• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

980 Ti

Unknown-One said:
I'm using an MSI GTX 980TI GAMING 6G, which comes with a beefy front-plate and VRM heatsinks pre-installed (these can be left on the card when installing the G10 bracket). The fan on the NZXT bracket + the MSI heatsinks seems to do the job nicely.
Click to expand...

Did you install a shim between your gpu and H75? It is probably not making good enough contact if you haven't. The MSI card probably needs a shim like the classified does for example... because of the vrm heatplate that the card has versus the reference model.
 
MSI card does not require a shim but still check if the contact is good.
Secondly plug pump directly to the PSU through 3 pin to molex connector.
Make sure that radiator is exhausting air outside the case in push/pull otherwise it will warm that case up very fast.

With H75 you should be about 55-60 C load.
 
AriciU said:
Did you install a shim between your gpu and H75? It is probably not making good enough contact if you haven't. The MSI card probably needs a shim like the classified does for example... because of the vrm heatplate that the card has versus the reference model.
Click to expand...
The MSI frontplate does not interfere with the H75, no shim is required.

KickAssCop said:
MSI card does not require a shim but still check if the contact is good.
Secondly plug pump directly to the PSU through 3 pin to molex connector.
Make sure that radiator is exhausting air outside the case in push/pull otherwise it will warm that case up very fast.

With H75 you should be about 55-60 C load.
Click to expand...
Contact looks good, I've already re-mounted once to swap thermal paste from NHT1 to MX4.
Pump is plugged directly into my PSU.
Radiator has a single fan (push), and is mounted so that it exhausts hot air out the back of the case.

Still seeing ~85c load temps...

Edit: Double-checked the surface temperature of the radiator and tubing with my IR temp gun. Both report-in as 56c, nowhere near the core temperature of 80c. It really seems like this little H75 simply can't get the heat away from the core fast enough.
 
Last edited:
Very strange. My H55 (same radiator as H75, just comes with one fan instead of 2 fans with H75) is doing 60 C max temperature (in Mad Max but typically 50-55 C in all other games).

You can see picture of my setup posted before. My hunch is that MSI plate really cuts it close so it maybe a contact issue. A shim could do wonders. I know at least a couple of people on OCN that used a shim with MSI cards and got 50-55 C temperatures as well.
 
KickAssCop said:
Very strange. My H55 (same radiator as H75, just comes with one fan instead of 2 fans with H75) is doing 60 C max temperature (in Mad Max but typically 50-55 C in all other games).
Click to expand...
Well, I'm seeing 85c, so I need some ideas... Even taking the side-panel off only dropped my temp 1c, which tells me it's not an airflow problem.

I'll try re-seating it again... I have a sneaking suspicion the MSI backplate is causing problems (there's an air-gap between it and the back of the card's PCB, making it difficult to get proper mount-pressure).

KickAssCop said:
My hunch is that MSI plate really cuts it close so it maybe a contact issue. A shim could do wonders. I know at least a couple of people on OCN that used a shim with MSI cards and got 50-55 C temperatures as well.
Click to expand...
I really don't know why I'd use a shim. The MSI frontplate has a huge hole in the middle around the GPU, there's no way it's interfering...

I ran my GTX 780, with an EVGA front-plate, without a shim, as well. All you have to do to fix interference is snap-off the four posts on the plate that extend into the GPU area.
 
Last edited:
The only other thing can be the pump is not functioning properly. This similar issue was happening to another guy on evga forums but using the hybrid. He tried all the things you have tried and was now asked to just do an RMA. Since you have a Corsair pump, RMA shouldn't be an issue at all.
 
KickAssCop said:
The only other thing can be the pump is not functioning properly. This similar issue was happening to another guy on evga forums but using the hybrid. He tried all the things you have tried and was now asked to just do an RMA. Since you have a Corsair pump, RMA shouldn't be an issue at all.
Click to expand...
I've already RMAed this cooler several times... They couldn't seem to send me one that didn't have horrific pump noise and/or unstoppable gurgling (Every H75 they sent me was louder than the the stock cooler at idle, which defeated 50% of the point of the H75)

If it turns out THIS one has a bad pump, I probably wont bother going Corsair ever again. How's the pump noise on NZXT AIO coolers?
 
Unknown-One said:
I've already RMAed this cooler several times... They couldn't seem to send me one that didn't have horrific pump noise and/or unstoppable gurgling (Every H75 they sent me was louder than the the stock cooler at idle, which defeated 50% of the point of the H75)

If it turns out THIS one has a bad pump, I probably wont bother going Corsair ever again. How's the pump noise on NZXT AIO coolers?
Click to expand...

the same as similar corsair models as they are both made by Asetek..

Pump noise in those units are always due to high movement of the unit in the installation and normally should take just a few days to fix as the bubbles get moved off, for this type of issue is always good to place the radiator with the reservoir side up so the bubbles will not interfere and will be accumulated in that place.

I would take advice of the shim.. a proper mounted shim tend to fix issues with AIO. but also you can simply test that cooler in another machine or another card and see how it behave. it shouldn't take too much time and can save a RMA process.. in the other hand i would check the revision of the Bracket, I know that the first revision the G10 have issues with the contact and with the screws, being too short make the false illusion that the cooler are making good contact however the contact is poor, If remember correctly you have just to contact NZXT and they will sent you a revised unit that may fix your problem.
 
Last edited:
Unknown-One said:
I've already RMAed this cooler several times... They couldn't seem to send me one that didn't have horrific pump noise and/or unstoppable gurgling (Every H75 they sent me was louder than the the stock cooler at idle, which defeated 50% of the point of the H75)

If it turns out THIS one has a bad pump, I probably wont bother going Corsair ever again. How's the pump noise on NZXT AIO coolers?
Click to expand...
I use an NZXT Kraken X31 on my CPU and I can't hear it at all when my system is idling. There is the usual noise on system startup, but that is it. I would recommned using a different fan from the included one, though, like a Noctua NF-F12.
 
No pump noise with my two x41's also. I have the radiators mounted up top of the pumps if that counts for anything.
 
Unknown-One said:
I've already RMAed this cooler several times... They couldn't seem to send me one that didn't have horrific pump noise and/or unstoppable gurgling (Every H75 they sent me was louder than the the stock cooler at idle, which defeated 50% of the point of the H75)

If it turns out THIS one has a bad pump, I probably wont bother going Corsair ever again. How's the pump noise on NZXT AIO coolers?
Click to expand...

You re-seated it, but did you check the contact? Even though the frontplate whole is huge, it still may be not contacting properly.
 
Is the rad hot to the touch? If you got good contact then that rad should be getting really warm if the core is hitting 85C. If it's not hot then you know it's not removing the heat efficiency, hence poor contact.
 
KickAssCop said:
Very strange. My H55 (same radiator as H75, just comes with one fan instead of 2 fans with H75) is doing 60 C max temperature (in Mad Max but typically 50-55 C in all other games).

You can see picture of my setup posted before. My hunch is that MSI plate really cuts it close so it maybe a contact issue. A shim could do wonders. I know at least a couple of people on OCN that used a shim with MSI cards and got 50-55 C temperatures as well.
Click to expand...

I was pissed to find out that the H75 was a thin rad after I ordered it. Makes no sense.

Wasn't a big deal since it came with two decent fans but still. I really wish that they still sold the old H70 core. I have enough good 120mm fans kicking around.
 
Killerxp100 said:
Is the rad hot to the touch? If you got good contact then that rad should be getting really warm if the core is hitting 85C. If it's not hot then you know it's not removing the heat efficiency, hence poor contact.
Click to expand...
I mentioned earlier that the radiator's surface-temperature under load is 56c (133 F). The air coming out the back of the PC is HOT, but nowhere near 85c.

I just quadruple-checked, and the plate doesn't interfere with the H75 in any way. There's no way it's causing the problem I'm experiencing.
 
Looks like it's a bad pump. I tapped it with my fingernail while running a benchmark, heard the rush of bubbles, and my temps suddenly dropped by 15c...

About 10 minutes later, temps shot back up again.

Sooo... suggestions on a decent replacement? I've RMAed so many Corsair H55's and H75's for pump / noise issues, and I really don't want to bother with it again...
 
Unknown-One said:
Looks like it's a bad pump. I tapped it with my fingernail while running a benchmark, heard the rush of bubbles, and my temps suddenly dropped by 15c...

About 10 minutes later, temps shot back up again.

Sooo... suggestions on a decent replacement? I've RMAed so many Corsair H55's and H75's for pump / noise issues, and I really don't want to bother with it again...
Click to expand...
NZXT X31
 
is there someone who upgraded from GTX980 to GTX980 Ti?
how good is the upgrade?
is there some games that benefits of the more RAM?
 
sblantipodi said:
is there someone who upgraded from GTX980 to GTX980 Ti?
how good is the upgrade?
is there some games that benefits of the more RAM?
Click to expand...
Why do you have to ask? Look at reviews and see for yourself as that cuts through the BS that some people spread. It was worth it for me as I like to crank all the settings I can at 1440 and even 4k in some games. Vram usage is never simple topic to discuss as its hard to tell what a game really needs since most allocate more vram if you have it.
 
I usually find anandtechs benchmark database helpful for comparisons like this.

The charts are shiite, but the data is good.


From theoretical figures based on number of cores and clock speeds a non-overclocked 980Ti should be ~22% faster than a non-overclocked 980.

980 Ti's seem to be overclocking pretty well though, especially with aftermarket cooling. ('ve never had a 980 though, so I can't compare how well they overclock.

The 980Ti at $650 is the best deal in high performance GPU's IMHO. It's what I would go with hands down if I were GPU shopping (and I have two of them, so that's what I did :p )

If I already had a 980, I'm not sure. I'd upgrade if - and only if - you are not getting the performance you want out of your 980 in titles you are currently playing, and a 20% or less improvement is enough to put you across the border into perfect vsync land of 60fps (or 120fps if that's your thing)

Then I'd look at what I could sell my 980 for, what the 980ti I want is going for, and decide if that 22% increase (probably more, but overclocking is never a guarantee) is worth the money to me, or if I'd be better off waiting 9 months or so for Pascal, which promises to be a pretty large leap.
 
Zarathustra[H];1041873434 said:
I usually find anandtechs benchmark database helpful for comparisons like this.

The charts are shiite, but the data is good.


From theoretical figures based on number of cores and clock speeds a non-overclocked 980Ti should be ~22% faster than a non-overclocked 980.

980 Ti's seem to be overclocking pretty well though, especially with aftermarket cooling. ('ve never had a 980 though, so I can't compare how well they overclock.

The 980Ti at $650 is the best deal in high performance GPU's IMHO. It's what I would go with hands down if I were GPU shopping (and I have two of them, so that's what I did :p )

If I already had a 980, I'm not sure. I'd upgrade if - and only if - you are not getting the performance you want out of your 980 in titles you are currently playing, and a 20% or less improvement is enough to put you across the border into perfect vsync land of 60fps (or 120fps if that's your thing)

Then I'd look at what I could sell my 980 for, what the 980ti I want is going for, and decide if that 22% increase (probably more, but overclocking is never a guarantee) is worth the money to me, or if I'd be better off waiting 9 months or so for Pascal, which promises to be a pretty large leap.
Click to expand...
How are you possibly calculating only 22% performance difference from the specs alone?

The 980 Ti has almost 38% more cores and tmus and the real world boost difference is only about 6-8% as a typical 980 runs at around 1240 while a typical 980 Ti seems to run around 1150-1175 on average. And then it has 50% more rops only running a typical 8% lower speed along with a 50% memory bandwidth increase. How you possibly came up with 22% based on the specs is puzzling.

Now of course some games will not be fully gpu limited so actual real world performance may only be around 25% difference in many titles.

Hardwarecanucks had a 980 Ti that seem to stay just over 1200 during testing which is a few % higher than what I see in most reviews. It ends up being quite a bit faster than a 980 here.


photo hosting sites
 
I had one of those Herculez Airboss nutter 980s, that was good for 1600MHz.
But it became unstable at stock clocks after a while so I sent it back.
My 980ti is a fair bit quicker even though it clocks to 1456MHz.
 
Michaelius said:
I went 970->980 ti and difference was almost generation jump worthy (50%+)
Click to expand...
Just about. I can attest to this, as going from 970 SLI to a single Titan X offers about the same level of performance in most games and better in some, while having the benefits of running a single card.
 
I went from 970 SLi to a single 980 ti for a bit. 980 ti was within 5-7% of 970 SLi performance. All overclocked of course. Adding a second 980 ti just put my system over 5000.
 
misterbobby said:
How are you possibly calculating only 22% performance difference from the specs alone?
Click to expand...

Going off of official specs.

2048 cores @ 1216Mhz boost vs 2816 cores @ 1076Mhz boost.

True the TI sees higher clocks than that in typical use, but so does the 980...
 
Is there someone who have a gtx980 ti acx 2.0 from evga non overclocked and can tell me what is the GPU boost frequency with that card?
As far as I know since it runs cool it should boost more than the 1075mhz advertised by nvidia.
 
That's the base clock, boost is at 1150mhz or something like that.

My 980ti acx without touching anything was boosting in the 1250-1300 range If my memory serves me right. Power limit will be the thing holding you back on a stock bios.
 
Killerxp100 said:
That's the base clock, boost is at 1150mhz or something like that.

My 980ti acx without touching anything was boosting in the 1250-1300 range If my memory serves me right. Power limit will be the thing holding you back on a stock bios.
Click to expand...
No, 1076 is the advertised reference boost clock. But yeah around 1150-1175 or so should be actual game boost.

You should not be hitting 1250-1300 on reference clocked card so your memory is serving you wrong for sure.
 
With reference clocks if I kept the temps really low I could reach 1200 something boost clock but not higher. But as soon as the temp hit around 70+ it would start dropping slowly though.
 
kalston said:
With reference clocks if I kept the temps really low I could reach 1200 something boost clock but not higher. But as soon as the temp hit around 70+ it would start dropping slowly though.
Click to expand...
Yes, a cold card is always going to Boost higher than in normal operation. It's why Brent advocates for warming up the cards first to give real word Boost clocks in their reviews.
 
Which is why my Prolimatech MK-26 has given me a noticeable performance boost without even overclocking (well it's not huge but in demanding games it's a little extra regardless) :) That was actually a surprise to me since I kept reading left and right that Maxwell doesn't throttle before 84c. Guess it's a matter of wording but in my book if it's not reaching max boost clock it's throttling...

On Furmark it downclocks a lot though even before temps rise. But I think it just stresses the card in a different way since temps do get hot and power is over 100%.
 
Last edited:
My card begins throttling at 56C. That's why I BIOS modded it not to do it. Also killed off the power limit. Obviously at your own risk but I've never heard of anything negative happening.

In my experience my card 980 / Titan X hit 1550 / 1525 BIOS modded, watercooled, overvolted. My normal operating freq is about 100 below that just to make sure crashes never happen. I can't give the wife a reason to bitch. :).
 
Dayaks said:
My card begins throttling at 56C. That's why I BIOS modded it not to do it. Also killed off the power limit. Obviously at your own risk but I've never heard of anything negative happening.

In my experience my card 980 / Titan X hit 1550 / 1525 BIOS modded, watercooled, overvolted. My normal operating freq is about 100 below that just to make sure crashes never happen. I can't give the wife a reason to bitch. :).
Click to expand...


What did you modify/where can you see the throttling occurring if you don't mind me asking?
 
Mine drops by 15 MHz @ 72C (1469 to 1455) and then stays there.
Seems like an odd decision.

DhU9CUF.png
 
You must log in or register to reply here.
Back
Top