New Posts  All Forums:Forum Nav:

Dual GTX 460s

post #1 of 11
Thread Starter 
OK, so I am folding on a x4 965 and a GTX 460. I just picked up another 460 and was wondering if I have to free up another core for the 2nd video card? Someone told me you should leave one core idle to handle Fermi folding, so my question is: Is that for one card or can one core manage two cards? Thanks.
post #2 of 11
I think they mean that you shouldn't have your CPU running @ 100% load across all cores. You just need enough CPU cycles free to feed the GPU what it needs, which isn't that much, which is why even at 100% CPU load you can still fold at high PPD (just not maximum PPD).

You don't need to dedicate an entire core to folding either one card or two, just leave one core < 100% load, with some free cycles you can dedicate to feeding your folding cards and you'll be fine.

EDIT: An alternative would be to give your folding higher priority so that it demands cpu cycles away from other programs if you do happen to hit 100% but want to maximize your folding PPD.
Bleeding Edge
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k @ 5.0Ghz 24/7 ASUS P8P67PRO AMD 5870x2 8GB GSkill Ripjaw X 1600 @ CL8 
Hard DriveOSMonitorKeyboard
ADATA S599 64GB SSD + 2x 1TB Spinpoint (Raid 0) Windows 7 64Bit Dell 23" Razer Lycosa 
PowerCase
Seasonic X650 HAF X 
  hide details  
Reply
Bleeding Edge
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k @ 5.0Ghz 24/7 ASUS P8P67PRO AMD 5870x2 8GB GSkill Ripjaw X 1600 @ CL8 
Hard DriveOSMonitorKeyboard
ADATA S599 64GB SSD + 2x 1TB Spinpoint (Raid 0) Windows 7 64Bit Dell 23" Razer Lycosa 
PowerCase
Seasonic X650 HAF X 
  hide details  
Reply
post #3 of 11
Quote:
Originally Posted by TheSchlaf View Post
OK, so I am folding on a x4 965 and a GTX 460. I just picked up another 460 and was wondering if I have to free up another core for the 2nd video card? Someone told me you should leave one core idle to handle Fermi folding, so my question is: Is that for one card or can one core manage two cards? Thanks.
if you have a X4 you want to fold X3 (smp -3) you wont need to free another core, because the GPU client only uses 12-25% per client, so you can run 4 more GPU clients on that 1 core.
Gaming 2016
(17 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 6700K Gigabyte Z170M-D3H MSI GTX1070 8G ARMOR Kingston HyperX FURY DDR4 
Hard DriveHard DriveHard DriveCooling
Samsung 840 Evo Samsung 850 Evo Seagate SHHD Cooler Master Nepton 240M 
OSMonitorKeyboardPower
Windows 10 Pro XB321HK Logitech 710 Corsair RM550x 
CaseMouseMouse PadAudio
Cooler Master Silencio 352 Logitech G500 Razer Goliathus Presonus Eris 4.5 
  hide details  
Reply
Gaming 2016
(17 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 6700K Gigabyte Z170M-D3H MSI GTX1070 8G ARMOR Kingston HyperX FURY DDR4 
Hard DriveHard DriveHard DriveCooling
Samsung 840 Evo Samsung 850 Evo Seagate SHHD Cooler Master Nepton 240M 
OSMonitorKeyboardPower
Windows 10 Pro XB321HK Logitech 710 Corsair RM550x 
CaseMouseMouse PadAudio
Cooler Master Silencio 352 Logitech G500 Razer Goliathus Presonus Eris 4.5 
  hide details  
Reply
post #4 of 11
Thread Starter 
Quote:
Originally Posted by 1337LutZ View Post
if you have a X4 you want to fold X3 (smp -3) you wont need to free another core, because the GPU client only uses 12-25% per client, so you can run 4 more GPU clients on that 1 core.
Thanks. That's what I have going now.
post #5 of 11
On the tests I've done, it pays off more (PPD wise) if you just fold on all of your cores you take a very small hit on the GPU folding, but you make it up in what the CPU puts out.
I have 2 systems folding on all cores, and fermi's installed and folding, and there is very little difference in PPD which ever way you go.

EDIT:
I should mention on install of the GPU client, I gave them a higher priority then I did the other folding clients.
Edited by KarmaKiller - 3/12/11 at 1:45am
Obsiden V2.0
(18 items)
 
HP Touchpad
(0 items)
Random Tech Stuff
(27 photos)
CPUMotherboardGraphicsRAM
Intel Core I7 970 @ 4.2Ghz Asus Rampage Formula III ASUS GTX780 12GB G.Skill DDR3 1600 
Hard DriveHard DriveCoolingOS
Samsung 840  Toshiba 5tb Thermalright Silver Arrow Windows 10  
MonitorMonitorKeyboardPower
Auria 27" 1440p IPS @85hz Acer H243Hbmid 24"  Ducky Shine Zero TKL MX Browns OCZ 1010watt 
CaseMouseMouse PadAudio
Syrillian Built Tech Bench (Obsiden Silence) Zowie EC2 Blue SteelSeries QcK Mini Creative ZxR 
AudioAudio
KRK Rokkit 5 Gen2 Tonar USB Condesnsor Mic 
  hide details  
Reply
Obsiden V2.0
(18 items)
 
HP Touchpad
(0 items)
Random Tech Stuff
(27 photos)
CPUMotherboardGraphicsRAM
Intel Core I7 970 @ 4.2Ghz Asus Rampage Formula III ASUS GTX780 12GB G.Skill DDR3 1600 
Hard DriveHard DriveCoolingOS
Samsung 840  Toshiba 5tb Thermalright Silver Arrow Windows 10  
MonitorMonitorKeyboardPower
Auria 27" 1440p IPS @85hz Acer H243Hbmid 24"  Ducky Shine Zero TKL MX Browns OCZ 1010watt 
CaseMouseMouse PadAudio
Syrillian Built Tech Bench (Obsiden Silence) Zowie EC2 Blue SteelSeries QcK Mini Creative ZxR 
AudioAudio
KRK Rokkit 5 Gen2 Tonar USB Condesnsor Mic 
  hide details  
Reply
post #6 of 11
Thread Starter 
Quote:
Originally Posted by KarmaKiller View Post
On the tests I've done, it pays off more (PPD wise) if you just fold on all of your cores you take a very small hit on the GPU folding, but you make it up in what the CPU puts out.
I have 2 systems folding on all cores, and fermi's installed and folding, and there is very little difference in PPD which ever way you go.
Little difference even with SMP bonuses?
post #7 of 11
No that's what I'm saying, you make up the small hit in the GPU performance by the getting the CPU WU's done faster. (thus more bonus points)
Obsiden V2.0
(18 items)
 
HP Touchpad
(0 items)
Random Tech Stuff
(27 photos)
CPUMotherboardGraphicsRAM
Intel Core I7 970 @ 4.2Ghz Asus Rampage Formula III ASUS GTX780 12GB G.Skill DDR3 1600 
Hard DriveHard DriveCoolingOS
Samsung 840  Toshiba 5tb Thermalright Silver Arrow Windows 10  
MonitorMonitorKeyboardPower
Auria 27" 1440p IPS @85hz Acer H243Hbmid 24"  Ducky Shine Zero TKL MX Browns OCZ 1010watt 
CaseMouseMouse PadAudio
Syrillian Built Tech Bench (Obsiden Silence) Zowie EC2 Blue SteelSeries QcK Mini Creative ZxR 
AudioAudio
KRK Rokkit 5 Gen2 Tonar USB Condesnsor Mic 
  hide details  
Reply
Obsiden V2.0
(18 items)
 
HP Touchpad
(0 items)
Random Tech Stuff
(27 photos)
CPUMotherboardGraphicsRAM
Intel Core I7 970 @ 4.2Ghz Asus Rampage Formula III ASUS GTX780 12GB G.Skill DDR3 1600 
Hard DriveHard DriveCoolingOS
Samsung 840  Toshiba 5tb Thermalright Silver Arrow Windows 10  
MonitorMonitorKeyboardPower
Auria 27" 1440p IPS @85hz Acer H243Hbmid 24"  Ducky Shine Zero TKL MX Browns OCZ 1010watt 
CaseMouseMouse PadAudio
Syrillian Built Tech Bench (Obsiden Silence) Zowie EC2 Blue SteelSeries QcK Mini Creative ZxR 
AudioAudio
KRK Rokkit 5 Gen2 Tonar USB Condesnsor Mic 
  hide details  
Reply
post #8 of 11
Thread Starter 
Quote:
Originally Posted by KarmaKiller View Post
No that's what I'm saying, you make up the small hit in the GPU performance by the getting the CPU WU's done faster. (thus more bonus points)
Oh, OK.
post #9 of 11
Most of the WU's that i get use max 2% of my cpu. The 15-25% was with the older wu's on fermi's.
Sane
(17 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3930K @ 4.9Ghz 1.46V ASUS Rampage IV Formula ASUS GTX780 1176Mhz 1.187V 16GB GeIL Enhance Corsa 1600Mhz CL9 
Hard DriveCoolingCoolingCooling
Samsung 256GB 830 EK Supremacy Nickel-Plexi XSPC RS360 XSPC RX240 
CoolingOSMonitorKeyboard
5x Scythe Gentle Typhoon AP-15 Windows 8 Pro x64 Asus VE228T Logitech G19s 
PowerCaseMouse
Corsair HX850 NZXT Phantom 630 Razor Abyssus 
  hide details  
Reply
Sane
(17 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3930K @ 4.9Ghz 1.46V ASUS Rampage IV Formula ASUS GTX780 1176Mhz 1.187V 16GB GeIL Enhance Corsa 1600Mhz CL9 
Hard DriveCoolingCoolingCooling
Samsung 256GB 830 EK Supremacy Nickel-Plexi XSPC RS360 XSPC RX240 
CoolingOSMonitorKeyboard
5x Scythe Gentle Typhoon AP-15 Windows 8 Pro x64 Asus VE228T Logitech G19s 
PowerCaseMouse
Corsair HX850 NZXT Phantom 630 Razor Abyssus 
  hide details  
Reply
post #10 of 11
Quote:
Originally Posted by KarmaKiller View Post
On the tests I've done, it pays off more (PPD wise) if you just fold on all of your cores you take a very small hit on the GPU folding, but you make it up in what the CPU puts out.
I have 2 systems folding on all cores, and fermi's installed and folding, and there is very little difference in PPD which ever way you go.

EDIT:
I should mention on install of the GPU client, I gave them a higher priority then I did the other folding clients.
I have done the same with my 2600k. Ran both ways and find myself better off running -smp than -smp 7 with affinity set to run both gpus on the unused core. I am glad somebody else has had the same experience. I posted this in another thread and got jumped on!
The Downgrade
(13 items)
 
  
CPUMotherboardGraphicsRAM
455 unlocked to quad ASRock 990FX Extreme4 GT430 Adata 1600 4GB 
Hard DriveOptical DriveOSMonitor
C300 SSD LITESCRIBE DVD RW 7 Home Premium 64 ASUS 24" 
PowerCase
ABS 850W Dark Beret Powder Coated Rocketfish 
  hide details  
Reply
The Downgrade
(13 items)
 
  
CPUMotherboardGraphicsRAM
455 unlocked to quad ASRock 990FX Extreme4 GT430 Adata 1600 4GB 
Hard DriveOptical DriveOSMonitor
C300 SSD LITESCRIBE DVD RW 7 Home Premium 64 ASUS 24" 
PowerCase
ABS 850W Dark Beret Powder Coated Rocketfish 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Overclock.net Folding@Home Team