Gigabyte G1 Gaming 970, only +25 Core Voltage

[Quentin Chalmers]

A couple days ago I received my GIGABYTE GV-N970G1 GAMING-4GD GeForce GTX 970. I have been screwing around with overclocking it quite a bit and have noticed two things that seem off:

1. Factory Default Voltage for the card is 1218 mv, but when I use MSI Afterburner to raise the voltage by +87 mv, the actual voltage rise from sensors has never gone past 1243 mv. This is only a +25 mv increase

2. Even under extreme overclocking and benchmarks, I have not seen the card TDP go above 89%, This doesn't seem right considering many reviews for GTX 970's were claiming the overclocking limitation would be TDP.

From my experience, this GPU is just as stable running 1500 Mhz core clock and 8000 Mhz memory on stock voltage and power limit than not. Thats right, raising the voltage and power limit to maximum does not help whatsoever on overclocking. I have pushed this card right up to where it starts to artifact on stock voltage and power limit, then raised the voltage and power limit to maximum (+87 mv and 112%) and the card gets no more stable!

Normally when overclocking you are able to go a fair amount farther when you raise the voltage and power limit...

Anybody else with experience know whats going on here?

 

[Quentin Chalmers]

thinking about overclocking my G1 970. (currently reading up on it, complete noob). haven't tried anything yet. but, prior to oc'ing...

Q. my GPU stock voltage is 862mv - tested with precisionX and GPUZ, is this potentially a good or bad thing? i see others are getting much higher @ 1200mv or so (at default/stock)

Something here doesn't seem right. You are saying that the observed voltage under maximum load is only 862 mv? How are you measuring this? That is a reasonable voltage at idle but not at load.

 

[alegelos]

My 970 G1 does:

Core: +185mhz
Mem: +500mhz
vCore: stock
Pwr Limit: +112%

 

[alegelos]

Does anyone know how to pass the 1600mhz?

 

[Quentin Chalmers]

Does anyone know how to pass the 1600mhz?

Well that all depends on your GPU's overclocking potential. There are many possible factors limiting your overclock. These can include thermal throttling, artifacts, and power and voltage limits. For me the limit is artifacts. My 970 trys to boost extremely high when under load but I have to limit it because I get artifacts 🙁 I only have my GPU set to +86 on the core clock and +500 on the memory clock, however under load my card runs at a constant 1490 core and 8000 effective memory. As you can see my GPU is boosting way above its rated core clock. However If I set the core clock any higher in MSI afterburner I get artifacts in games... makes me sad 🙁

 

[Jesserson]

Thought I'd share the oc results of my Gigabyte GTX 970 G1 Gaming. Upping the voltage to +87 mv, the core clock to +170 and the mem clock to +350 I'm stable running fire strike, valley, heaven, furmark etc.... However even if I lower the core voltage down to a measly +50 I cannot get the mem clock to reach above 390 without artifacts or the software crashing. So I have settled for a higher core clock and lower mem clock because of the fact that the core clock affects performance much more. Quite honestly I expected better results considering Gigabyte cherry picked these chip. But I digress. I would love to know if anybody else is experiencing the same thing that I am.

 

[Quentin Chalmers]

Thought I'd share the oc results of my Gigabyte GTX 970 G1 Gaming. Upping the voltage to +87 mv, the core clock to +170 and the mem clock to +350 I'm stable running fire strike, valley, heaven, furmark etc.... However even if I lower the core voltage down to a measly +50 I cannot get the mem clock to reach above 390 without artifacts or the software crashing. So I have settled for a higher core clock and lower mem clock because of the fact that the core clock affects performance much more. Quite honestly I expected better results considering Gigabyte cherry picked these chip. But I digress. I would love to know if anybody else is experiencing the same thing that I am.

What sustained boost clock does your card run at under load?

I find it weird that you need the voltage boost, with me it has no affect on overclocking. Guess everybody has different results.

If you read reviews on newegg about our GPU everybody talks about how it hits 1500/8000 easy. Unfortunately not all of the "cherry picked" gpus are this good.

 

[Jesserson]

Thought I'd share the oc results of my Gigabyte GTX 970 G1 Gaming. Upping the voltage to +87 mv, the core clock to +170 and the mem clock to +350 I'm stable running fire strike, valley, heaven, furmark etc.... However even if I lower the core voltage down to a measly +50 I cannot get the mem clock to reach above 390 without artifacts or the software crashing. So I have settled for a higher core clock and lower mem clock because of the fact that the core clock affects performance much more. Quite honestly I expected better results considering Gigabyte cherry picked these chip. But I digress. I would love to know if anybody else is experiencing the same thing that I am.

What sustained boost clock does your card run at under load?

I find it weird that you need the voltage boost, with me it has no affect on overclocking. Guess everybody has different results.

If you read reviews on newegg about our GPU everybody talks about how it hits 1500/8000 easy. Unfortunately not all of the "cherry picked" gpus are this good.

My sustained core clock boost is around 1560mhz

 

[Quentin Chalmers]

My sustained core clock boost is around 1560mhz

Wow, thats extremely good. Just makes me more disappointed that my G1 can only sustain 1490 without artifacts haha.

 

[Seraph21]

Is there no way to increase the voltage? Mine runs at 1575/8050 temps are only 58C max and the TDP doesnt go over 80% with 1.2500V There is so much more left in the card but I need the power to run it up there.

 

[DmC10]

Thought I'd share the oc results of my Gigabyte GTX 970 G1 Gaming. Upping the voltage to +87 mv, the core clock to +170 and the mem clock to +350 I'm stable running fire strike, valley, heaven, furmark etc.... However even if I lower the core voltage down to a measly +50 I cannot get the mem clock to reach above 390 without artifacts or the software crashing. So I have settled for a higher core clock and lower mem clock because of the fact that the core clock affects performance much more. Quite honestly I expected better results considering Gigabyte cherry picked these chip. But I digress. I would love to know if anybody else is experiencing the same thing that I am.

in the past the "the core clock affects performance much more" quote would be right, but not with the 900 series...
i've done some tests and safely running +80core and +500mem on the g1 970 (TDP 112%)
the memory artifacts at +605, so i think +580 is my limit, the core has gone up to +160 without problems.
PS. i dont know if it matters but i have ver. 1.0

 

[Quentin Chalmers]

Is there no way to increase the voltage? Mine runs at 1575/8050 temps are only 58C max and the TDP doesnt go over 80% with 1.2500V There is so much more left in the card but I need the power to run it up there.

Are you saying that the maximum voltage your card can reach when overclocked is 1.25 volts? That seems unreasonably low.

 

[sz0ty0l4]

my gtx 980's power consumption don't go over 60% normally. why is this? i see people on youtube with 110% tdp. however my core clock is only 1316 and memory is 3504 but still 60% is really low. if i set +20mv voltage it gets to 70% but i still consider that low. why is this? can be fault of my PSU?
Please help! I Opened a thread about this but no answers so far 🙁
http://www.tomshardware.co.uk/answers/id-2553257/gtx-980-gpu-load-voltage-question.html

thanks in advance

 

[Quentin Chalmers]

my gtx 980's power consumption don't go over 60% normally. why is this? i see people on youtube with 110% tdp. however my core clock is only 1316 and memory is 3504 but still 60% is really low. if i set +20mv voltage it gets to 70% but i still consider that low. why is this? can be fault of my PSU?
Please help! I Opened a thread about this but no answers so far 🙁
http://www.tomshardware.co.uk/answers/id-2553257/gtx-980-gpu-load-voltage-question.html

thanks in advance

Hey I just checked your thread, I can't help that much but I can answer a couple questions.

1. Yes, the VDDC is showing the current voltage your card is running at. 1.206 volts is perfectly fine. If you overclock you card and raise the voltage, you will see this number rise. On my GTX 970 I can only go up to 1.243 volts, but depending on your card you may be able to go higher. I would not be too concerned with raising the voltage on your card. Adding voltage for overclocking is very common, and it is unlikely to damage your card.

2. The reason you are not seeing full TDP percentage could be because you aren't running a program/game that is stressing it hard enough. Different games and programs will result in different TDP use. For example, when I run Dota 2, I stay near 50-60% of my cards TDP. However if I run Far Cry 4, I will be up closer to 90-100%. What game/software are you running when you get a TDP of 60%?

Another thing to mention, is 1316 your boost clock? Or base clock? If it is your boost clock that is a bit low for a 980 and would help explain your low TDP percentage. TDP rises as you stress your card more with a higher core clock.

 

[sz0ty0l4]

Hello first of all thank you for your answer to me. actually i had the 60% tdp in games like far cry 4, dying light with totally maxed settings on cards stock clock.
i only reached 75% in far cry 4 maxed no AA +4k DSR..... thats why i was worried, because i see others 100-110% tdp on 1080p.... however now that i added 20mv and 100mhz core clock( its at 1410 now) TDP went up to like 68-70% on 1080p but i still consider it very low.. anyway the frametimes are much better now. like 15ms and i see rarely 20ms.

I can only hope it's not the psu's fault

 

[Quentin Chalmers]

Hello first of all thank you for your answer to me. actually i had the 60% tdp in games like far cry 4, dying light with totally maxed settings on cards stock clock.
i only reached 75% in far cry 4 maxed no AA +4k DSR..... thats why i was worried, because i see others 100-110% tdp on 1080p.... however now that i added 20mv and 100mhz core clock( its at 1410 now) TDP went up to like 68-70% on 1080p but i still consider it very low.. anyway the frametimes are much better now. like 15ms and i see rarely 20ms.

I can only hope it's not the psu's fault

Hmmm that does seem like unusually low TDP use. How are your frame rates in games? 15-20 milliseconds for frame time variance is considered to be fairly bad, assuming that is 99% percentile.

Has your PC experienced any crashes or other problems? I ask because if the problem really was a faulty PSU you would be experiencing much worse issues than low gpu TDP percentage. What is the model of your PSU? It has sufficient amps on the 12 volt rail for your 980, right?

 

[sz0ty0l4]

Fsp raider 650w 80+ silver The box says " 12 single rail design designed for power-intensive vga cards, prodive strong power to any overloading conditions " not im sceptic about this. maybe it's a lie and its not providing the promised wattage? or i missed over something?

i thought it has sufficient amps, but i'm not sure anymore. maybe could you check for me?...

my pc didn't crash a single time since i have it. fps numbers are pretty fine in game, but sometimes i have drops and sometimes also "gpu usage " drops too.. but its not like games are unenjoyable or anything. just its pretty wierd and iwould like to understand whats the problem , because as you said aswell other people has 80%+ tdp on their cards
my firestrike combined score was 14500 everything on stock clocks
can it be that the 15-20ms frametimes are because of the tdp %?
in games like farcry 4 the gpu usage goes to 98-99% but the tdp still stays at maximum of 70%.. and i also see drops sometimes in both.

i apperciate your help

 

[Quentin Chalmers]

Fsp raider 650w 80+ silver The box says " 12 single rail design designed for power-intensive vga cards, prodive strong power to any overloading conditions " not im sceptic about this. maybe it's a lie and its not providing the promised wattage? or i missed over something?

i thought it has sufficient amps, but i'm not sure anymore. maybe could you check for me?...

my pc didn't crash a single time since i have it. fps numbers are pretty fine in game, but sometimes i have drops and sometimes also "gpu usage " drops too.. but its not like games are unenjoyable or anything. just its pretty wierd and iwould like to understand whats the problem , because as you said aswell other people has 80%+ tdp on their cards
my firestrike combined score was 14500 everything on stock clocks
can it be that the 15-20ms frametimes are because of the tdp %?
in games like farcry 4 the gpu usage goes to 98-99% but the tdp still stays at maximum of 70%.. and i also see drops sometimes in both.

i apperciate your help

Glad to help,

I just checked your power supply and it has 50 amps on a single 12 volt rail, the gtx 980 requires a minimum of 38 amps so your power supply should be just fine.

I honestly don't know much about what causes microstutter, it is often something you experience in SLI or when the game you are playing is requiring more VRAM than your gpu has. I don't think your power supply is the cause of the microstutter but I am not 100% sure. Is the microstutter actually bad enough to be noticeable when you play games?

Changes in TDP use in a game are normal. Different parts of the game are more or less demanding and your TDP may fluctuate a little.

There is one suggestion I have for you to see if your TDP use is normal, download Furmark, a GPU stressing program and run it. When I run furmark on my GTX 970 I see TDP use of 100%.

http://www.ozone3d.net/benchmarks/fur/

If you run furmark and do not see high TDP percentage then something is unusual for sure. If furmark shows high TDP percentage, then it will show your gpu is normal and you shouldn't worry.

 

[sz0ty0l4]

no its not that bad, btwi just ran firestrike to see and here is the result:
http://www.3dmark.com/3dm/6045386
the score seems fine, but i was monitoring the tdp,temps,gpu usage, voltages
and the maximum TDP i had was 75%, even though gpu was running on 99% during the graphics tests with 1.206v .
i'll think on a psu upgrade maybe that will solve it, but low TDP % can't damage the gpu in any way righT?

 

[Quentin Chalmers]

no its not that bad, btwi just ran firestrike to see and here is the result:
http://www.3dmark.com/3dm/6045386
the score seems fine, but i was monitoring the tdp,temps,gpu usage, voltages
and the maximum TDP i had was 75%, even though gpu was running on 99% during the graphics tests with 1.206v .
i'll think on a psu upgrade maybe that will solve it, but low TDP % can't damage the gpu in any way righT?

Dang, you've got yourself a nice gaming pc there, those results for a gtx 980 are great.

When I run Firestrike tests on my gtx 970 my TDP percentage during the test ranges between about 75% to 85%. My TDP percentage should be higher than yours because I have overclocked my card quite a bit. What this means is that the TDP percentage you are seeing during Firestrike seems totally normal.

Most likely other people are showing higher TDP percentage than you because they have overclocked their cards quite a bit. Is there any way you can compare your TDP percentage in firestrike to somebody else with a gtx 980 at similar clock speeds?

If you are still concerned I would recommend furmark as another good gpu stress test.

I would not recommend you replace your power supply. Most likely you won't notice any difference in microstutter or TDP percentage on your gpu. Low TDP percentage will not damage you card at all, it simply means that you GPU is efficient and is doing a good job at running games while not producing much heat.

From what I can tell, you don't need to be concerned about your TDP percentage use as long as your gpu is performing as expected. If you think the microstutter you experience is worse than other people with the same gpu as you then you can be concerned, but it is unlikely that your psu would be the cause of the microstutter.

 

[sz0ty0l4]

I think finally after 3 hours i found the answer for this wierd situation. i found 4 people on gtx 980 forum talking about this and the answer seem to be that each 3d party gtx 980( msi , gigabyte etc) has a different TDP as 100%, but i use afterburner or gpu-z the programms measure my TDP to the TDP given by the manufacturer( gigabyte in my case ) which is like 220-230 if i count with 230 it turn out 230*0.75= 172,5 is already higher than 100% of nvidia given TDP( 165W). so i'm already over 100% tdp when i'm at 75%. and it seems to just be a terrible missunderstanding of these given TDP numbers from my part... 😀
anyway thank you for your time and help at least now i know how to measure these things

 

[Quentin Chalmers]

I think finally after 3 hours i found the answer for this wierd situation. i found 4 people on gtx 980 forum talking about this and the answer seem to be that each 3d party gtx 980( msi , gigabyte etc) has a different TDP as 100%, but i use afterburner or gpu-z the programms measure my TDP to the TDP given by the manufacturer( gigabyte in my case ) which is like 220-230 if i count with 230 it turn out 230*0.75= 172,5 is already higher than 100% of nvidia given TDP( 165W). so i'm already over 100% tdp when i'm at 75%. and it seems to just be a terrible missunderstanding of these given TDP numbers from my part... 😀
anyway thank you for your time and help at least now i know how to measure these things

Awesome, glad you found a solution

Happy gaming!

 

[Donizs]

Hey guys, I'm not solving anyone's problem here, but more like coming with mine... I have windforce edition, not the gaming one, but it still overclocks like a champ, I have it running 1563MHz stable with no voltage increase, however, my problem is my memory. Not such a problem, just that I see at everyone runs at around 500MHz overclock, and mine does only 200... 220 and Valley crashes... Any solution?

 

[Quentin Chalmers]

Hey guys, I'm not solving anyone's problem here, but more like coming with mine... I have windforce edition, not the gaming one, but it still overclocks like a champ, I have it running 1563MHz stable with no voltage increase, however, my problem is my memory. Not such a problem, just that I see at everyone runs at around 500MHz overclock, and mine does only 200... 220 and Valley crashes... Any solution?

What brand is the memory on your gpu? If you can only overclock 200-220 Mhz my guess is you have the Hynix RAM which is much worse at overclocking. The people who are getting 500 Mhz overclock (like me) have the Samsung RAM which is higher quality.

 

[chrysalis]

Hi guys, I recently been playing around with my gpu settings mostly due to some instability.

I currently am annoyed with evga skimping on the voltage controller but it is what it is.

Previously I had my boost clocked pumped which would give me about 1465mhz max and mem pumped so it was approx 2000 (as reported by gpuz). Finally power set to 110% limit.

However this was causing driver crashes. Very frequent in valley, occasionally in games.
I then observed if I left everything at stock, the crashes still occured albeit less frequently, gpuz reported lots of vrel on the perfcap reason which I think is related to unable to get requested voltage reliably.

Took my rig to my friend's shop and we tried out one of the spare asus 970s in my rig, issues went away, also tried my 970 in one of his rigs which had a massive psu, to rule out my psu, issues still occurred.

For reference my gpu is an evga 970 FTW (not the fixed FTW+ which has a better voltage chip).

I then added +25mv to the voltage, this boosts the stock boost from 1405 to 1418 interestingly, even with power set to 90% the capable stock boost is higher than with stock voltage at 110%.

We observed no more vrel with +25mv with power capped at 90%. At 95% also no vrel, at 100% occasional vrel, at 110% lots of vrel.

Running benchmarks such as valley and 3dmark firestrike extreme, the score was no different between 90% and 110%. The difference been tho at 90% the gpu gets no vrel and limits itself, but at 110% it tries to get more voltage but seems to fail and can cause instability. So now I have +25mv, and 90% power cap. No crashes in games or bench apps. I have also set my vram again to 2000 and is stable, but I think I have been seeing very occasional minor artefacts driving around in gta5 so may drop the vram a bit (is samsung for those who curious).

Increasing the vram clock does improve my fire strike extreme score (for those curious).

My conclusion is these first gen evga 970 cards do indeed have voltage issues as some suspected.