. Assuming it uses the full GM204, it would probably end up around 25% faster than 980 Ti, but overclock worse with has less vram, and comes with all the pitfalls of a multi-GPU setup, likely for a higher price than the 980 Ti as well. If such a card was released, it would pretty much be DoA because it serves no purpose whatsoever.

Why? IIRC, a single GTX 980 Ti performs close to SLI GTX 970s. GM204, both on the cut-down 970 and 980, can OC higher than GM200 (how often do you see a 980 Ti or Titan X @ 1600 core?). I do agree that such a card wouldn't be worth it over a 980 Ti since the price would almost definitely be higher and perform just slightly better + have the headaches of SLI. However, Nvidia could counter AMD's Fury X2 with dual GM204:





That was from Jan 22: http://www.techpowerup.com/reviews/Sapphire/R9_390_Nitro/23.html

The Oculus Rift's res is 2160x1200 or 2,592,000 total pixels. FHD is 2,073,600 total pixels and WQHD is 3,686,400 total pixels so it's a bit closer to FHD (1920x1080) than WQHD (2560x1440). I could definitely see a dual GM204 card countering the Fury X2 (Since Fiji XT gets destroyed by GM200 at lower resolutions). Hell, they could potentially use the cut-down GM204 used in the 970 and clock it a bit higher + price it even cheaper than the Fury X2. I bet it would use less power as well and match the performance of the card.

Edit: Keep in mind that there have been 970s as small as this that have been created:





Compared to a Fury Nano:



I can definitely see a dual GM204 card from Nvidia being close in size to the Fury X2. And we all know how much performance per inch matters these days. thumb.gif
Edited by xxdarkreap3rxx - 2/4/16 at 6:08am
post #60 of 80
Quote:
Originally Posted by xxdarkreap3rxx View Post

Why? IIRC, a single GTX 980 Ti performs close to SLI GTX 970s. GM204, both on the cut-down 970 and 980, can OC higher than GM200 (how often do you see a 980 Ti or Titan X @ 1600 core?). I do agree that such a card wouldn't be worth it over a 980 Ti since the price would almost definitely be higher and perform just slightly better + have the headaches of SLI. However, Nvidia could counter AMD's Fury X2 with dual GM204:

-Snip-

That was from Jan 22: http://www.techpowerup.com/reviews/Sapphire/R9_390_Nitro/23.html

The Oculus Rift's res is 2160x1200 or 2,592,000 total pixels. FHD is 2,073,600 total pixels and WQHD is 3,686,400 total pixels so it's a bit closer to FHD (1920x1080) than WQHD (2560x1440). I could definitely see a dual GM204 card countering the Fury X2 (Since Fiji XT gets destroyed by GM200 at lower resolutions). Hell, they could potentially use the cut-down GM204 used in the 970 and clock it a bit higher + price it even cheaper than the Fury X2. I bet it would use less power as well and match the performance of the card.

Edit: Keep in mind that there have been 970s as small as this that have been created:

-Snip-

Compared to a Fury Nano:

-Snip-

I can definitely see a dual GM204 card from Nvidia being close in size to the Fury X2. And we all know how much performance per inch matters these days. thumb.gif

I don't think their main problem with GTX 970 sli is the raw performance. It's probably the troubles that come with multi GPU setups in general, along-side questionable performance scaling and the 4GB partitioned VRAM.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Rumors and Unconfirmed Articles
Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [tweaktown] AMD Radeon R9 Fury X2 should see 12 TFLOPS of compute performance