MSI Launches The GTX 980 Ti GAMING 6G

Why's that? If there's no artifacting under max game stress at 1492 (has to be 1500 or so for then to show) ? Worried about long term issues or just not worth 1-2 FPS if that?
Well to me its just common sense not to be right on the edge of where you know you have issues. Dropping another 25 mhz is not even going to make but 1 fps at most but its your card so do what you want I guess.
 
I've come to the conclusion that the mv slider does absolutely nothing. Tried it out with lower settings to see if it even affected the temps but I can't see this doing anything at all at any level. I did see the Voltage Limit graph in AB hit 1 @ 1550 when it reverted to default OC mode though.

So far I'm up to 1506/3758 @ 75-76c (68% fan inaudible @ auto settings) with a Valley score of 4257. Going to play some Skyrim and see if that holds. On a quick note I benched Metro LL Redux and the frames were all over the place like a mad womans shit with an average of 70 but that was stock clocks.
 
I've come to the conclusion that the mv slider does absolutely nothing. Tried it out with lower settings to see if it even affected the temps but I can't see this doing anything at all at any level. I did see the Voltage Limit graph in AB hit 1 @ 1550 when it reverted to default OC mode though.

So far I'm up to 1506/3758 @ 75-76c (68% fan inaudible @ auto settings) with a Valley score of 4257. Going to play some Skyrim and see if that holds. On a quick note I benched Metro LL Redux and the frames were all over the place like a mad womans shit with an average of 70 but that was stock clocks.
The Metro benchmarks have wildly varying framerates on any card. They can fully push the power limit too in spots even though its under 99% gpu usage for most of the run.
 
Been playing Skyrim since I last posted, It reverted back to default at 1506/3758 in 5 minutes. Although that was stable in Valley but meh. I dropped the clock to +145 and that gave me 1494 and 1482 was the throttled clock but only seen that a handful of times. So 1494/3758 seems like the daily, although I reckon the memory can go up. Max power was 102% that I saw. EDIT - And temps never exceeded 75c, pretty much 72 was the average outdoors.

I find it interesting that when it doesn't like the clock it just brings on a black screen, reverts to default and resumes game. My AMD cards would likely bring on BSOD or complete lockup so this little feature is fantastic I thinks. Looks like I won't be seeing 1500 though.
 
Because its best not to run at the brink 24/7, at least cards I've had degrade are the ones I ran full tilt out of the gate.
 
I've come to the conclusion that the mv slider does absolutely nothing. Tried it out with lower settings to see if it even affected the temps but I can't see this doing anything at all at any level.

I had a look at Kombuster, on the sphere OC test I noticed the clocks bounce all the way around 1400-1494 and so does the voltage. In fact the voltage goes as high as 1.16 and as low as 1.102 which i find bizarre but anyway. So when I increase mv to 25 the clock doesn't bounce around anywhere near as much and tends to sit on 1.187 @ 1494 with the occasional drop to 1480 like in skyrim without volts. Sometimes it goes to 1.212. Also this program reads 105% TDP at it's peak, and 78c with the volts (tested for 10-15mins). I'm really trying to understand how this thing works lol. I'll probably just stick to the clocks I'm at now though, makes my 7970 look ancient in Skyrim, solid 60's in places I could barely scrap 30 lol.
 
Welp, seems like 1492 MHz is my "Witcher 3" stable core clock at +25 mV, 109% power, +150/+400, and ~80% fan. Temps ~68-70 C...any further on the core and I start seeing wierd red/black artifacts in game.
My stable clock for witcher is 1444 (+140). Anything more and I see red dots - and that even happens on max voltage and max fan, so I reverted voltage to default (ASIC is 66%, if that makes any difference)
 
I had a look at Kombuster, on the sphere OC test I noticed the clocks bounce all the way around 1400-1494 and so does the voltage. In fact the voltage goes as high as 1.16 and as low as 1.102 which i find bizarre but anyway. So when I increase mv to 25 the clock doesn't bounce around anywhere near as much and tends to sit on 1.187 @ 1494 with the occasional drop to 1480 like in skyrim without volts. Sometimes it goes to 1.212. Also this program reads 105% TDP at it's peak, and 78c with the volts (tested for 10-15mins). I'm really trying to understand how this thing works lol. I'll probably just stick to the clocks I'm at now though, makes my 7970 look ancient in Skyrim, solid 60's in places I could barely scrap 30 lol.

From what I've read Nvidia might throttle Kombustor because people have been able to break their cards running it. I don't recommend using Kombustor or Furmark for testing.
 
Just got my 2 MSI Gaming 6G installed last night.

How are you folks with SLI dealing with heat? My case has pretty decent airflow and my upper card reaches a steady state 85 deg even with no overclock with the stock fan profile. Even setting fans to 85% only gets the one down to about 77 deg in while running Heaven.

A bit of OCing (+100 core) drives the one card all the way up 88. With any voltage bump to speak of it throttles off the temp limit of 91 deg even with fans at 100%.

I was planning on water cooling these cards anyway, but the blocks haven't been released yet, and frankly it makes me worry that this is just too many Watts for my 5x120 rad setup to cool (with my 4770k).

On the upside, if I can actually remove the heat, it looks like there's TONS of power to play with. Even with the voltage bumped to +87mV, my power didn't go over 100%, and it was only around 70% or so with stock voltage.

The fans *are* very quiet though - even at 100% they're not terribly loud. Not that I'd actually want to run them that high for any length of time simply for durability. My MSI mining cards back in the day almost all had fan bearing failures.
 
Just got my 2 MSI Gaming 6G installed last night.

How are you folks with SLI dealing with heat? My case has pretty decent airflow and my upper card reaches a steady state 85 deg even with no overclock with the stock fan profile. Even setting fans to 85% only gets the one down to about 77 deg in while running Heaven.

A bit of OCing (+100 core) drives the one card all the way up 88. With any voltage bump to speak of it throttles off the temp limit of 91 deg even with fans at 100%.

I was planning on water cooling these cards anyway, but the blocks haven't been released yet, and frankly it makes me worry that this is just too many Watts for my 5x120 rad setup to cool (with my 4770k).

On the upside, if I can actually remove the heat, it looks like there's TONS of power to play with. Even with the voltage bumped to +87mV, my power didn't go over 100%, and it was only around 70% or so with stock voltage.

The fans *are* very quiet though - even at 100% they're not terribly loud. Not that I'd actually want to run them that high for any length of time simply for durability. My MSI mining cards back in the day almost all had fan bearing failures.

Mate I don't know about SLI but I can only seem to get +140 stable in games at 1080P. Temperatures and noise with the auto fans are fantastic though, very happy there. I'm settling for a daily OC of 1494/3758, I know my memory could do more though. Same shit different arse as far as I'm concerned with the voltages on these, I don't see any benefit from it at all, other than the Kombuster test which I see mentioned above is no good anyway.
 
+140 gives you a boost clock of 1494? I also use +140 but that gives me boost clock of 1444 MHz.
 
+140 gives you a boost clock of 1494? I also use +140 but that gives me boost clock of 1444 MHz.

+120 gives me 1500. I settled on +103 at stock voltage which gives me +1457 which then throttles to 1444 or so after 65 degrees. I don't think the OC number means anything. Whatever it thinks the chip can handle is where it ends up at.
 
+120 gives me 1500. I settled on +103 at stock voltage which gives me +1457 which then throttles to 1444 or so after 65 degrees. I don't think the OC number means anything. Whatever it thinks the chip can handle is where it ends up at.

It should control the adjustment to the base clock. Boost 2.0 (or whatever it's called now) increases the clock speed dynamically depending on the heat, power and voltage. If none of these three are at the specified limit (which can be adjusted also with afterburner or similar tools), then it increases the core clock beyond the base speed. They call this boost clock. Ultimately what it means is that very demanding applications will have a lower boost clock than less demanding ones, so the clock speeds are going to be higher in Doom 3 than they would be in Crysis 3.

Here's a little info on it: http://www.geforce.com/hardware/technology/gpu-boost-2/technology

The monitoring ability of Afterburner allows you to see which parameter is limiting the boost clock by indicating a value of 1. So if your card hit the voltage maximum specified, your Voltage Limit would show a 1. If you then increased your voltage offset slightly to give it more headroom, you may run into the temperature limit with the extra voltage heating up the chip more.
 
+120 gives me 1500. I settled on +103 at stock voltage which gives me +1457 which then throttles to 1444 or so after 65 degrees. I don't think the OC number means anything. Whatever it thinks the chip can handle is where it ends up at.
That is exactly the number I settled on (1457 which throttles to 1444) but I have no idea why you achieved this with +103 while I have to set it to 140. Can anyone explain this difference?
 
That is exactly the number I settled on (1457 which throttles to 1444) but I have no idea why you achieved this with +103 while I have to set it to 140. Can anyone explain this difference?

because silicon lottery
 
Good to know, I thought the boost values are similar across models from the same manufacturer.

Cyph, what is your Asic?
 
73.9%. However, I've seen reports of lower ASIC that can reach a higher OC.

I'm sure modding the BIOS can get me higher with a little more heat, but I'm happy with the clock, especially since my temperature maxes out at 72C but mostly hovers around 68C at 70% fan. I wouldn't worry too much about the max OC since having a stable OC and a lower temperature is preferable. There are some lucky individuals who hit 1550 on stock voltage. I think if you can get 1450 at stock voltage, be happy and call it a day.
 
My ASIC is 74.4. Sometimes i see it throttle to 1482 for a few seconds but stays on 1494. If i above 1500 at all, even with more volts, it justs goes to a black screen and reverts it to stock oc mode which gives me 1342. It doesnt even display a driver crash message, although it did once when i tried that Kombuster program with the volts at +25 with 1506 and that only hit 78c so idk but im happy with 1494 at stock volts.

EDIT Actually im at +135 to get 1494.
 
Last edited:
Some observations this morning, i was playing FC3 in 4K (DSR) and it throttled down to 1476. Mostly fps was around the 45-50 mark sometimes shooting to 80. Any amount of voltage increase resulted in crashes and wildly varying voltages even without changing clocks. Temps sit around 75 occasionally going up to 76.

Skyrim, any amount of voltage increase results in a throttle to 1482 but without it will just sit at 1494. Crashes over +25mv. Sits on 1.187 by itself, voltage increases in AB make it drop. Had a play with the fan now, added a point in AB at 66c to go to 66%. Interestingly, the temps sit at 67c now at those same clocks and the fan likes to sit at 68-69%. Seems if the fans spin up a bit earlier they manage the temps better. I played for about an hour or so and that was where it stayed.
 
I found adding voltage only increased temperatures on my MSI gaming and also on my classified with no benefit in increasing clocks.
 
I found adding voltage only increased temperatures on my MSI gaming and also on my classified with no benefit in increasing clocks.

Sure does seem pretty useless, especially on that classified with the 14+3. If the asus dcu3 is anything like the dcu2 7970 i had i reckon it will be the worst off out of all of them bar reference.
 
YGvea79.jpg


Here is what happens with my classified.
60 C 1505 boost, 1.212 v
65 C 1492 boost, 1.2 v
68 C 1480 boost, 1.175 v

So much for 14+3 power delivery.
 
141% power eh? Guess that aint good for much either. Mine wont get to 1.2v easily it seems, likes to sit at 1.187 all the time.
 
My MSI arrived. Fan noise is much less and more pleasant tone compared to the Gigabyte G1 980 Ti I had. More air noise and less whine (not talking about coil whine, just fans), though at above 60% it gets audible and loud at around 80%.

ASIC quality is 73.8% so should be fine since I'm slapping an AIO cooler on this. With stock voltages I managed around 1450 MHz overclock at about 75°C. I'd say that's probably a sweet spot for this card on the stock cooler for speed, temps and noise.

I'll see how it handles the AIO cooler.
 
My poor cpu is running 10 C hotter than when I had a reference 980. I am hitting temps as high as 72 in Crysis 3 where the norm was 62 max in the same spot.
 
I moved from a 650d case to a HAF x yesterday as I anticipate SLI 980 tis to raise temperature significantly in my case. The move improved both my graphics card and Cpu temperature by more than 10c.
 
I moved from a 650d case to a HAF x yesterday as I anticipate SLI 980 tis to raise temperature significantly in my case. The move improved both my graphics card and Cpu temperature by more than 10c.
The HAF X has probably the best airflow out there but its just so ugly and dated looking to me though.
 
The HAF X has probably the best airflow out there but its just so ugly and dated looking to me though.

Ain't that the truth; I wouldn't be caught dead with that abomination, as good as its cooling is. :p

As for me, my CPU hasn't had its temp raised by much if at all. No more than 3C. Still runs quite frosty even at full load gaming session of Metro LL.
 
Mounted G10 + Corsair H55 on my MSI. There was a lot of thermal paste from the factory, enough to spill it over the chip edges. Took a while to clean that up. Also fitting the Corsair radiator into my case next to a huge Phanteks CPU cooler was a bit tricky. I ended up putting the radiator fan outside the case (no problem since I had already cut out the grille from the fan slot).

Temps dropped in idle from 60 -> 30 and load temps from 75 -> 60. Ran fine at least at 1509 overclock, will see if I can get more.

An AIO cooler seems to be a good fit for the 980 Ti but I would recommend using a fan controller (or if your motherboard has good fan control then that) to control the speeds on the radiator fan.
 
Just got home from work and tried my modded skyrim in 4k. Frames arent great but they are consistent which is a good sign, only 25-35 in demanding areas outdoors and 45+ indoors. What impressed me the most is that the card still stays at 1494 and the temps only slightly higher at 70c. CPU looked like it wanted to cook after an 20 mins though, went up to 79c but didnt go much higher. Unfortunately even 4k cant stop the jaggies as good as MSAA on those stick fences... Too bad thats not possible with ENB.
 
Mounted G10 + Corsair H55 on my MSI. There was a lot of thermal paste from the factory, enough to spill it over the chip edges. Took a while to clean that up. Also fitting the Corsair radiator into my case next to a huge Phanteks CPU cooler was a bit tricky. I ended up putting the radiator fan outside the case (no problem since I had already cut out the grille from the fan slot).

Temps dropped in idle from 60 -> 30 and load temps from 75 -> 60. Ran fine at least at 1509 overclock, will see if I can get more.

An AIO cooler seems to be a good fit for the 980 Ti but I would recommend using a fan controller (or if your motherboard has good fan control then that) to control the speeds on the radiator fan.

I use a VGA to PWM adapter to go from the 980 Ti fan port on the pcb to the fan on my h90. Works like a charm and you can use Afterburner or stock NV software to control fan curve. Retains a motherboard fan header if you are in short supply of them also.

This is what I am talking about. I got mine from NE as shipping was cheaper in case anyone wants to go this route as well.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
I've had my card a couple of weeks now. I guess I'm happy overall (it's so quiet) but my card has an ASIC of 57.5% and can only just manage 1400mhz at stock volts.
Increasing the volts to max gets me around 1430mhz but not sure it's worth it.
 
I use a VGA to PWM adapter to go from the 980 Ti fan port on the pcb to the fan on my h90. Works like a charm and you can use Afterburner or stock NV software to control fan curve. Retains a motherboard fan header if you are in short supply of them also.

I looked into those but couldn't find any available quickly over here. Don't really care as most of my fans are connected to a hardware fan controller. Now the only things on the motherboard headers are the CPU fans, GPU cooler pump and G10 bracket VRM fan as they are pretty much silent. The fan control (excluding for the CPU) on my Gigabyte board is awful.
 
I've had my card a couple of weeks now. I guess I'm happy overall (it's so quiet) but my card has an ASIC of 57.5% and can only just manage 1400mhz at stock volts.
Increasing the volts to max gets me around 1430mhz but not sure it's worth it.

Its apparent that the voltage slider only brings you pain, ie weird throttles, inconsistent voltages and oc failure on clocks that otherwise work on stock volts. I think just forget about that slider and just see what you can get on stock
 
I've had my card a couple of weeks now. I guess I'm happy overall (it's so quiet) but my card has an ASIC of 57.5% and can only just manage 1400mhz at stock volts.
Increasing the volts to max gets me around 1430mhz but not sure it's worth it.

You're not doing so bad.
I have the EVGA SC+ ACX 2.0+, Asic 68.9%

My card hits 1440 at stock and higher volts (maxed at 1.199V), makes no difference because the card is power limited anyway. (Afterburner can tell you this if you add power limit to the graph)
I blew a high power (425W) bios on it and was very soon voltage limited at 1.223V, 1478MHz with a quiet fan of around 53% (I think, cant quite remember).
So I blew a higher voltage (1.25V 425W) bios on it and got up to 1500MHz with a moderately noisy 75% fan.


Afterburner can tell you what is limiting the card, go into its options and have a look at the monitoring section.
I run my card at stock anyway cos I'm only 1080p, the clocking is just for kicks.
What res you at?
 
Its apparent that the voltage slider only brings you pain, ie weird throttles, inconsistent voltages and oc failure on clocks that otherwise work on stock volts. I think just forget about that slider and just see what you can get on stock


What about power limit bar?
 
With stock voltages I managed around 1450 MHz overclock at about 75°C. I'd say that's probably a sweet spot for this card on the stock cooler for speed, temps and noise.

What's your memory clock at?
 
Back
Top