Overclock.net › Forums › Graphics Cards › AMD/ATI › [Official] AMD Radeon HD 7950/7970/7990 Owners Thread
New Posts  All Forums:Forum Nav:

[Official] AMD Radeon HD 7950/7970/7990 Owners Thread - Page 823

post #8221 of 38701
Sold my DCII, I'll hopefully have my Lightning by next week (stupid Easter weekend frown.gif). Then, with the Easter paycheck, I might be tempted to buy a second straight away wink.gif
MiniCaptein
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 7700K MSI Z270I Gaming Pro Carbon AC mITX EVGA GTX980Ti SC+ 2x8GB G.Skill Ripjaws V 3200MHz 16-16-16-36 
Hard DriveHard DriveHard DriveCooling
120GB Samsung 840 EVO 500GB Samsung Spinpoint F3 (Storage) 2TB WD Caviar Black (Games) Corsair H80i 
OSMonitorKeyboardPower
Win7 Home Premium 64bit Dell U2711 2560x1440 60Hz Das Keyboard 4 Cherry Brown Seasonic X-Series 750W 
CaseMouseMouse Pad
Corsair Obsidian 250D Razer Taipan Razer Goliathus Fragged Extended Speed Edition 
  hide details  
Reply
MiniCaptein
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 7700K MSI Z270I Gaming Pro Carbon AC mITX EVGA GTX980Ti SC+ 2x8GB G.Skill Ripjaws V 3200MHz 16-16-16-36 
Hard DriveHard DriveHard DriveCooling
120GB Samsung 840 EVO 500GB Samsung Spinpoint F3 (Storage) 2TB WD Caviar Black (Games) Corsair H80i 
OSMonitorKeyboardPower
Win7 Home Premium 64bit Dell U2711 2560x1440 60Hz Das Keyboard 4 Cherry Brown Seasonic X-Series 750W 
CaseMouseMouse Pad
Corsair Obsidian 250D Razer Taipan Razer Goliathus Fragged Extended Speed Edition 
  hide details  
Reply
post #8222 of 38701
Quote:
Originally Posted by tsm106 View Post

Quote:
Originally Posted by Ken1649 View Post

 

 

It doesn't work that way. If every cards in Crossfire to Quafire config has to wait for the other cards finish rendering for the other card to render next frame, we are not going to see any gain in FPS. If that's how they work, we all will be sticking to one card.

 

Crossfiring between different card of the same series


He's talking about the time to render frames between cards. Think about it for a sec cuz you've argued for this before in terms of cfx on differing lane widths and how 4x is bad juju. I believe its the same effect if you end up with cards with big gaps in performance.

Yeah. Also microstutter is a product of irregular intervals between frames, I imagine clocking cards differently would only make this problem worse.

edit: looking at that link though, maybe it wouldnt
Edited by Tslm - 4/2/12 at 1:01am
no noise
(9 items)
 
trinity protop
(6 items)
 
 
CPUGraphicsRAMHard Drive
AMD A6-4400M AMD Radeon 7520G 8GB Corsair Vengeance 1600 Samsung 830 128gb 
Optical DriveOS
DVD R/W Windows 7 
  hide details  
Reply
no noise
(9 items)
 
trinity protop
(6 items)
 
 
CPUGraphicsRAMHard Drive
AMD A6-4400M AMD Radeon 7520G 8GB Corsair Vengeance 1600 Samsung 830 128gb 
Optical DriveOS
DVD R/W Windows 7 
  hide details  
Reply
post #8223 of 38701
I am getting a Graphics score of P9438 in 3D Mark 11 Performance at clocks of 1199mhz ,1796mhz memory clock and a voltage of 1.275.

Is that good? Also m i5 2500K at 4.5ghz is at P8199 which brings the score total to P9097


Thanks
    
CPUMotherboardGraphicsRAM
i5 2500K @ 4.7Ghz 1.36v Asus P8P67 pro MSI GTX480 Lightning @920Mhz Corsair Vengeance 8GB 
Hard DriveOSMonitorPower
Pheonix Pro 128GB SSD + Samsung F3 Win7 64bit Dell U2412M Corsair HX750 
Case
Corsair 600T WE 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 2500K @ 4.7Ghz 1.36v Asus P8P67 pro MSI GTX480 Lightning @920Mhz Corsair Vengeance 8GB 
Hard DriveOSMonitorPower
Pheonix Pro 128GB SSD + Samsung F3 Win7 64bit Dell U2412M Corsair HX750 
Case
Corsair 600T WE 
  hide details  
Reply
post #8224 of 38701
Quote:
Originally Posted by Tslm View Post

Quote:
Originally Posted by tsm106 View Post
Quote:
Originally Posted by Ken1649 View Post

 

 

It doesn't work that way. If every cards in Crossfire to Quafire config has to wait for the other cards finish rendering for the other card to render next frame, we are not going to see any gain in FPS. If that's how they work, we all will be sticking to one card.

 

Crossfiring between different card of the same series


He's talking about the time to render frames between cards. Think about it for a sec cuz you've argued for this before in terms of cfx on differing lane widths and how 4x is bad juju. I believe its the same effect if you end up with cards with big gaps in performance.

Yeah. Also microstutter is a product of irregular intervals between frames, I imagine clocking cards differently would only make this problem worse.
 


OK if ya all say how it works then that's how it works biggrin.gif

ケンジ
(16 items)
 
  
CPUCPUMotherboardGraphics
i7-3960X i7-980X GA-X79-UD7 / GA-X58A-UD9 7970 + 465 
GraphicsRAMRAMHard Drive
2x 6990 + 580 16GB 2133 MHz 24GB 2000 MHz 120GB x 2 RAID0 
Hard DriveHard DriveCoolingCooling
120GB x 4 RAID0 Raptor 300GB x 2 RAID0 CPU-370 CPU-370 
CoolingCoolingOSPower
VID-AR699 EK FB X58A-UD9 Win 7 Ultimate x64 1500W / 1250W+1200W 
  hide details  
Reply
ケンジ
(16 items)
 
  
CPUCPUMotherboardGraphics
i7-3960X i7-980X GA-X79-UD7 / GA-X58A-UD9 7970 + 465 
GraphicsRAMRAMHard Drive
2x 6990 + 580 16GB 2133 MHz 24GB 2000 MHz 120GB x 2 RAID0 
Hard DriveHard DriveCoolingCooling
120GB x 4 RAID0 Raptor 300GB x 2 RAID0 CPU-370 CPU-370 
CoolingCoolingOSPower
VID-AR699 EK FB X58A-UD9 Win 7 Ultimate x64 1500W / 1250W+1200W 
  hide details  
Reply
post #8225 of 38701
Just finished my first stress test (no oc yet was getting random bsod (116 & 24)thought it may have been the 7970's installed a dozen drivers + old CCC and i'm not getting them any more smile.gif )
Get great temps
GPU1 41C 99% load
GPU2 40C 96% load

Under 'full' load for 1h smile.gif

How long would you run a stability test after a OC?
    
CPUMotherboardGraphicsGraphics
3930k w/ XSPC raystorm @ 4.8Ghz Rampage Extreme 4 w/ EK WB HD 7970 w/ EK WB Parallel HD 7970 w/ EK WB Parallel 
RAMHard DriveHard DriveHard Drive
32GB of Mushkin PC3-14900 1866mhz 9-10-9-27 Red... Seagate 1tb x2 raid 0 M4 256gb SSD 2tb WD Green 
Optical DriveCoolingCoolingCooling
Samsung External SE-208 Phobya v.2 480 x 3 Phobya v1.2 480  Dual XSPC top w/ 2 x DDC -1 pump w/ DIYINHK Pum... 
CoolingOSMonitorKeyboard
2 x AC AQ5 lt w/ WB's 64bit windows 7 pro Samsung 2443BW+ Sidewinder X6 
PowerCaseMouseAudio
Enermax Platimax 1200w PSU Case Labs TH10 DeathAdder 3500dpi AKG Q701 
AudioAudioAudio
ASUS Xonar One Dac Audioengine A5 Beyerdynamic MMX 300 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
3930k w/ XSPC raystorm @ 4.8Ghz Rampage Extreme 4 w/ EK WB HD 7970 w/ EK WB Parallel HD 7970 w/ EK WB Parallel 
RAMHard DriveHard DriveHard Drive
32GB of Mushkin PC3-14900 1866mhz 9-10-9-27 Red... Seagate 1tb x2 raid 0 M4 256gb SSD 2tb WD Green 
Optical DriveCoolingCoolingCooling
Samsung External SE-208 Phobya v.2 480 x 3 Phobya v1.2 480  Dual XSPC top w/ 2 x DDC -1 pump w/ DIYINHK Pum... 
CoolingOSMonitorKeyboard
2 x AC AQ5 lt w/ WB's 64bit windows 7 pro Samsung 2443BW+ Sidewinder X6 
PowerCaseMouseAudio
Enermax Platimax 1200w PSU Case Labs TH10 DeathAdder 3500dpi AKG Q701 
AudioAudioAudio
ASUS Xonar One Dac Audioengine A5 Beyerdynamic MMX 300 
  hide details  
Reply
post #8226 of 38701
Quote:
Originally Posted by 4514kaiser View Post

Just finished my first stress test (no oc yet was getting random bsod (116 & 24)thought it may have been the 7970's installed a dozen drivers + old CCC and i'm not getting them any more smile.gif )
Get great temps
GPU1 41C 99% load
GPU2 40C 96% load

Under 'full' load for 1h smile.gif

How long would you run a stability test after a OC?
 



Restore CPU + RAM to BIOS Optimized Defaults then give this a try post #28 biggrin.gif

 

 

ケンジ
(16 items)
 
  
CPUCPUMotherboardGraphics
i7-3960X i7-980X GA-X79-UD7 / GA-X58A-UD9 7970 + 465 
GraphicsRAMRAMHard Drive
2x 6990 + 580 16GB 2133 MHz 24GB 2000 MHz 120GB x 2 RAID0 
Hard DriveHard DriveCoolingCooling
120GB x 4 RAID0 Raptor 300GB x 2 RAID0 CPU-370 CPU-370 
CoolingCoolingOSPower
VID-AR699 EK FB X58A-UD9 Win 7 Ultimate x64 1500W / 1250W+1200W 
  hide details  
Reply
ケンジ
(16 items)
 
  
CPUCPUMotherboardGraphics
i7-3960X i7-980X GA-X79-UD7 / GA-X58A-UD9 7970 + 465 
GraphicsRAMRAMHard Drive
2x 6990 + 580 16GB 2133 MHz 24GB 2000 MHz 120GB x 2 RAID0 
Hard DriveHard DriveCoolingCooling
120GB x 4 RAID0 Raptor 300GB x 2 RAID0 CPU-370 CPU-370 
CoolingCoolingOSPower
VID-AR699 EK FB X58A-UD9 Win 7 Ultimate x64 1500W / 1250W+1200W 
  hide details  
Reply
post #8227 of 38701
Quote:
Originally Posted by m3t4lh34d View Post

Most of the time with BF3 you'll notice artifacting, such as random objects popping out of characters and general random shapes appearing on textures. If you're WAY beyond where you should be, you'll get a hard freeze or a sound loop, where you hear buzzing and there is no video and you have to either hard reboot or use CTRL+ALT+DEL, IF you still can at that point. If the driver has recovered.
Everyone has their own method. When you test each card individually they do not OC the same as they do with the added heat and stress of crossfire/tri/quad. As long as the cards are still stable at a higher voltage and you can keep the heat down during testing, there is no reason not to eliminate that variable to find the maximums in crossfire. You can always adjust down afterwards.
Sure, if you have a card that can do the same clock at a much lower voltage than the rest, you would want to keep it at or near that voltage, as long as its stable in crossfire. It doesn't hurt to write down the voltages and clocks of each card so that you can try similar settings in crossfire and reduce heat/power consumption. However for the beginner OC'er, there's no reason not to use my method as long as its stable and you can keep heat down, just to simply find the maximums of your cards.
Btw, my method has worked for me on air to attain 1300mhz in tri-fire, so I'm not talking out of my ass lol

I can see your point with regards to extra heat etc. But in my experience, cards in crossfire can clock higher than they might be stable in a single gpu config. I personally always test the cards individually before I play with crossfire.
Quote:
Originally Posted by 4514kaiser View Post

Just finished my first stress test (no oc yet was getting random bsod (116 & 24)thought it may have been the 7970's installed a dozen drivers + old CCC and i'm not getting them any more smile.gif )
Get great temps
GPU1 41C 99% load
GPU2 40C 96% load
Under 'full' load for 1h smile.gif
How long would you run a stability test after a OC?

I think you will find crysis and warhead will be the most stressful on the gpu temps wise. For stability, use crysis 2.
post #8228 of 38701
Tested two of my reference ASUS 7970s so far. Using stock shrouds

1st: ASIC 73.7%, does 1250/1900 @ 1.25 V / 1.75 V (GPU/Mem voltages)
2nd: ASIC 76.2%, does 1275/1850 @ 1.25 V / 1.75 V

paired with a 3930K @ 5G, cards in CF at 1200/1850, got 59773 in Vantage - a UK record for 2x GPU
time to push for 1250/1850 on both
Edited by borandi - 4/2/12 at 4:50am
post #8229 of 38701
Quote:
Originally Posted by borandi View Post

Tested two of my reference ASUS 7970s so far. Using stock shrouds
1st: ASIC 73.7%, does 1250/1900 @ 1.25 V / 1.75 V (GPU/Mem voltages)
2nd: ASIC 76.2%, does 1275/1850 @ 1.25 V / 1.75 V
paired with a 3930K @ 5G, cards in CF at 1200/1850, got 59773 in Vantage - a UK record for 2x GPU
time to push for 1250/1850 on both

Nice job man, 2 monstor overclocks on the memory there!
Both my cards do 1775 with a .25 bump. May have to revisit.
post #8230 of 38701
Quote:
Originally Posted by Newbie2009 View Post

Quote:
Originally Posted by borandi View Post

Tested two of my reference ASUS 7970s so far. Using stock shrouds
1st: ASIC 73.7%, does 1250/1900 @ 1.25 V / 1.75 V (GPU/Mem voltages)
2nd: ASIC 76.2%, does 1275/1850 @ 1.25 V / 1.75 V
paired with a 3930K @ 5G, cards in CF at 1200/1850, got 59773 in Vantage - a UK record for 2x GPU
time to push for 1250/1850 on both

Nice job man, 2 monstor overclocks on the memory there!
Both my cards do 1775 with a .25 bump. May have to revisit.

Still got two more cards to test wink.gif
Will also need to flash with the latest KPC ASUS BIOS to increase that vcore limit. Only 1.25 V max, should be able to dump 1.35 V through them for benchmarking, see if they'll budge on the MHz then
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › [Official] AMD Radeon HD 7950/7970/7990 Owners Thread