Yeah, 800W at 12 volts is nearly 67 ampres, which is an insane amount of power draw. How was that measured, exactly?
AMD: Volcanic Islands R1100/1200 (8***/9*** series) Speculation/ Rumour Thread
DuckThor Evil
Legend
800?
Gah, I assume liquid nitrogen cooling then or surely your voltage regs would blow to hell from such a load... Out of curiosity, what vcore was that at?
No, just a custom water cooling loop. It's EVGA 780Ti Classified Hydro Copper. It has quite a beefy PCB. I had it set a bit over 1.5v for the core on a software tool, but the actual voltage in the Ti-models has been reported to be higher than what the software says, so probably around 1.6v. I got two of these puppies.
Precision X power draw readings have been reported to be quite accurate. That's what I used. It's very easy to make my AX860 shutdown with stock CPU and only GPU OC when you crank up the voltages, with both OC'd I managed to shutdown an AX1200. Overclock.net has plenty of similar stories. These cards can zip some serious power when the voltages are cranked up.
Last edited by a moderator:
UniversalTruth
Veteran
Price?
I'm amazed the voltage regulator hardware on your vidcard could cope with that kind of ampearage (500A@1.6V!!!) You'd think those tiny little ICs would be reduced to smoking ruins in a flash and a bang, like that old dual-GPU board NV launched a bunch of years ago now. 
Still, I wouldn't expect your GPU to last very long at 1.6V, that can't be healthy at today's geometries... For curiosity's sake, how high did it clock at that voltage level?
Still, I wouldn't expect your GPU to last very long at 1.6V, that can't be healthy at today's geometries... For curiosity's sake, how high did it clock at that voltage level?
DuckThor Evil
Legend
I'm amazed the voltage regulator hardware on your vidcard could cope with that kind of ampearage (500A@1.6V!!!) You'd think those tiny little ICs would be reduced to smoking ruins in a flash and a bang, like that old dual-GPU board NV launched a bunch of years ago now.
Still, I wouldn't expect your GPU to last very long at 1.6V, that can't be healthy at today's geometries... For curiosity's sake, how high did it clock at that voltage level?
That particular card was only able to hit around 1390Mhz on the core. My second card passed Firestrike with 1441Mhz on the core at 1.35v in the software, but that's also probably closer to 1.45 or at least 1.4v in reality. I'm not done with that card yet as I've only run it with the AX860 so far and it begins to shut down at around 1.37v in the software if the CPU is also OC'd. I'll probably do some runs during the weekend.
I'm currently 38th in 1 card Firestrike hall of fame with 13451 points.
http://www.3dmark.com/hall-of-fame-2/fire+strike+3dmark+score+performance+preset/version+1.1/1+gpu
Perhaps these should be in some other thread though
I'm amazed the voltage regulator hardware on your vidcard could cope with that kind of ampearage (500A@1.6V!!!) You'd think those tiny little ICs would be reduced to smoking ruins in a flash and a bang, like that old dual-GPU board NV launched a bunch of years ago now.
Still, I wouldn't expect your GPU to last very long at 1.6V, that can't be healthy at today's geometries... For curiosity's sake, how high did it clock at that voltage level?
You should look more what overclockers done to them GPU's or CPU..... dont forget we are speaking about high rush run, not stability of the transistors ..
38th with just 1441mhz is a bit surprising grall... what tell Hwbot on this ? ( i have not look, and dont want loose time for this anyway ) ( i have check your score on3dmark, but dont do research )
Last edited by a moderator:
A1xLLcqAgt0qc2RyMz0y
Veteran
Breaking the 375 Watt PCIe Limit
Talk here is sure different now that AMD is breaking the 375 Watt Limit than when Nvidia was rumored to be releasing a card that was speculated to need more than 375 watts.
What changed, has there been some hardware change that now allows greater than 375 watts?
Talk here is sure different now that AMD is breaking the 375 Watt Limit than when Nvidia was rumored to be releasing a card that was speculated to need more than 375 watts.
What changed, has there been some hardware change that now allows greater than 375 watts?
A1xLLcqAgt0qc2RyMz0y
Veteran
Where in that preview are those cards that went "over the limit of officially allowed power through PCI Express powerplugs" because I don't see any mentioned in the article?
I didn't say they were mentioned in the article, but 6990 or 7990, can't remember which, had 375W and 450W BIOSes with only 375W powerplugs
First reviews are up for the 295x2 ..
http://www.guru3d.com/articles_pages/amd_radeon_r9_295x2_review,1.html
http://www.anandtech.com/show/7930/the-amd-radeon-r9-295x2-review
Im surprised, in good..
http://www.guru3d.com/articles_pages/amd_radeon_r9_295x2_review,1.html
http://www.anandtech.com/show/7930/the-amd-radeon-r9-295x2-review
Im surprised, in good..
Don't get me wrong, I'm a current AMD owner and have been since ... what, probably 2006? I moved away from NV's 7800GT and into a pair of crossfired 3870's, then a pair of 4850's, a single 5850, and currently a single 7970 OC edition.
But nothing in these benches is impressing me any more. I'm handwaving-off the general challenges of AFR methods, and just looking at the power, price and performance here. The 295x's frame pacing is near-always better than it has been with the GCN 1.0 cards, but it's continually worse than NV's offering. Absolute performance is certainly even with NV, but at a higher price and a higher power consumption rate. The only thing really going for AMD here is the watercooling, which keeps the noise quite further down than the competitors.
Nothing here strikes me as truly impressive. Not that it should, I suppose, it's CF-on-a-stick and I wouldn't buy an equivalent SLI-on-a-stick from NV either. Still, it's a LOT of power, a LOT of price, and the performance just isn't the same to me.
Guess it's time to go back to NV again after the long hiatus.
But nothing in these benches is impressing me any more. I'm handwaving-off the general challenges of AFR methods, and just looking at the power, price and performance here. The 295x's frame pacing is near-always better than it has been with the GCN 1.0 cards, but it's continually worse than NV's offering. Absolute performance is certainly even with NV, but at a higher price and a higher power consumption rate. The only thing really going for AMD here is the watercooling, which keeps the noise quite further down than the competitors.
Nothing here strikes me as truly impressive. Not that it should, I suppose, it's CF-on-a-stick and I wouldn't buy an equivalent SLI-on-a-stick from NV either. Still, it's a LOT of power, a LOT of price, and the performance just isn't the same to me.
Guess it's time to go back to NV again after the long hiatus.
UniversalTruth
Veteran
http://www.techpowerup.com/reviews/AMD/R9_295_X2/24.html
The fastest single card solution now?
What about the NVidia's Titan-Z?
Why 1500$ and not 1100-1200$?
The fastest single card solution now?
What about the NVidia's Titan-Z?
Why 1500$ and not 1100-1200$?
What about the Titan-Z? Only in very low resolution cases will the "Z" be faster than the 295x, unless you're talking compute performance with double precision math.
UniversalTruth
Veteran
It is missing from the review which is astonishing given the part of the market these products target.
At least it could give a better idea on the distribution of forces between both camps, now it just gives the wrong impression that there is no competition whatsoever.
Every article i have read do mention Titan Z in the conclusion,... but anyway we have no benchmark for compare it and this card will not be released untill the end of the month.
Anandtech... resume it enough well i think..
This said, maybe Nvidia will release really soon a 790 version in the same price range ..
It looks like the water pump and fan run at a raised level at idle, so the noise floor is higher at low power.
Coil noise is being commented on. Perhaps AMD's engineers and QA testers all have high-frequency hearing loss?
Coil noise is being commented on. Perhaps AMD's engineers and QA testers all have high-frequency hearing loss?