Overclock.net › Forums › Graphics Cards › AMD/ATI › ATI Drivers and Overclocking Software › How to unlock Gigabyte HD 7970 OC voltage?
New Posts  All Forums:Forum Nav:

How to unlock Gigabyte HD 7970 OC voltage? - Page 3

post #21 of 26
Mathematical fact: clock for clock, 7950 is 0-14% slower than 7970.

Real life fact: clock for clock, 7950 is usually 3-6% slower than 7970.
The Air Tunnel
(10 items)
 
   
CPUMotherboardGraphicsRAM
i7-4820K (4500MHz@1.28V) P9X79 GTX 970 G1 1536/8002MHz @1.218V Team Group Vulcan 4x4GB 2133MHz 
Hard DriveCoolingOSPower
2x240GB SSD @RAID 0 Noctua NH-D14 SE2011 Windows 8.1 Pro Corsair AX750 
CaseAudio
Rosewill Armor EVO hiFace+AudioGD+DT770Pro 
  hide details  
Reply
The Air Tunnel
(10 items)
 
   
CPUMotherboardGraphicsRAM
i7-4820K (4500MHz@1.28V) P9X79 GTX 970 G1 1536/8002MHz @1.218V Team Group Vulcan 4x4GB 2133MHz 
Hard DriveCoolingOSPower
2x240GB SSD @RAID 0 Noctua NH-D14 SE2011 Windows 8.1 Pro Corsair AX750 
CaseAudio
Rosewill Armor EVO hiFace+AudioGD+DT770Pro 
  hide details  
Reply
post #22 of 26
Quote:
Originally Posted by twisted1 View Post

There's not even a reference to what driver they were using with the 7970, if they re-tested or just borrowed the stats from the initial test. And they're using diffrent settings, 2xMSAA, no MSAA on the other cards. And their performance graphs is just a mess, diffrent settings, clocks. A mess. This is how you make a graph, all tests done with the same driver( Catalyst 8.921.2.0), 10-15runs per test to get accurate numbers and a good mix of tests so no particular card gets any extra advantage.. So here you have the performance index, all stock clocks, game tests only, no benchmarks.
( excuse the non-english language in the graph )
3220?k=d439f775142ac8b3c94a70395e91171b
And again, the test rigg does not measure up.
The serious tests I've read, done by serious sites. On a decent test rigg( Intel Core i7 3960X @ +4,0 GHz ), with the same drivers for all cards simply land around 10% performance diffrence. That's all I'm saying, If the 7950's where in fact faster, trust me I would have a couple of those instead.
The biggest physical diffrence between Tahiti PRO/XT is the suffix(ie they're exactly the same). Only 4 compute units disabled sounds great, in practice however that means 1792 shaders vs 2048 shaders, about 15% less for the 7950. Under the same circumstances a card with 15% less raw sader compute perf. can not be faster, it defy's all logic and probably some physics too.
I have had a few cards through the years, atm. I only have two 7970's atm. tho, one xfx and the gigabyte. By tomorrow I'll hopefully have a third tho, (matrix :DD

They were using Catalyst 12.2. Your graph has no relevancy and you seriously think you need an i7-3960X to eliminate bottleneck? lachen.gif

A 7950 'can' be faster than its bigger brother. Look around the forum, do some research, educate yourself my friend. At the same clock frequencies, 14% less shaders realistically equate to a 3-5% performance difference between the two

The same can be said of the 670 and 680. There is a 3% real world performance difference between the two per clock.

Sorry to disappoint you further but here is a third review using an i7-3960X lmaosmiley.gif: (I would suggest you reading the entire thing thumb.gif)
http://www.xbitlabs.com/articles/graphics/display/sapphire-radeon-hd-7950-oc_9.html#sect0
Xbitlabs: "The gap is much smaller now: 3-4% on average at both resolutions. When clocked at the same frequencies as the HD 7970, the Radeon HD 7950 delivers almost the same performance..."

post #23 of 26
Quote:
Originally Posted by Bruennis View Post

A 7950 'can' be faster than its bigger brother. Look around the forum, do some research, educate yourself my friend. At the same clock frequencies, 14% less shaders realistically equate to a 3-5% performance difference between the two The same can be said of the 670 and 680. There is a 3% real world performance difference between the two per clock. "The gap is much smaller now: 3-4% on average at both resolutions. When clocked at the same frequencies as the HD 7970, the Radeon HD 7950 delivers almost the same performance..."

correct. compute benchmarks might show a larger difference at same clocks. But for games the difference is 3 - 5% at same clocks . thumb.gif
Fragbox
(14 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i5 2400 DH67BL AMD Radeon HD 6900 Series XFX HD 6950 2GB 
RAMRAMHard DriveOptical Drive
Corsair Corsair DDR3 1333 Mhz 2 GB Corsair DDR3 1333 Mhz 2 GB Western Digital Caviar Green SONY DVD-RW AD-7260S 
OSMonitorKeyboardPower
Windows 7 Professional 32 bit BENQ G2420HD Logitech K200 Seasonic VX550 psu 
CaseMouse
ANTEC 200 V2 Logitech mouse 
  hide details  
Reply
Fragbox
(14 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i5 2400 DH67BL AMD Radeon HD 6900 Series XFX HD 6950 2GB 
RAMRAMHard DriveOptical Drive
Corsair Corsair DDR3 1333 Mhz 2 GB Corsair DDR3 1333 Mhz 2 GB Western Digital Caviar Green SONY DVD-RW AD-7260S 
OSMonitorKeyboardPower
Windows 7 Professional 32 bit BENQ G2420HD Logitech K200 Seasonic VX550 psu 
CaseMouse
ANTEC 200 V2 Logitech mouse 
  hide details  
Reply
post #24 of 26
Quote:
Originally Posted by Bruennis View Post

They were using Catalyst 12.2. Your graph has no relevancy and you seriously think you need an i7-3960X to eliminate bottleneck? lachen.gif
A 7950 'can' be faster than its bigger brother. Look around the forum, do some research, educate yourself my friend. At the same clock frequencies, 14% less shaders realistically equate to a 3-5% performance difference between the two
The same can be said of the 670 and 680. There is a 3% real world performance difference between the two per clock.
Sorry to disappoint you further but here is a third review using an i7-3960X lmaosmiley.gif: (I would suggest you reading the entire thing thumb.gif)
http://www.xbitlabs.com/articles/graphics/display/sapphire-radeon-hd-7950-oc_9.html#sect0
Xbitlabs: "The gap is much smaller now: 3-4% on average at both resolutions. When clocked at the same frequencies as the HD 7970, the Radeon HD 7950 delivers almost the same performance..."

Using the same driver in all tests is highly relevant, if you use different drivers however it becomes totally irrelevant. The test in your link uses the Catalyst 11.12 from 01/25/2012 wink.gif

Yes, you need an extremely powerful system to absolutely 100% eliminate every possible bottleneck. Since the difference is so small, even the slightest random bottleneck or hickup will have an impact on the end results.

For regular gaming? No, of course not.

BTW. how's you'r Matrix doing? I'm in love wih mine, lol. Haven't quite figured out all of the voltages and stuff, there's so damn many things you can adjust. Anyways, got it semi-bench stable @1340Mhz. That's very semi-stable, wont do a heaven run.

I've also noticed that my power-supply is not up to the task, getting voltage drops and high ripple when i OC cpu @+5Ghz and the GPU at the same time.
Edited by twisted1 - 11/18/12 at 11:49am
post #25 of 26
So Back at the Ranch, lol..... Actually I too have a gigabyte 7970 OC which I just bought new at the FRY's down here the 1st of last month, thank you tax return! Anyways after comparing the differences between this and the ghz edition I decided to manually up the clocks to GHZ specs of 1100 Core / 1500 memory on the stock cooler that comes with it to see what happens. The good news is it worked great on the OC Guru II program but started to run 64C under load on the GPU. That's when I did some more research and found out the Arctic Accelero Extreme 7970 helps lowers the temps about 25C. So I got it and put the sucker in( after 2 days and all types of adhesives for those darn memory chips later!), The good news is card idles around 35C and loads around 48C, however the new problem is The Card freezes up and crashes the system even at 48C! I'm hoping I didn't ruin anything during the installation of this beastly cooler and heatsinks, figured if I did I wouldn't be able to start and boot into windows or even start up a game. I'm thinking it might even be the new 13.3 beta 2 drivers too. That's why I'm thinking about trying the bios flash to at least the ghz edition if that'll help keep the volts up for more room, I'm seriously hoping this works.
post #26 of 26
Ok so this might sound kinda stupid but................ I noticed on the card before I did any possibly permanent bios mods, I wanted to check to see which bios number we were running, looks like it was set to "2" , so what I did was turn it over to 1 and do a little crysis 3 real quick at 1100/ 1500 , from the looks of it not only did my temps get up to 50C and stay right around there, but there were no signs of crashing! well at the moment, usually it would do it withing 5 minutes of playing, now seems to run pretty good, don't know for sure yet so I'll do some more testing to make sure and post some more results. I guess the moral of this post is, maybe try a different bios on the switch just to see what happens....... maybe thumb.gif
New Posts  All Forums:Forum Nav:
  Return Home
Overclock.net › Forums › Graphics Cards › AMD/ATI › ATI Drivers and Overclocking Software › How to unlock Gigabyte HD 7970 OC voltage?