Overclock.net › Forums › Graphics Cards › NVIDIA › GTX 600 Series Unlocked Voltage Bios' downloads and tools.
New Posts  All Forums:Forum Nav:

GTX 600 Series Unlocked Voltage Bios' downloads and tools. - Page 733

post #7321 of 7931
Ok, so i finally had the time to try the new mod. Got it up and running. Whats the difference in this one compared to the other mod? Just curious. Thanks again DJ. Also, rep for your help. Never used the rep system before and didn't realize it was that relevant or I would have done it sooner lol.
post #7322 of 7931
No worries. You can go back and plus one any post he made in this thread that helped others.

Plus 1 to u too BTW
     
Reply
     
Reply
post #7323 of 7931
Quote:
Originally Posted by battleaxe View Post

No worries. You can go back and plus one any post he made in this thread that helped others.

Plus 1 to u too BTW
Already did thumb.gif Back at ya.
post #7324 of 7931
They should put DJ on payroll. Lol
     
Reply
     
Reply
post #7325 of 7931
Quote:
Originally Posted by kevindd992002 View Post

I'm assuming this is using KBT 1.27, right? How did you come up with those values?

Well, if I keep the BIOS with the max core clock and max mem clock Heaven 4.0 tends to crash. I never experienced any power throttle while monitoring PrecisionX performance monitor. It's always a crash/artifact that's being my basis of unstability.

How do you personally monitor throttling because of power limitation?
1.27
It was mostly after testing i got to these values. Logic for PCI-e here (Blah-blah-blah, no need to read biggrin.gif): Warning: Spoiler! (Click to show)
Like i said i had major headache with my *power problem* so i after some time i was changing them up and down, flashing, benching, nothing was working just because i did not realize that i had so much headroom actually on total power, changing a little up or down for testing did nothing, i have power meter on table too so i see how much my system consumes all the time.
Then discovered that giving more to the pci-e was doing miracles in my case but ofc at same time the power meter was showing much larger numbers... And lowering it a little was like choking my card, from 66 to 60 was sending speeds under 1000MHz, so i raised it to 75, card sucking more and more power, become afraid that maybe it will throw all that extra power through pic-e and will eventually damage mosfets if a large spike should occur... So i removed cooling, took out some multimeters, traced around card how the power flows through it etc, soldered 3 multimeters to measure all three power inputs in real time, after more flashing and testing with original and modified bioses and logging my results i came to a conclusion that gpu power runs evenly through 6 pwm-s at any given time, and if even one of them is limited (that being the pci-e power limit i triggered), all 6 are ordered to lower power not just one...
So after that i was sure that pwm-s should be safe and they are regulated evenly and i'm not pushing all that extra power through only one of them, the only thing to worry about was that i would not burn slot or motherboard traces or so, but given that standard is 75W and actually is probably way over it and all my needs are coming from 12V line, that meaning pretty low amperes, was brave enough to raise it a little over biggrin.gif, and just in case lowered maximum, to trigger limit faster if it really would suck too much from there, but that should never happen ingame anyway, maybe only furmark etc unnatural things... and changing the max had no effect before anyway...

Other changes to 260 and 280 are just my previous calculations and measures that how much my card actually draws, and i dont mind it limiting itself during some X situation, if it starts to make a difference in actual gaming (more than 13MHz downclock), then i probably will raise them a little...
(Bla-bla...) Warning: Spoiler! (Click to show)
After all my goal was to OC it for everyday use, flash bios with ideal setting for my needs and forget it... Since i am surround gamer, every drop is precious. But at the same time i would like it last long and do not want to see it crashing sometimes, that's why i have at least 13-26MHz safe zone for every speed and voltage, probably even too much on lower voltage range than 1,1V (friend did a little battlefield and crashed there during early oc stages, it throttled down there due to my *power problem* with oc-d memory)

Not going for any records etc, so the up to 1% i will gain by using 1320 instead of 1306 is not so important to me, 40fps vs 40,4 fps is not so big difference that i need it while it might sometimes result with crash, voltage is not so stable after all, i get 1,215-1,222V on multimeter while running heaven, drops to 1,204 when i move window (pauses heaven), up to 1,226 at least in skyrim, but ofc multimeter will not show very fast changes, so better to have small room to play with than get into crash by running near absolute max, it might happen just when changing scene in heaven triggers lower gpu usage, therefore less power needed, automatic voltage dropping due to load decrease and then next scene kicks in and triggers very fast low-voltage event?, GPU-Z render test will get 1,212V etc.

If you have it otherwise stable then you might be too near to max speed and need to back off at least some MHz... and if you are crashing already in heaven, then what about more demanding things like battlefield or so... for me skyrim surround with enb is far more demanding. Warning: Spoiler! (Click to show)
Only other thing i can think atm is that you will not see throttling is that when it wants to throttle it will try to lower voltage to 1,2V first and crashes or when it wants to reduce speed it runs into problems, or both together... like for lower speed it is specified to run at lower voltage... i have set voltage table manually for every speed just to be safer, but i'm not 100% sure that bios follows it always, but so far i seem to be having it almost my way, except on lower voltages it seems i need to specify 0,025V lower voltage than it actually sets etc... Default bios will run last 2 speeds with same voltage usually, so it should downclock first if having power problems, but i did not see any sign of throttling too when i crashed and lowering memory speed resulted being stable when i was having pcie power problems

Long and complicated stuff with many possible reasons, but 95% i would say you have just set it too high and need to back off just a little... If giving pci-e even few W more is not making any difference, then this is not the problem, like i said my +9W for pcie gives around 60W for total average realtime power usage and added 3W removes all throttling in skyrim and that with 4G board, you should have at least same already available with 2G board without touching anything if i'm not very off with memory power req.

Personally i left pretty much spare room for card to breathe.
And did test 1346 with added power... still crash, so it is just my limit to have 1333 heaven stable gpu @1,212V, 1320 will probably run everything, so i chose 1306 to be extra extra extra safe while losing up to 1%...

Damn i write a lot :S Sorry, my bad! biggrin.gif
Might as well write a book "Overclocking Gigabyte 4GB Windforce 3X 670/680 and 760/770 ver.1 OC graphics cards with non-reference 6+2 pwm" wink.gif
Edit: The book would say only "Raise pci-e default power and then do the usual things"

*And since we do not know that if someone else is really running into same problem and if rising pci-e power helps them, i might as well be the only one...
If it is confirmed then this info might as well be in new thread for gigabyte 4G boards or something to save someone from same headache. I did come here to seek help too.
Edited by maxufeiss - 3/6/14 at 11:57am
post #7326 of 7931
Quote:
Originally Posted by maxufeiss View Post

1.27
It was mostly after testing i got to these values. Logic for PCI-e here (Blah-blah-blah, no need to read biggrin.gif): Warning: Spoiler! (Click to show)
Like i said i had major headache with my *power problem* so i after some time i was changing them up and down, flashing, benching, nothing was working just because i did not realize that i had so much headroom actually on total power, changing a little up or down for testing did nothing, i have power meter on table too so i see how much my system consumes all the time.
Then discovered that giving more to the pci-e was doing miracles in my case but ofc at same time the power meter was showing much larger numbers... And lowering it a little was like choking my card, from 66 to 60 was sending speeds under 1000MHz, so i raised it to 75, card sucking more and more power, become afraid that maybe it will throw all that extra power through pic-e and will eventually damage mosfets if a large spike should occur... So i removed cooling, took out some multimeters, traced around card how the power flows through it etc, soldered 3 multimeters to measure all three power inputs in real time, after more flashing and testing with original and modified bioses and logging my results i came to a conclusion that gpu power runs evenly through 6 pwm-s at any given time, and if even one of them is limited (that being the pci-e power limit i triggered), all 6 are ordered to lower power not just one...
So after that i was sure that pwm-s should be safe and they are regulated evenly and i'm not pushing all that extra power through only one of them, the only thing to worry about was that i would not burn slot or motherboard traces or so, but given that standard is 75W and actually is probably way over it and all my needs are coming from 12V line, that meaning pretty low amperes, was brave enough to raise it a little over biggrin.gif, and just in case lowered maximum, to trigger limit faster if it really would suck too much from there, but that should never happen ingame anyway, maybe only furmark etc unnatural things... and changing the max had no effect before anyway...

Other changes to 260 and 280 are just my previous calculations and measures that how much my card actually draws, and i dont mind it limiting itself during some X situation, if it starts to make a difference in actual gaming (more than 13MHz downclock), then i probably will raise them a little...
(Bla-bla...) Warning: Spoiler! (Click to show)
After all my goal was to OC it for everyday use, flash bios with ideal setting for my needs and forget it... Since i am surround gamer, every drop is precious. But at the same time i would like it last long and do not want to see it crashing sometimes, that's why i have at least 13-26MHz safe zone for every speed and voltage, probably even too much on lower voltage range than 1,1V (friend did a little battlefield and crashed there during early oc stages, it throttled down there due to my *power problem* with oc-d memory)

Not going for any records etc, so the up to 1% i will gain by using 1320 instead of 1306 is not so important to me, 40fps vs 40,4 fps is not so big difference that i need it while it might sometimes result with crash, voltage is not so stable after all, i get 1,215-1,222V on multimeter while running heaven, drops to 1,204 when i move window (pauses heaven), up to 1,226 at least in skyrim, but ofc multimeter will not show very fast changes, so better to have small room to play with than get into crash by running near absolute max, it might happen just when changing scene in heaven triggers lower gpu usage, therefore less power needed, automatic voltage dropping due to load decrease and then next scene kicks in and triggers very fast low-voltage event?, GPU-Z render test will get 1,212V etc.

If you have it otherwise stable then you might be too near to max speed and need to back off at least some MHz... and if you are crashing already in heaven, then what about more demanding things like battlefield or so... for me skyrim surround with enb is far more demanding. Warning: Spoiler! (Click to show)
Only other thing i can think atm is that you will not see throttling is that when it wants to throttle it will try to lower voltage to 1,2V first and crashes or when it wants to reduce speed it runs into problems, or both together... like for lower speed it is specified to run at lower voltage... i have set voltage table manually for every speed just to be safer, but i'm not 100% sure that bios follows it always, but so far i seem to be having it almost my way, except on lower voltages it seems i need to specify 0,025V lower voltage than it actually sets etc... Default bios will run last 2 speeds with same voltage usually, so it should downclock first if having power problems, but i did not see any sign of throttling too when i crashed and lowering memory speed resulted being stable when i was having pcie power problems

Long and complicated stuff with many possible reasons, but 95% i would say you have just set it too high and need to back off just a little... If giving pci-e even few W more is not making any difference, then this is not the problem, like i said my +9W for pcie gives around 60W for total average realtime power usage and added 3W removes all throttling in skyrim and that with 4G board, you should have at least same already available with 2G board without touching anything if i'm not very off with memory power req.

Personally i left pretty much spare room for card to breathe.
And did test 1346 with added power... still crash, so it is just my limit to have 1333 heaven stable gpu @1,212V, 1320 will probably run everything, so i chose 1306 to be extra extra extra safe while losing up to 1%...

Damn i write a lot :S Sorry, my bad! biggrin.gif
Might as well write a book "Overclocking Gigabyte 4GB Windforce 3X 670/680 and 760/770 ver.1 OC graphics cards with non-reference 6+2 pwm" wink.gif
Edit: The book would say only "Raise pci-e default power and then do the usual things"

*And since we do not know that if someone else is really running into same problem and if rising pci-e power helps them, i might as well be the only one...
If it is confirmed then this info might as well be in new thread for gigabyte 4G boards or something to save someone from same headache. I did come here to seek help too.

Thank you for this VERY informative post smile.gif

Can't I just raise all the power sections there to a value that is known to be the max for that specific section? Let's say, I set the PCI-e power to 75W since it is known to be the max and then go ahead and set the others to 300W or something? Will that produce any potential damage to the card or the pcie slot?
post #7327 of 7931
Quote:
Originally Posted by kevindd992002 View Post

Thank you for this VERY informative post smile.gif

Can't I just raise all the power sections there to a value that is known to be the max for that specific section? Let's say, I set the PCI-e power to 75W since it is known to be the max and then go ahead and set the others to 300W or something? Will that produce any potential damage to the card or the pcie slot?
If i'm not very off then pci-e should handle much more easily, saw some bios'es with 600, and think that there are many just shorting measurement completely etc... and few watts should be always accounted for (unless with übercheapo motherboard...)
Even overvolted and overclocked until smoke comes out should not damage pcie slot with your card, it would be still probably under 100W usage from there..., any decent motherboard should be fine with that..
for 2G card i'd raise it 75W for pci-e (probably is not limiting at 66 yet), 80 maybe if afraid about it limiting, leave max on 100 (really should not matter much)...
If you are using it just for gaming and aint getting it to throttle down, there is no need for anything really... like i said i like it to throttle down a little in furmark etc, i do not plan to play a game with hairy donuts yet...
6-pin is by default over the limit (87W instead of 75W), 8-pin 150 is ok, 1&3 to 260 is under 10% more than original 240, 300 should be ok if you do not play with donuts too... it is just me being probably too cautious of sudden spikes, have not heard of massive 680 etc pwm burnings and reference 680 pcb has only 4 lower-amperage pwm-s compared to our cooled 6, so with 1,212V it is highly unlikely to burn our card, others might be 300 as well, but i'm not seeing any downclocking while gaming so no point for me...

If you are doing different things with your card and you need all the power it can push through then i do not know, probably all is safe until volting...
I would not push anything higher than i actually need...

Easiest way to find out if your pci-e might be limiting is to flash bios with 60W pci-e power, restart, run heaven, and if it is throttling some, you might be near limit and need to raise it to 70-75... it would be knowledge about 2G cards too, but it might not do any real fps damage even at 55W setting for you biggrin.gif
And i said my card is actually drawing over 90W from there with 78W/85W setting...
260-270W average total power for hours without throttling, so the card is drawing basically all it can use within *almost* stock power settings in game...
hmm... will test what surround battlefield 4 can draw or does it make me throttle...
EDIT: seems to be using a tiny bit more and might run into my total power limit if given enough time, might up my that just a little if i get proof to that someday, until then i'm fine. So 300W setting might be useful... (my data might be not very accurate on powers, but should be not very far off... with 260/300 setting i was sometimes seeing slightly larger usage sometimes, but i have 3 screens, receiver, laptop and pc behind power display, so enough room for error there... probably am getting higher than actual results and we are even safer, given usual googling for power usage on these)
Edited by maxufeiss - 3/6/14 at 3:38pm
post #7328 of 7931
Quote:
Originally Posted by evilkitty View Post


i changed the max values like you said to, leaving the def values as they were, i was not able to get another 13MHz

now start pushing the card and see what its results are then you up the voltage a little at a time in the bios then push again till you find a stable point as some cards cant hanlde  higher voltages and some hate low voltages.

Quote:
Originally Posted by Anti Hero View Post

Ok, so i finally had the time to try the new mod. Got it up and running. Whats the difference in this one compared to the other mod? Just curious. Thanks again DJ. Also, rep for your help. Never used the rep system before and didn't realize it was that relevant or I would have done it sooner lol.

i have more control with kepler bios tweaker 1.27 so it allows me to define a low value as well as a high value to balance the card. so if it cant run the high value it should still run the low value. I also upped the rail values.


Edited by djthrottleboi - 3/6/14 at 2:39pm
throttlemeister
(30 items)
 
  
CPUMotherboardRAMHard Drive
i7-4790k gigabyte GA-Z97X-Gaming G1 WIFI-BK corsair vengeance 2x8 1866Mhz seagate barracuda 
Hard DriveHard DriveHard DriveHard Drive
seagate barracuda  seagate barracuda seagate barracuda seagate barracuda 
Hard DriveOptical DriveCoolingCooling
Crucial M500 SSD random dvd multirecorder / cd drive fast sata 6... ex240 rad xspc ax240 kit 
CoolingCoolingCoolingOS
3 140mm fans plus 6 120mm fans Arctic Cooling MX4 4g Thermal Paste XSPC EX360 crossflow windows 8.1 pro 
OSOSMonitorKeyboard
ZorinOS Mac Os X Yosemite Monoprice MPCP28UHD logitech g510s 
PowerCaseMouseMouse Pad
evga supernova 1300 g2 Corsair 900D logitech g500s none  
AudioAudioAudioOther
Logitech X-540 5.1 Surround Sound Speaker w/ su... diamond xtreme sound XS71U Logitech H390 Headset xbox 360 wireless 
OtherOther
logitech c110 webcam logitech driving force gt 
  hide details  
Reply
throttlemeister
(30 items)
 
  
CPUMotherboardRAMHard Drive
i7-4790k gigabyte GA-Z97X-Gaming G1 WIFI-BK corsair vengeance 2x8 1866Mhz seagate barracuda 
Hard DriveHard DriveHard DriveHard Drive
seagate barracuda  seagate barracuda seagate barracuda seagate barracuda 
Hard DriveOptical DriveCoolingCooling
Crucial M500 SSD random dvd multirecorder / cd drive fast sata 6... ex240 rad xspc ax240 kit 
CoolingCoolingCoolingOS
3 140mm fans plus 6 120mm fans Arctic Cooling MX4 4g Thermal Paste XSPC EX360 crossflow windows 8.1 pro 
OSOSMonitorKeyboard
ZorinOS Mac Os X Yosemite Monoprice MPCP28UHD logitech g510s 
PowerCaseMouseMouse Pad
evga supernova 1300 g2 Corsair 900D logitech g500s none  
AudioAudioAudioOther
Logitech X-540 5.1 Surround Sound Speaker w/ su... diamond xtreme sound XS71U Logitech H390 Headset xbox 360 wireless 
OtherOther
logitech c110 webcam logitech driving force gt 
  hide details  
Reply
post #7329 of 7931
Quote:
Originally Posted by maxufeiss View Post

If i'm not very off then pci-e should handle much more easily, saw some bios'es with 600, and think that there are many just shorting measurement completely etc... and few watts should be always accounted for (unless with übercheapo motherboard...)
Even overvolted and overclocked until smoke comes out should not damage pcie slot with your card, it would be still probably under 100W usage from there..., any decent motherboard should be fine with that..
for 2G card i'd raise it 75W for pci-e (probably is not limiting at 66 yet), 80 maybe if afraid about it limiting, leave max on 100 (really should not matter much)...
If you are using it just for gaming and aint getting it to throttle down, there is no need for anything really... like i said i like it to throttle down a little in furmark etc, i do not plan to play a game with hairy donuts yet...
6-pin is by default over the limit (87W instead of 75W), 8-pin 150 is ok, 1&3 to 260 is under 10% more than original 240, 300 should be ok if you do not play with donuts too... it is just me being probably too cautious of sudden spikes, have not heard of massive 680 etc pwm burnings and reference 680 pcb has only 4 lower-amperage pwm-s compared to our cooled 6, so with 1,212V it is highly unlikely to burn our card, others might be 300 as well, but i'm not seeing any downclocking while gaming so no point for me...

If you are doing different things with your card and you need all the power it can push through then i do not know, probably all is safe until volting...
I would not push anything higher than i actually need...

Easiest way to find out if your pci-e might be limiting is to flash bios with 60W pci-e power, restart, run heaven, and if it is throttling some, you might be near limit and need to raise it to 70-75... it would be knowledge about 2G cards too, but it might not do any real fps damage even at 55W setting for you biggrin.gif
And i said my card is actually drawing over 90W from there with 78W/85W setting...
260-270W average total power for hours without throttling, so the card is drawing basically all it can use within *almost* stock power settings in game...
hmm... will test what surround battlefield can draw or does it make me throttle...

Ah.. But how do you know if your GPU is throttled by the power limit?

Can you please help me out here. Basically, the pictures below show the default power values of the stock BIOS of my card using KBT 1.27 in that order:







Which of those should I tinker?
post #7330 of 7931
Quote:
Originally Posted by djthrottleboi View Post

now start pushing the card and see what its results are then you up the voltage a little at a time in the bios then push again till you find a stable point as some cards cant handle higher voltages and some hate low voltages.
Ok, which voltages do i adjust and in what increments (zoom does not work very well so here in the raw image)

if any one wants to use my extended screenshots to make a tutorial go for it (i did screenshot merging in gimp)
Edited by evilkitty - 3/6/14 at 4:46pm
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: NVIDIA
Overclock.net › Forums › Graphics Cards › NVIDIA › GTX 600 Series Unlocked Voltage Bios' downloads and tools.