Overclock.net › Forums › Graphics Cards › AMD/ATI › ATI Drivers and Overclocking Software › *Puts on shades* 5850 overclocking time, advice?
New Posts  All Forums:Forum Nav:

*Puts on shades* 5850 overclocking time, advice? - Page 3

post #21 of 33
don't forget source games, they will help with stability testing too.
    
CPUMotherboardGraphicsGraphics
i7 5930k MSI X99A SLI Krait Edition MSI TF III 7950 MSI TF III 7950 
RAMHard DriveHard DriveHard Drive
Rip Jaws 16GB @2400MHz Samsung 850 Evo SSD Seagate 250GB Hitachi 1TB 
Optical DriveCoolingOSMonitor
El-generico Cooler Master Hyper 212X 7 ultimate 64bit Tripe Viewsonic 22", 5760x1080 eyefinity secks. 
KeyboardPowerCaseMouse
Logitech G15 HX750i Phanteks Enthoo Pro Logitech G402 
Mouse Pad
ಠ_ಠSolid oak deskಠ_ಠ 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
i7 5930k MSI X99A SLI Krait Edition MSI TF III 7950 MSI TF III 7950 
RAMHard DriveHard DriveHard Drive
Rip Jaws 16GB @2400MHz Samsung 850 Evo SSD Seagate 250GB Hitachi 1TB 
Optical DriveCoolingOSMonitor
El-generico Cooler Master Hyper 212X 7 ultimate 64bit Tripe Viewsonic 22", 5760x1080 eyefinity secks. 
KeyboardPowerCaseMouse
Logitech G15 HX750i Phanteks Enthoo Pro Logitech G402 
Mouse Pad
ಠ_ಠSolid oak deskಠ_ಠ 
  hide details  
Reply
post #22 of 33
Thread Starter 
Quote:
Originally Posted by Kratthew D. Kraids View Post
don't forget source games, they will help with stability testing too.
Of course! TF2/DoD Source are one of the many games I love to play the most
post #23 of 33
You don't really need to run Furmark for 15 mins for each interval IMO, but that's up to you.

You'll see artifacts once you've pushed core clocks too high (w/o increasing voltage), but you won't see any errors with the memory; it will correct itself and by doing so, decrease performance.
What you'll need to look for is either stagnating performance or even decreases; from there you'll want to knock it down a bit and try to find the happy medium
Stugots III
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K @ 4.5GHz ASUS Maximus V Gene EVGA GTX 1060 SC 2x8GB G.Skill 2133MHz 
Hard DriveOptical DriveCoolingOS
250GB 850-EVO | 1TB EZEX | 2TB F4 It BURNS! H100i | NB Pros Windows 10 Pro x64 
MonitorKeyboardPowerCase
ASUS VE258Q Filco Majestouch I Seasonic X-560w Fractal Arc Mini R2 
MouseMouse PadAudioAudio
Logitech G502 Vengeance MM200 HT Omega eClaro SHURE SRH-840 | Audioengine A2+S8 
  hide details  
Reply
Stugots III
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K @ 4.5GHz ASUS Maximus V Gene EVGA GTX 1060 SC 2x8GB G.Skill 2133MHz 
Hard DriveOptical DriveCoolingOS
250GB 850-EVO | 1TB EZEX | 2TB F4 It BURNS! H100i | NB Pros Windows 10 Pro x64 
MonitorKeyboardPowerCase
ASUS VE258Q Filco Majestouch I Seasonic X-560w Fractal Arc Mini R2 
MouseMouse PadAudioAudio
Logitech G502 Vengeance MM200 HT Omega eClaro SHURE SRH-840 | Audioengine A2+S8 
  hide details  
Reply
post #24 of 33
Thread Starter 
Quote:
Originally Posted by candy_van View Post
You don't really need to run Furmark for 15 mins for each interval IMO, but that's up to you.

You'll see artifacts once you've pushed core clocks too high (w/o increasing voltage), but you won't see any errors with the memory; it will correct itself and by doing so, decrease performance.
What you'll need to look for is either stagnating performance or even decreases; from there you'll want to knock it down a bit and try to find the happy medium
Sounds good. I'm going to try for 1250 memory, see what happens. If I get no errors/artifacts I'll stop right there for memory and start doing the core.

Then the gaming/tweaking begins

Quick question!: Is it okay to have other windows over a stability test such as Furmark? Or does it need to be the foreground and uninterrupted by other tasks?
post #25 of 33
I don't think it matters too much, but I would leave it alone running fullscreen
    
CPUMotherboardGraphicsGraphics
i7 5930k MSI X99A SLI Krait Edition MSI TF III 7950 MSI TF III 7950 
RAMHard DriveHard DriveHard Drive
Rip Jaws 16GB @2400MHz Samsung 850 Evo SSD Seagate 250GB Hitachi 1TB 
Optical DriveCoolingOSMonitor
El-generico Cooler Master Hyper 212X 7 ultimate 64bit Tripe Viewsonic 22", 5760x1080 eyefinity secks. 
KeyboardPowerCaseMouse
Logitech G15 HX750i Phanteks Enthoo Pro Logitech G402 
Mouse Pad
ಠ_ಠSolid oak deskಠ_ಠ 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
i7 5930k MSI X99A SLI Krait Edition MSI TF III 7950 MSI TF III 7950 
RAMHard DriveHard DriveHard Drive
Rip Jaws 16GB @2400MHz Samsung 850 Evo SSD Seagate 250GB Hitachi 1TB 
Optical DriveCoolingOSMonitor
El-generico Cooler Master Hyper 212X 7 ultimate 64bit Tripe Viewsonic 22", 5760x1080 eyefinity secks. 
KeyboardPowerCaseMouse
Logitech G15 HX750i Phanteks Enthoo Pro Logitech G402 
Mouse Pad
ಠ_ಠSolid oak deskಠ_ಠ 
  hide details  
Reply
post #26 of 33
Quote:
Originally Posted by alex98uk View Post
Does the OCCT tool check for memory errors? I can't do GPU Mem because I don't have a CUDA card, but will the standard GPU: OCCT work?

I maxed the core at about 1000mhz, but I want to work on the memory which is atm, only 1100mhz.
Yes there is a specific test on OCCT that tests GPU memory.
post #27 of 33
Quote:
Originally Posted by E-Peen View Post
Sounds good. I'm going to try for 1250 memory, see what happens. If I get no errors/artifacts I'll stop right there for memory and start doing the core.

Then the gaming/tweaking begins

Quick question!: Is it okay to have other windows over a stability test such as Furmark? Or does it need to be the foreground and uninterrupted by other tasks?
You should work on the core clock first IMO; it's going to give you more performance in contrast from what you'll get on the memory.
I'd suggest just keeping the memory at stock while you push the core, then bring it back up to about 1200 and work from there.

The memory on 5850 and 5870 are identical, so 1200 is typically a pretty safe bet for stability (prior to ECC correction).

In regards to your testing software, you should keep it in the foreground simply for the fact that you'll want to monitor it while it's running.
Whether that be looking for artifacts or keeping an eye on performance while adjusting memory; having another window over it won't harm anything - but you will want to keep close eye on it though.
Stugots III
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K @ 4.5GHz ASUS Maximus V Gene EVGA GTX 1060 SC 2x8GB G.Skill 2133MHz 
Hard DriveOptical DriveCoolingOS
250GB 850-EVO | 1TB EZEX | 2TB F4 It BURNS! H100i | NB Pros Windows 10 Pro x64 
MonitorKeyboardPowerCase
ASUS VE258Q Filco Majestouch I Seasonic X-560w Fractal Arc Mini R2 
MouseMouse PadAudioAudio
Logitech G502 Vengeance MM200 HT Omega eClaro SHURE SRH-840 | Audioengine A2+S8 
  hide details  
Reply
Stugots III
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K @ 4.5GHz ASUS Maximus V Gene EVGA GTX 1060 SC 2x8GB G.Skill 2133MHz 
Hard DriveOptical DriveCoolingOS
250GB 850-EVO | 1TB EZEX | 2TB F4 It BURNS! H100i | NB Pros Windows 10 Pro x64 
MonitorKeyboardPowerCase
ASUS VE258Q Filco Majestouch I Seasonic X-560w Fractal Arc Mini R2 
MouseMouse PadAudioAudio
Logitech G502 Vengeance MM200 HT Omega eClaro SHURE SRH-840 | Audioengine A2+S8 
  hide details  
Reply
post #28 of 33
Thread Starter 
Quote:
Originally Posted by candy_van View Post
You should work on the core clock first IMO; it's going to give you more performance in contrast from what you'll get on the memory.
I'd suggest just keeping the memory at stock while you push the core, then bring it back up to about 1200 and work from there.

The memory on 5850 and 5870 are identical, so 1200 is typically a pretty safe bet for stability (prior to ECC correction).

In regards to your testing software, you should keep it in the foreground simply for the fact that you'll want to monitor it while it's running.
Whether that be looking for artifacts or keeping an eye on performance while adjusting memory; having another window over it won't harm anything - but you will want to keep close eye on it though.
Well said! I'll get to work and report back later with results and any questions I may have.

Cheers for all the suggestions & help everyone
post #29 of 33
Quote:
Originally Posted by VanillaCena View Post
Yes there is a specific test on OCCT that tests GPU memory.
That only works with CUDA enabled cards. So ATI's won't work with it.
post #30 of 33
Quote:
Originally Posted by wumpus View Post
furmark and kombustor are useless
^This, I dont know why people continue to use these because they are garbage. Evga OC Scanner falls into this list as well.
OCCT is by far the best way to test your core clock. If you can pass 15 minutes of that with fullscreen, shader complexity 8, with error checking on your core clock is good to go. None of those test the memory very well at all though so I always run a couple loops of crysis bench afterwards with ridiculous amounts of AA to push that memory.
 
Media Server
(12 items)
 
 
CPUMotherboardGraphicsGraphics
i7 4770K MSI Z97S Krait XFX RX 480 Black Edition XFX RX 480 Black Edition 
RAMHard DriveHard DriveOptical Drive
G.Skill Trident X 2666mhz Intel 530 Sandisk Extreme Pro LG WH14NS40 
Optical DriveCoolingOSMonitor
LG WH16NS40 Noctua NH-D14 Win 10 64-Bit BenQ XL2730Z 
KeyboardPowerCaseMouse
Logitech G710+ Seasonic Platinum 1000w Fractal Design R5 Zowie FK1+ 
Mouse PadOtherOther
XTrac Ripper XXL  4 x Vardar F-4 120mm (2x NH-D14, 2x Bottom Intake) 3 x Vardar F-3 140mm Intake (two front, one side) 
CPUMotherboardGraphicsRAM
Intel Celeron J1800 QNAP Proprietary  Intel HD Graphics Crucial 8GB Kit 
Hard DriveHard DriveHard DriveHard Drive
4TB Western Digital Red Pro 4TB Western Digital Red Pro 4TB Western Digital Red Pro 4TB Western Digital Red Pro 
CoolingOSPowerCase
Stock Embedded Linux External, 96 W, 100-240 V TS-451 
  hide details  
Reply
 
Media Server
(12 items)
 
 
CPUMotherboardGraphicsGraphics
i7 4770K MSI Z97S Krait XFX RX 480 Black Edition XFX RX 480 Black Edition 
RAMHard DriveHard DriveOptical Drive
G.Skill Trident X 2666mhz Intel 530 Sandisk Extreme Pro LG WH14NS40 
Optical DriveCoolingOSMonitor
LG WH16NS40 Noctua NH-D14 Win 10 64-Bit BenQ XL2730Z 
KeyboardPowerCaseMouse
Logitech G710+ Seasonic Platinum 1000w Fractal Design R5 Zowie FK1+ 
Mouse PadOtherOther
XTrac Ripper XXL  4 x Vardar F-4 120mm (2x NH-D14, 2x Bottom Intake) 3 x Vardar F-3 140mm Intake (two front, one side) 
CPUMotherboardGraphicsRAM
Intel Celeron J1800 QNAP Proprietary  Intel HD Graphics Crucial 8GB Kit 
Hard DriveHard DriveHard DriveHard Drive
4TB Western Digital Red Pro 4TB Western Digital Red Pro 4TB Western Digital Red Pro 4TB Western Digital Red Pro 
CoolingOSPowerCase
Stock Embedded Linux External, 96 W, 100-240 V TS-451 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Overclock.net › Forums › Graphics Cards › AMD/ATI › ATI Drivers and Overclocking Software › *Puts on shades* 5850 overclocking time, advice?