Overclock.net › Forums › Graphics Cards › AMD/ATI › 6900 series: how do Disable powerthrottle!
New Posts  All Forums:Forum Nav:

6900 series: how do Disable powerthrottle! - Page 3

post #21 of 28
Quote:
Originally Posted by xioros View Post
Thx for testing, rep+
+
Yes, it does disable it, at least I think:
Mode 1 + furmark: throttle
Mode 2 + furmark: no trottle



huh? xD,
It seems to work on my cards...
I tested it, works for me as well.

http://www.overclock.net/amd-ati/108...l#post14426711
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 Logitech G402 Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 @ 4.4/3.2GHz core/uncore, 1.36 vcore, 1.2... Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 2x Samsung MV-3V4G3D/US @ 2000, 10-11-11-30-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 2000, 10-11-11-3... OCZ (Toshiba) Trion 150 120GB Hyundai Sapphire 120GB 3x Hitachi Deskstar 7k1000.C 1TB 
CoolingOSPowerCase
Noctua NH-D14 Windows 7 Pro x64 SP1 Antec TP-750 Fractal Design R5 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
i7-6800K @ 4.3/3.5GHz core/uncore, 1.36/1.2v ASRock X99 OC Formula (P3.10) GTX 780 (temporary) 4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v 
Hard DriveHard DriveCoolingOS
Intel 600p 256GB NVMe 2x HGST Travelstar 7k1000 1TB Corsair H55 (temporary) Windows Server 2016 Datacenter 
PowerCase
Seasonic SS-860XP2 Corsair Carbide Air 540 
  hide details  
Reply
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 Logitech G402 Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 @ 4.4/3.2GHz core/uncore, 1.36 vcore, 1.2... Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 2x Samsung MV-3V4G3D/US @ 2000, 10-11-11-30-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 2000, 10-11-11-3... OCZ (Toshiba) Trion 150 120GB Hyundai Sapphire 120GB 3x Hitachi Deskstar 7k1000.C 1TB 
CoolingOSPowerCase
Noctua NH-D14 Windows 7 Pro x64 SP1 Antec TP-750 Fractal Design R5 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
i7-6800K @ 4.3/3.5GHz core/uncore, 1.36/1.2v ASRock X99 OC Formula (P3.10) GTX 780 (temporary) 4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v 
Hard DriveHard DriveCoolingOS
Intel 600p 256GB NVMe 2x HGST Travelstar 7k1000 1TB Corsair H55 (temporary) Windows Server 2016 Datacenter 
PowerCase
Seasonic SS-860XP2 Corsair Carbide Air 540 
  hide details  
Reply
post #22 of 28
Thread Starter 
Quote:
Originally Posted by gonX View Post
Once you go past a certain threshold it will limit anyway.
Love this emotion: ?
But yea, probably... after it reaches the 20% amount, I could start throttling...

Quote:
Originally Posted by LethalRise750 View Post
Heh, Guess I should be happy my 6970 has a bios without a limit.
Huh? Showme

Quote:
Originally Posted by Blameless View Post
thx for testing rep +
   
The Singularity
(19 items)
 
CPUMotherboardGraphicsRAM
Core i7 4790K @4GHz with stock cooler ASUS Maximus VI Gene GTX 1080 Ti G.Skill RipjawsX 16 GB 1600 MHz 
Hard DriveHard DriveOSMonitor
840 Pro Series 256 GB WD Black 3TB Windows 7 Pro x64 Dell U3415W 
KeyboardPowerCase
Dell U2715H Corsair AX860 with terrible MDPC-x sleeving job Corsair 350D 
CPUMotherboardGraphicsRAM
Undecided: Coffee Lake 6c? Undecided Nvidia GeForce GTX 1080 Ti 16 GB high-speed DDR4 
Hard DriveCoolingMonitorMonitor
1TB Samsung 960 EVO NVMe Silent custom liquid cooling Dell U3415W Dell U2715H 
KeyboardPowerCaseMouse
Corsair K70 Cherry MX Red Corsair AX860 /w custom length MDPC-x cables CaseLabs SMA8 Mionix Naos 7000 
CPUMotherboardGraphicsGraphics
[Intel] Core i5 2500K [ASUS] Maximus IV Extreme [Gainward] GeForce GTX 560 Ti Phantom 2 GB [Gainward] GeForce GTX 560 Ti Phantom 2 GB 
RAMHard DriveHard DriveOptical Drive
[G.Skill] Ripjaws-X [Corsair] Force Series 3 120GB [Samsung] F3 [Asus] DRW 24B3LT 
CoolingCoolingCoolingOS
[Corsair] H100 [Gainward] Phantom [Custom] Red Cooling [Microsoft] Windows 7 Ultimate 
MonitorKeyboardPowerCase
[Acer] 3x S221HQ EyeFinity [Logitech] G110 [Corsair] AX1200 [NZXT] Phantom Singularity mod by Xioros 
MouseMouse PadAudio
[Roccat] Kone [+] [Razer] Vespula [Logitech] X-530 5.1 Surround 
  hide details  
Reply
   
The Singularity
(19 items)
 
CPUMotherboardGraphicsRAM
Core i7 4790K @4GHz with stock cooler ASUS Maximus VI Gene GTX 1080 Ti G.Skill RipjawsX 16 GB 1600 MHz 
Hard DriveHard DriveOSMonitor
840 Pro Series 256 GB WD Black 3TB Windows 7 Pro x64 Dell U3415W 
KeyboardPowerCase
Dell U2715H Corsair AX860 with terrible MDPC-x sleeving job Corsair 350D 
CPUMotherboardGraphicsRAM
Undecided: Coffee Lake 6c? Undecided Nvidia GeForce GTX 1080 Ti 16 GB high-speed DDR4 
Hard DriveCoolingMonitorMonitor
1TB Samsung 960 EVO NVMe Silent custom liquid cooling Dell U3415W Dell U2715H 
KeyboardPowerCaseMouse
Corsair K70 Cherry MX Red Corsair AX860 /w custom length MDPC-x cables CaseLabs SMA8 Mionix Naos 7000 
CPUMotherboardGraphicsGraphics
[Intel] Core i5 2500K [ASUS] Maximus IV Extreme [Gainward] GeForce GTX 560 Ti Phantom 2 GB [Gainward] GeForce GTX 560 Ti Phantom 2 GB 
RAMHard DriveHard DriveOptical Drive
[G.Skill] Ripjaws-X [Corsair] Force Series 3 120GB [Samsung] F3 [Asus] DRW 24B3LT 
CoolingCoolingCoolingOS
[Corsair] H100 [Gainward] Phantom [Custom] Red Cooling [Microsoft] Windows 7 Ultimate 
MonitorKeyboardPowerCase
[Acer] 3x S221HQ EyeFinity [Logitech] G110 [Corsair] AX1200 [NZXT] Phantom Singularity mod by Xioros 
MouseMouse PadAudio
[Roccat] Kone [+] [Razer] Vespula [Logitech] X-530 5.1 Surround 
  hide details  
Reply
post #23 of 28
Fellas, you would want your cards to throttle under certain conditions. There's' a reason the Powertune slide goes up to +20%. There's also a reason, at default, it isn't set as such. I've run 6970 crossfire, triple 6970, quadfire 6990 + 6970 x 2, & quadfire 6990s - never, with the throttle lifted to +20% & whilst using Afterburner did I see an instance of throttling. Adjusting the MSI config was never a requisite for this accomplishment. I pushed my clocks as high as 1025, in fact, yet never succumbed to power restraint in games (with slider in +20%).

It's really frivolous, though, to run furmark (or whichever so-called power virus) on a regular basis. Powertune is an exceptional implementation for keeping the PCB circuitry safe. It safety mechanism isn't limited to the program in use or an arbitrary clock settings - it polls the entire activity throughput of the chip & power phases to determine potential subsequent "throttling". The throttling itself is in no means traditional; more aptly put, it gracefully caps the original clock and gradually reels down speed until the polling process determines the reduction has reached a safe-equilibrium.

In my experience, Powertune's maximum settings will allow intensely high (super 1 GHz) in-game clocks with nary a hint of recoil. If furmark, though, illicits the process to effect, there's a good reason for it. The feature, again, doesn't identify programs by executable names..it's not biased in this regard. It's working on the degree of chip activity! Under such circumstances, let it do it's thing.
Quadfire Turbo
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 - 2600K / 5.4 GHz Gigabyte P67A-UD7-B3 #2 Sapphie Radeon 6990 @ 1030 / 1520 Ripjaws 4GB x 4 (16GB) @ 2200 MHz 
Hard DriveOptical DriveOSMonitor
OCZ Vertex3,Corsair Force GT,Intel X25-M, +4TB HDD LG Bluray Burner Windows 7 64 5x1 P(NEC ips), S750 Samsung 120hz HD3D, HP zr30w 
KeyboardPowerCaseMouse
Logitech G13 #2 Kingwin Mach 1 / 1220 Watts XSPC H2 R.A.T. 7 
Mouse Pad
Razer Goliath 
  hide details  
Reply
Quadfire Turbo
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 - 2600K / 5.4 GHz Gigabyte P67A-UD7-B3 #2 Sapphie Radeon 6990 @ 1030 / 1520 Ripjaws 4GB x 4 (16GB) @ 2200 MHz 
Hard DriveOptical DriveOSMonitor
OCZ Vertex3,Corsair Force GT,Intel X25-M, +4TB HDD LG Bluray Burner Windows 7 64 5x1 P(NEC ips), S750 Samsung 120hz HD3D, HP zr30w 
KeyboardPowerCaseMouse
Logitech G13 #2 Kingwin Mach 1 / 1220 Watts XSPC H2 R.A.T. 7 
Mouse Pad
Razer Goliath 
  hide details  
Reply
post #24 of 28
Quote:
Originally Posted by XXXfire View Post
Fellas, you would want your cards to throttle under certain conditions. There's' a reason the Powertune slide goes up to +20%. There's also a reason, at default, it isn't set as such. I've run 6970 crossfire, triple 6970, quadfire 6990 + 6970 x 2, & quadfire 6990s - never, with the throttle lifted to +20% & whilst using Afterburner did I see an instance of throttling. Adjusting the MSI config was never a requisite for this accomplishment. I pushed my clocks as high as 1025, in fact, yet never succumbed to power restraint in games (with slider in +20%).

It's really frivolous, though, to run furmark (or whichever so-called power virus) on a regular basis. Powertune is an exceptional implementation for keeping the PCB circuitry safe. It safety mechanism isn't limited to the program in use or an arbitrary clock settings - it polls the entire activity throughput of the chip & power phases to determine potential subsequent "throttling". The throttling itself is in no means traditional; more aptly put, it gracefully caps the original clock and gradually reels down speed until the polling process determines the reduction has reached a safe-equilibrium.

In my experience, Powertune's maximum settings will allow intensely high (super 1 GHz) in-game clocks with nary a hint of recoil. If furmark, though, illicits the process to effect, there's a good reason for it. The feature, again, doesn't identify programs by executable names..it's not biased in this regard. It's working on the degree of chip activity! Under such circumstances, let it do it's thing.
This.
It's there for a reason. I would rather it downclock a couple times in Metro 2033 and have almost no fps change than risk damage.

+rep
My comp
(14 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II x4 955 GA-870A-UD3  Sapphire HD6950 Gskill 4GB 
Hard DriveCoolingOSMonitor
Western Digital Stock Win 7 x64 Samsung 
PowerCaseMouse
Corsair TX750 Antec 900 G5 
  hide details  
Reply
My comp
(14 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II x4 955 GA-870A-UD3  Sapphire HD6950 Gskill 4GB 
Hard DriveCoolingOSMonitor
Western Digital Stock Win 7 x64 Samsung 
PowerCaseMouse
Corsair TX750 Antec 900 G5 
  hide details  
Reply
post #25 of 28
I would rather risk damage than knowing that I don't get the maximum potential for my hardware. Probably the reason why I had been running at 1.49v on my 2600k for quite some time.

Hardware is pretty cheap, y'know. Setting PowerTune to higher than standard should still be fairly within spec, and not destroy anything. I don't think anyone has actually had their 69xx explode on them by setting PowerTune to 20%. If it doesn't explode within the first 15 minutes of Furmark I'd say you're pretty sure that it doesn't explode a couple of months down the road just playing games.
#well
(19 items)
 
Lenovo L530
(8 items)
 
 
CPUMotherboardGraphicsRAM
Intel Core i7 4770k Gigabyte Z87X-UD4H XFX Radeon HD 6950 Corsair CMX8GX3M2A2000C9 
Hard DriveHard DriveOptical DriveCooling
Samsung EVO 840 Hitachi HDS722020ALA330 Generic DVD±RW Burner Noctua NH-D14 
OSMonitorMonitorKeyboard
Windows 8.1 Dell U2711 LG W2453 Ducky DK9008 Overclock.net Edition, Cherry MX B... 
PowerCaseMouseMouse Pad
Corsair TX850 Antec Three Hundred Corsair Raptor M40 QPAD HeatoN M 
AudioAudioAudio
E-MU Tracker|pre Beyerdynamic DT-770 250 Ohm AntLion ModMIc 
CPUGraphicsRAMRAM
Intel Ivy Bridge 3210M Intel HD 4000 Graphics Soldered Corsair Vengeance  
Hard DriveOptical DriveOSMonitor
500GB DVD-+RW Windows 7 Professional 1600x900 
  hide details  
Reply
#well
(19 items)
 
Lenovo L530
(8 items)
 
 
CPUMotherboardGraphicsRAM
Intel Core i7 4770k Gigabyte Z87X-UD4H XFX Radeon HD 6950 Corsair CMX8GX3M2A2000C9 
Hard DriveHard DriveOptical DriveCooling
Samsung EVO 840 Hitachi HDS722020ALA330 Generic DVD±RW Burner Noctua NH-D14 
OSMonitorMonitorKeyboard
Windows 8.1 Dell U2711 LG W2453 Ducky DK9008 Overclock.net Edition, Cherry MX B... 
PowerCaseMouseMouse Pad
Corsair TX850 Antec Three Hundred Corsair Raptor M40 QPAD HeatoN M 
AudioAudioAudio
E-MU Tracker|pre Beyerdynamic DT-770 250 Ohm AntLion ModMIc 
CPUGraphicsRAMRAM
Intel Ivy Bridge 3210M Intel HD 4000 Graphics Soldered Corsair Vengeance  
Hard DriveOptical DriveOSMonitor
500GB DVD-+RW Windows 7 Professional 1600x900 
  hide details  
Reply
post #26 of 28
Quote:
Originally Posted by XXXfire View Post
Fellas, you would want your cards to throttle under certain conditions. There's' a reason the Powertune slide goes up to +20%. There's also a reason, at default, it isn't set as such. I've run 6970 crossfire, triple 6970, quadfire 6990 + 6970 x 2, & quadfire 6990s - never, with the throttle lifted to +20% & whilst using Afterburner did I see an instance of throttling. Adjusting the MSI config was never a requisite for this accomplishment. I pushed my clocks as high as 1025, in fact, yet never succumbed to power restraint in games (with slider in +20%).

It's really frivolous, though, to run furmark (or whichever so-called power virus) on a regular basis. Powertune is an exceptional implementation for keeping the PCB circuitry safe. It safety mechanism isn't limited to the program in use or an arbitrary clock settings - it polls the entire activity throughput of the chip & power phases to determine potential subsequent "throttling". The throttling itself is in no means traditional; more aptly put, it gracefully caps the original clock and gradually reels down speed until the polling process determines the reduction has reached a safe-equilibrium.
Using the term "power virus" for stress tests like OCCT or FurMark is simply an underhanded attempt by GPU makers at deflecting blame for their own malicious penny pinching.

Instead of spending another five dollars per card to equip them with VRMs that can handle whatever current the GPU is capable of drawing, they would rather limit performance in high demand situations.

If this isn't dishonest, I don't know what is. They are essentially advertising a performance level that can never be reached, because when you get anywhere near the theoretical maximum of what the card can do, PowerTune kicks in and knocks you down.

GPUs aren't toys any more. I run distributed computing projects on mine 24/7 and some of these are nearly as demanding on the hardware as FurMark.

I'm willing to accept that OCing is not a given, but at stock speeds, there is no software that any piece of hardware should not be able to run 24/7/365.25 for the entire duration of it's warranty, without hiccup, fuss, or failure, and without throttling.

If I want to run FurMark for 5 years straight, my card should be up to it. My CPUs are all up to doing the equivalent.

Quote:
Originally Posted by XXXfire View Post
In my experience, Powertune's maximum settings will allow intensely high (super 1 GHz) in-game clocks with nary a hint of recoil. If furmark, though, illicits the process to effect, there's a good reason for it. The feature, again, doesn't identify programs by executable names..it's not biased in this regard. It's working on the degree of chip activity! Under such circumstances, let it do it's thing.
The fact you are using a multi-GPU config that is almost certainly being bottlenecked by your CPU and Crossfire's overhead is distorting your results. Try again with one single GPU card (which is what the vast majority of the market is using).

Single cards with fast CPUs will throttle at STOCK speeds and PowerTune settings, in real games. I've seen it myself on 6950s and 6970s. It's also been documented in reviews:

http://www.anandtech.com/show/4061/a...deon-hd-6950/9

http://www.guru3d.com/article/radeon...6970-review/24

Even with power tune at +20% moderate OCs (well under 1GHz) will throttle in game.

Quote:
Originally Posted by Fortivo View Post
This.
It's there for a reason. I would rather it downclock a couple times in Metro 2033 and have almost no fps change than risk damage.
This is the sort of attitude that lets NVIDIA and ATI get away with selling you substandard garbage.

I'd rather have my card break once a month so I can RMA it and punish them to the tune of 1000% of what I spent until they either die and go out of business or fork over the few dollar per card it would take for them to be honest.

No one would accept a CPU they could only use at 80% load for a few hours a day, and they shouldn't accept GPU makers trying to pass off such limitations on their hardware as "normal" or okay.
Edited by Blameless - 8/2/11 at 4:00am
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 Logitech G402 Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 @ 4.4/3.2GHz core/uncore, 1.36 vcore, 1.2... Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 2x Samsung MV-3V4G3D/US @ 2000, 10-11-11-30-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 2000, 10-11-11-3... OCZ (Toshiba) Trion 150 120GB Hyundai Sapphire 120GB 3x Hitachi Deskstar 7k1000.C 1TB 
CoolingOSPowerCase
Noctua NH-D14 Windows 7 Pro x64 SP1 Antec TP-750 Fractal Design R5 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
i7-6800K @ 4.3/3.5GHz core/uncore, 1.36/1.2v ASRock X99 OC Formula (P3.10) GTX 780 (temporary) 4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v 
Hard DriveHard DriveCoolingOS
Intel 600p 256GB NVMe 2x HGST Travelstar 7k1000 1TB Corsair H55 (temporary) Windows Server 2016 Datacenter 
PowerCase
Seasonic SS-860XP2 Corsair Carbide Air 540 
  hide details  
Reply
Primary
(15 items)
 
Secondary
(13 items)
 
In progress
(10 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3/3.6GHz core/uncore, 1.225/1.2v Gigabyte X99 SOC Champion (F22n) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-11-12-27-T1, 1.37v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 Logitech G402 Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 @ 4.4/3.2GHz core/uncore, 1.36 vcore, 1.2... Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 2x Samsung MV-3V4G3D/US @ 2000, 10-11-11-30-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 2000, 10-11-11-3... OCZ (Toshiba) Trion 150 120GB Hyundai Sapphire 120GB 3x Hitachi Deskstar 7k1000.C 1TB 
CoolingOSPowerCase
Noctua NH-D14 Windows 7 Pro x64 SP1 Antec TP-750 Fractal Design R5 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
i7-6800K @ 4.3/3.5GHz core/uncore, 1.36/1.2v ASRock X99 OC Formula (P3.10) GTX 780 (temporary) 4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v 
Hard DriveHard DriveCoolingOS
Intel 600p 256GB NVMe 2x HGST Travelstar 7k1000 1TB Corsair H55 (temporary) Windows Server 2016 Datacenter 
PowerCase
Seasonic SS-860XP2 Corsair Carbide Air 540 
  hide details  
Reply
post #27 of 28
Quote:
Originally Posted by Blameless View Post
The fact you are using a multi-GPU config that is almost certainly being bottlenecked by your CPU and Crossfire's overhead is distorting your results. Try again with one single GPU card (which is what the vast majority of the market is using).
I've run single, cross, triple & quad setups with 6970s. The fact that I'm running quadfire, however, isn't skewing results from a "cpu bottleneck" - because I'm finding typical GPU usage @ 90%-99% in virtually every title I've engrossed myself & my time in. That's low-balling it, though, as the few titles below 94% tend to skew averages down. That's the benefit of high IPC coupled with extraordinary clock speeds (5.2 GHz), especially at 3x1 & 5x1 portrait modes. 6.82 megapixel & 11.4, respectively.

Your assumption of bottlenecked GPUs is fallacious. There are no distorted results. It shouldn't surprise, though, as the Powertune function of the performance bios is effectively unlocked for 6990s.

That can be set aside, though, as I was running 6990 with a pair of 6970s incidentally clocked even higher than current (1015/1525). And preceding that? I was running triple 6970s on water clocked higher still (1025/1525). In every instance of game-play, with typical GPU usage effectively identical despite the component differentiation. Hop over to the 6990 club thread to see my 30 minute run of furmark @ 97-99% usage & perfectly stable clocks for further support of these claims.

Regarding the various 6970 builds, Powertune +20% never demonstrated a single instance of throttle. Not in performance, clocks, at the wall power draw, whatever. Your, uh, exhaustive citation list does nothing to convince me otherwise. You're welcome to poll 6970 users or propose a case study, but considering my extensive experience (alongside my near maximum clock achievements with multiple Cayman processors) and having been throttled NEVER in-game at those clocks, I'm already convinced.

Quote:
Using the term "power virus" for stress tests like OCCT or FurMark is simply an underhanded attempt by GPU makers at deflecting blame for their own malicious penny pinching.

Instead of spending another five dollars per card to equip them with VRMs that can handle whatever current the GPU is capable of drawing, they would rather limit performance in high demand situations.

If this isn't dishonest, I don't know what is. They are essentially advertising a performance level that can never be reached, because when you get anywhere near the theoretical maximum of what the card can do, PowerTune kicks in and knocks you down.

GPUs aren't toys any more. I run distributed computing projects on mine 24/7 and some of these are nearly as demanding on the hardware as FurMark.

I'm willing to accept that OCing is not a given, but at stock speeds, there is no software that any piece of hardware should not be able to run 24/7/365.25 for the entire duration of it's warranty, without hiccup, fuss, or failure, and without throttling.

If I want to run FurMark for 5 years straight, my card should be up to it. My CPUs are all up to doing the equivalent.

I didn't use the term "powervirus" in any other regard other than Furmark. Furmark is not distributed computing. It creates a condition of exaggerated power draw that is, in respect of games, not remotely analogous. I've not seen any Furmark throttling with a pair of 6990s, though refrained from implementing the worthless benchmark with my 6970s.

I'm not entirely sure where you've researched the quality of AMD's Cayman VRM selection, but I'd suggest expanding said research a bit further out. AMD's component selection for the power phase is totally contrary to what you've implied with your "five dollar per card" remark.

http://semiaccurate.com/2010/12/22/a...er-microscope/

This review best details the Powertune function, and does enough to support my argument. Granted, testing does not include overclocked results .. however, the fact that throttling isn't occuring at 0% Powertune (it exclusively illicits at negative PT settings) can be extrapolated out very easily.

Quote:
This is the sort of attitude that lets NVIDIA and ATI get away with selling you substandard garbage.

I'd rather have my card break once a month so I can RMA it and punish them to the tune of 1000% of what I spent until they either die and go out of business or fork over the few dollar per card it would take for them to be honest.
Considering your assertions regarding throttling is utterly wrong, this follow-up remark seems to indicate a slight hint of .. vindictive hatred. Perhaps you don't have deep-seeded bias impacting your perspective, but what erroneous conclusions you make doesn't alter the fundamental reality of the implementation, by nature, is a feature that PROTECTS THE CARD from excessive power draw! How did we get from there to the device being "substandard garbage"? And that, even though AMD STATES the INTENDED TDP is 190 watts, them capping power-draw when it reaches 285 watts (many reviews have pulled 285 watts in Furmark) is dishonest?

Let me reiterate a point, here. Ask around, it's not tough to verify. At stock clocks & at substantial overclocks, the Powertune +20% feature DOES NOT THROTTLE THE CARD. I've tested it with three separate 6970 cards & two 6990s in dual, triple, & quad crossfire modes with various combinations of the chips.

Quote:
No one would accept a CPU they could only use at 80% load for a few hours a day, and they shouldn't accept GPU makers trying to pass off such limitations on their hardware as "normal" or okay.
I run 95-99% GPU usage in, for example, Metro 2033 @ 1015/1500 clocks for hours on end @ without a hint of throttling. Nobody will (or should) accept the flawed imaginations & thick rhetoric, such as "GPU makers trying to pass off such limitations as normal" .. because these limitations don't exist; especially in the degree which you've declared. Sorry.
Quadfire Turbo
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 - 2600K / 5.4 GHz Gigabyte P67A-UD7-B3 #2 Sapphie Radeon 6990 @ 1030 / 1520 Ripjaws 4GB x 4 (16GB) @ 2200 MHz 
Hard DriveOptical DriveOSMonitor
OCZ Vertex3,Corsair Force GT,Intel X25-M, +4TB HDD LG Bluray Burner Windows 7 64 5x1 P(NEC ips), S750 Samsung 120hz HD3D, HP zr30w 
KeyboardPowerCaseMouse
Logitech G13 #2 Kingwin Mach 1 / 1220 Watts XSPC H2 R.A.T. 7 
Mouse Pad
Razer Goliath 
  hide details  
Reply
Quadfire Turbo
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 - 2600K / 5.4 GHz Gigabyte P67A-UD7-B3 #2 Sapphie Radeon 6990 @ 1030 / 1520 Ripjaws 4GB x 4 (16GB) @ 2200 MHz 
Hard DriveOptical DriveOSMonitor
OCZ Vertex3,Corsair Force GT,Intel X25-M, +4TB HDD LG Bluray Burner Windows 7 64 5x1 P(NEC ips), S750 Samsung 120hz HD3D, HP zr30w 
KeyboardPowerCaseMouse
Logitech G13 #2 Kingwin Mach 1 / 1220 Watts XSPC H2 R.A.T. 7 
Mouse Pad
Razer Goliath 
  hide details  
Reply
post #28 of 28
Quote:
Originally Posted by Blameless View Post
In my testing on a pair of 6950s, the 6970 BIOSes do increase the TDP limit to the full 6970 values.

Still, even with 6970 TDP + 20% in CCC, more than roughly 900MHz core will throttle in FurMark, stock clocks will throttle in OCCT 4.x, and more than 910-920 will throttle in more strenuous games.

I've been looking for a away to completely disable PowerTune's throttling, but have yet to see anything that reliably does so. Can anyone confirm that AB allows you to do it?

And by confirm I mean run OCCT 4 at max shader quality for 20-30 minutes and record the lowest GPU clock in GPU-Z's sensor tab.
Are you sure? The power draw for the modded 6950 +20% is considerably less than the real 6970 +20%. FPS is the same, but that's not the point.



From this, I was under the impression that even when the 6950 is flashed to the 6970, the powertune chip behaves like a 6950. The chip that controls powertune is hardware and not software.

Quote:
Originally Posted by XXXfire View Post
I run 95-99% GPU usage in, for example, Metro 2033 @ 1015/1500 clocks for hours on end @ without a hint of throttling. Nobody will (or should) accept the flawed imaginations & thick rhetoric, such as "GPU makers trying to pass off such limitations as normal" .. because these limitations don't exist; especially in the degree which you've declared. Sorry.
Powertune working its magic does not only manifest itself in the core throttling. You can have powertune "holding your card back from peak performance" even if the core stays constant.

Do this test. Run a benchmark that you know will not throttle the core with powertune at 0%. The clocks need to be high and at the tipping point. That is, if the core is increased 10-20 more, then the card will throttle. Run the test again with powertune at +20%. You will get more fps even though the core does not throttle at 0% or at +20%. In short, powertune does not seem to be doing its thing as shown in the core throttle. But limiting the card's power consumption limits its peak performance.
Silent But Deadly
(15 items)
 
  
CPUMotherboardGraphicsRAM
3570k @4.4 Gigabyte Z77X-UD3H Powercolor AMD 470 8GB Samsung 
Hard DriveCoolingCoolingOS
Cricual M4 64 / WD6400AAKS Silver Arrow SB-E Arctic Cooling 7970 7 64 Ultimate 
MonitorPowerCaseMouse
BenQ XL2410T Corsair AX650 Fractal Define R4 Logitech G5 
  hide details  
Reply
Silent But Deadly
(15 items)
 
  
CPUMotherboardGraphicsRAM
3570k @4.4 Gigabyte Z77X-UD3H Powercolor AMD 470 8GB Samsung 
Hard DriveCoolingCoolingOS
Cricual M4 64 / WD6400AAKS Silver Arrow SB-E Arctic Cooling 7970 7 64 Ultimate 
MonitorPowerCaseMouse
BenQ XL2410T Corsair AX650 Fractal Define R4 Logitech G5 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › 6900 series: how do Disable powerthrottle!