ASUS GTX 980 Ti STRIX DCIII OC Video Card Review @ [H]

I got one of these recently, when gaming, my card goes to 1444mhz and stays there (yet to see it down-clock) and I have left everything on stock, just using the default OC profile. Typically the fan doesn't go much over 1500rpm

I finally ditched my 7950 crossfire rig and went to this, the questions I have are:
1) I notice the GPU clock and memory clock only drop significantly when my monitors are off, should they not have other desktop idle states? - At desktop memory is always at 100% and GPU sits at 963mhz

2) I've never seen the fan turn off, what's the threshold?

3) Only really weird issue I've had with the card, is one time the card somehow got "stuck" at 400mhz or so, and it would not change despite any attempt to OC - had to reboot to fix. Is this the issue mentioned with 1.0.6.9 of the Asus util?

4) Sorry if this is answered elsewhere, but how much of a difference is the memory OC doing? I'm wondering how much of an effect it has on the overall performance.

5) Any idea on the MINOR variance between the Hardocp card and my own at stock settings? just simple asic qual?

settings at idle, you can see the max and min
27wCooM.png
 
I realize the 980 Ti vanilla has been covered in a number of articles already; but I'd like to have those numbers in the same chart when looking at this review.

I found myself searching around for comparison results vs the stock 980 Ti.

This looks like a pretty good buy for the money.
The Fury is a dud, big time....you can't spit-polish that turd.:eek:
 
I got one of these recently, when gaming, my card goes to 1444mhz and stays there (yet to see it down-clock) and I have left everything on stock, just using the default OC profile. Typically the fan doesn't go much over 1500rpm

I finally ditched my 7950 crossfire rig and went to this, the questions I have are:
1) I notice the GPU clock and memory clock only drop significantly when my monitors are off, should they not have other desktop idle states? - At desktop memory is always at 100% and GPU sits at 963mhz

2) I've never seen the fan turn off, what's the threshold?

3) Only really weird issue I've had with the card, is one time the card somehow got "stuck" at 400mhz or so, and it would not change despite any attempt to OC - had to reboot to fix. Is this the issue mentioned with 1.0.6.9 of the Asus util?

4) Sorry if this is answered elsewhere, but how much of a difference is the memory OC doing? I'm wondering how much of an effect it has on the overall performance.

5) Any idea on the MINOR variance between the Hardocp card and my own at stock settings? just simple asic qual?

settings at idle, you can see the max and min
27wCooM.png

Just curious, what settings are custom in the Nvidia panel?
 
I realize the 980 Ti vanilla has been covered in a number of articles already; but I'd like to have those numbers in the same chart when looking at this review.

I found myself searching around for comparison results vs the stock 980 Ti.

This looks like a pretty good buy for the money.
The Fury is a dud, big time....you can't spit-polish that turd.:eek:

yes, that would be my one desire for this article, to show a stock card for comparison
 
Multi LCD setups where it shines. Thats where you would see huge difference since more ram is required.
 
Because it is pointless.

Humm... Maybe you want to decide if the $20-30 premium over the stock 980ti is worth it?

Perhaps you would prefer a blower cooler but want to know the performance delta to make your final decision?

Others in this thread have also requested thus comparison.

Perhaps such a comparison is worthless to you. Congratulations, you have an opinion.

My opinion, since we are sharing, is that your post was pointless and also rude.

I'd give you a hug if I could. I'm sorry your life is so hard Johnny.
 
The fact of the matter is that if we included every GPU that everyone wanted in every single article we would have about 8 GPU reviews a year.

If you cannot read previous reviews and figure out a purchasing decision between the two, then I would suggest there are other websites for you to read.
 
The only problem I can see with this card, is I can get two 970 superclocked cards for the same price and they will run circles around this one card. Granted, with this card you could upgrade to a second one, down the road, but if you are going for bang for buck now, two 970s are the ways to go.
 
The only problem I can see with this card, is I can get two 970 superclocked cards for the same price and they will run circles around this one card. Granted, with this card you could upgrade to a second one, down the road, but if you are going for bang for buck now, two 970s are the ways to go.

what are you talking about? a single 980TI OC'd even at factory OC like this one will be way better than 970s in SLI configuration..
 
As long as the Fury X remains at $649 it is price competitive with the GTX 980 Ti and therefore the proper competition price competitive card to compare with. If you are not happy with the Fury X's positioning, take it up with AMD.

We added something different in this review, that the Gigabyte review did not have, if you noticed or not, 4K results comparing the 980 Ti and Fury X. So there was something new to look at and compare with the Fury X in this review, but it seems you have glossed over that little fact.
 
i have stopped reading 4K results after your Witcher 3 wild hunt test where you concluded that you need a 1000$ card to get the most out of the game @ 4K. Also, you concluded in ather reviews aswell that the currnt 980Ti and FuryX (witch are top tier cards) are not enaugh for 4k gaming in a single setup. I mean that is a reasonable conclusion because it is a bit frustrating to spend that kind of money and you cant max out all the games @ 4K and SLI/Xfire is needed.
 
while we know how a reference 980Ti is better than the FuryX, i would want to know the differences in performance between different 980Ti's. The comparison to the FuryX is irelevant now.

Feel free to look back at our other 980 Ti reviews. So far we have a reference 980 Ti review, a Gigabyte 980 Ti review, and this one. Quite easy to cross compare results. We also make sure to reference past reviews to compare things like overclock achievements etc...
 
Any reason why benchmarking was done at 1440p?

Both the 980 Ti and Fury X are 4K cards. In 1440p both of them become CPU bottlenecked. You aren't using overclocked Skylake i7s either.

I mean, in 1440P it looks like the 980 Ti is between 25-35% faster than a Fury X. I know this is because various driver bottlenecking. But in 4K the CPU part is taken out of the equation.

Be wary of what is "marketed" as a 4K card, and what the actual physical results are.

I can call this cup a duck, but it doesn't make it a duck.

We have performed a specific 4K review on the Fury X, as well as included ap2ap results in reviews comparing the Fury X and 980 Ti at 4K. The Fury X is no way close to being a capable 4K card, the results prove this. The 980 Ti is much faster than the Fury X at 4K, but even it is not really a perfect 4K card either.

Quite frankly, we are adiment that single-GPU video cards are not yet powerful for a good 4K gaming experience. That is possible with multi-GPU for now. There will be a time that single-GPU video cards are good 4K cards, but that time is not today.

At 1440p a more enjoyable experience can be had with both video cards, you are able to maximize most game settings at 1440p with a 980 Ti. For single-GPU gaming, 1440p seems to be the sweet spot according to my testing, and I've done a lot of it.
 
$864 on sale. Just insane. Since I have 2 regular titans, I will wait for Pascal. If it was $649 in Canada I would buy.

Did you know, and I was just testing this the other day, Two Kepler Titans in SLI is only slightly faster than a highly overclocked 980 Ti single card, or TITAN X. Only slightly faster, they are very close in performance. Quite shocking, but it shows how much of a performance improvement Maxwell really was. You could put an overclocked 980 Ti or TITAN X as a replacement for two Kepler TITAN's and it'd only be slightly less performance, probably wouldn't notice it.
 
Solid review as usual, I was surprised by some of the results. I wasn't expecting the card to be boosting to 1430mhz out of the box, I had to take a second look. I also wasn't expecting Asus to factory overclock the ram... I know that 200mhz is pretty negligible, but I'm sure it's enough to snatch a few top spots (vs similarly priced competitors) in stock vs stock benchmarks. It's also nice to see that voltage has made a difference in overclocking the last two 980 ti's that [H] has reviewed. It was starting to look like we were going towards voltage having no added effect (other than increased heat).

I do have one question however... Why doesn't [H] run benches at 1920 x 1080 on high-end, single gpu reviews anymore? I'm sure it was answered in a review sometime in the past, but I can't seem to find it. I know it's hardocp and not softocp and I'm sure time restraints have a lot to do with it... But I bet the majority of users here are still gaming at 1080... Certainly more than at 4k. It also seems like it has been pretty much decided that no single card out there can really handle AAA titles at 4k currently. On top of that, I can think of numerous titles that would benefit from the power a 980 ti has to offer (especially when you're a msaa fanatic like many of us are).

I've got my eye on an Asus PG279Q or Acer X34 Predator (with the original Swift as my fallback if the other two flop), so this question will likely lose relevance with me shortly (did that just answer my own question, lol?), but I was curious and it seems like a few other members are too.
 
Last edited:
Solid review as usual, I was surprised by some of the results. I wasn't expecting the card to be boosting to 1430mhz out of the box, I had to take a second look. I also wasn't expecting Asus to factory overclock the ram... I know that 200mhz is pretty negligible, but I'm sure it's enough to snatch a few top spots (vs similarly priced competitors) in stock vs stock benchmarks. It's also nice to see that voltage has made a difference in overclocking the last two 980 ti's that [H] has reviewed. It was starting to look like we were going towards voltage having no added effect (other than increased heat).

I do have one question however... Why doesn't [H] run benches at 1920 x 1080 on high-end, single gpu reviews anymore? I'm sure it was answered in a review sometime in the past, but I can't seem to find it. I know it's hardocp and not softocp and I'm sure time restraints have a lot to do with it... But I bet the majority of users here are still gaming at 1080... Certainly more than at 4k. It also seems like it has been pretty much decided that no single card out there can really handle AAA titles at 4k curently. On top of that, I can think of numerous titles that would benefit from the power a 980 ti has to offer (especially when you're a msaa fanatic like many of us are).

I've got my eye on an Asus PG279Q or Acer X34 Predator (with the original Swift as my fallback if the other two flop), so this question will likely lose relevance with me shortly (did that just answer my own question, lol?), but I was curious and it seems like a few other members are too.

It's funny, there was another post earlier in the thread that was arguing for why we don't do all our testing at 4K and why do we still test at 1440p. Everyone has an opinion what is "best" to test at.

The way I look at it, we test at what fits best for the card we are testing and the resolution of display a person buying said priced card would be using. If spending $650+ on a video card, most likely (but not always) you've got a decent sized display like 24/27" at 1440p. The kind of person that would spend that much money on a video card isn't going to cheap out on the display.

Now, gamers exist in all forms, from 1080p display, to 1440p, and 4K is on the rise with lower prices. We cannot, and do not, have the time to test every resolution in every review. We must be selective, or we'd only be able to push out one review a month. Keep in mind, it is just 1 person doing the review, not a team of people.

Therefore, we are selective in what price level the card is, and what display size makes sense in that price range and the best overall resolution to get the most out of games for that video card and have an enjoyable gameplay experience.
 
It's funny, there was another post earlier in the thread that was arguing for why we don't do all our testing at 4K and why do we still test at 1440p. Everyone has an opinion what is "best" to test at.

Hahaha, I had just finished reading through the review and hadn't waded through the comments yet when I posted... I'll await the incoming hate from 4k pioneers.

...we test at what fits best for the card we are testing and the resolution of display a person buying said priced card would be using. If spending $650+ on a video card, most likely (but not always) you've got a decent sized display like 24/27" at 1440p...and 4K is on the rise with lower prices...

This makes sense, thanks. I guess Asus and Acer just need to hurry up and release some monitors so I can stop being an outlier then.
 
If they're bottlenecked by the CPU, then the framerates would all be the same. "Hey, we notice these three cards are performing exactly the same, because the CPU can't feed them data fast enough."

If the cards are showing differences, especially drastic differences, then no, they are not CPU bottlenecked.

Also, AMD's terrible marketing aside, what about these cards makes them 4K cards and not 1440P cards? The 980Ti can only just barely provide a playable 4K experience, and the Fury X doesn't even come close. But they do great at 1440P.

Besides, [H]ardOCP is all about providing info on the "real world" experience. What percentage of "real world" players are gaming at 4K?

1) "What percentage of "real world" players are gaming at 4K?"? What percentage of people even buys 650$ worth of GPUs? And you can use downsampling, via DSR or VSR on both parties to simulate 4K conditions.

2) Nvidia is known to dodge the CPU limitation better than AMD. That doesn't mean either party isn't CPU bottlenecked one way or another. Eurogamer's review of the Intel 6600K and 6700K Skylake CPUs showed difference in performance at 4.4 GHz across the last 4 Intel generations using a Titan X at 1440p. So yes, even the 980 Ti is CPU limited at 1440p. Skylake at 4.4 GHz is some 25% faster than Sandy Bridge at 4.4 GHz. At 1440p and using those CPU frequencies, we should only be GPU limited. But the results showed we were still CPU bound.

3) At 4K, the difference between 980 Ti and Fury X is between 2 to 5 fps. So "not even close" is a stretch.
 
Humm... Maybe you want to decide if the $20-30 premium over the stock 980ti is worth it?

Perhaps you would prefer a blower cooler but want to know the performance delta to make your final decision?

Others in this thread have also requested thus comparison.

Perhaps such a comparison is worthless to you. Congratulations, you have an opinion.

My opinion, since we are sharing, is that your post was pointless and also rude.

I'd give you a hug if I could. I'm sorry your life is so hard Johnny.
Wants to spend 650 on a card and worries about 20 bucks lol.
Also I think there are 20+ reviews every where showing the 980 ti reference performance and telling everyone that card runs at 84 C at full load with fans that are noisy. Putting it here again is pointless.

Also Johnny is in my pants and is hard. You can hug him if you are into that sort of a thing.
 
From the first page of the review. From what I have seen from other reviews that take the card apart, only 3 of the heat pipes are contacting the GPU while two are in contact with nothing.

What say you?
To me it looks like the DCUIII was designed for the larger footprint of Fury.

I got one of these recently, when gaming, my card goes to 1444mhz and stays there (yet to see it down-clock) and I have left everything on stock, just using the default OC profile. Typically the fan doesn't go much over 1500rpm

I finally ditched my 7950 crossfire rig and went to this, the questions I have are:
1) I notice the GPU clock and memory clock only drop significantly when my monitors are off, should they not have other desktop idle states? - At desktop memory is always at 100% and GPU sits at 963mhz

2) I've never seen the fan turn off, what's the threshold?

3) Only really weird issue I've had with the card, is one time the card somehow got "stuck" at 400mhz or so, and it would not change despite any attempt to OC - had to reboot to fix. Is this the issue mentioned with 1.0.6.9 of the Asus util?

4) Sorry if this is answered elsewhere, but how much of a difference is the memory OC doing? I'm wondering how much of an effect it has on the overall performance.

5) Any idea on the MINOR variance between the Hardocp card and my own at stock settings? just simple asic qual?

settings at idle, you can see the max and min
  1. If you're running multiple monitors your video card will not idle. NVIDIA's explanation for keeping core clock high with multiple monitors or at 144 Hz refresh rates was to prevent glitches and potential crashing. AMD used to behave the same way, if I'm not mistaken.
  2. For the DCUII STRIX cooler on my GTX 970s it was 60C, IIRC. You can alter the threshold with a custom fan curve if you want to. The thermal threshold may be set lower on the 980 Ti due to its higher power demands, in which case temperature will ramp up more quickly.
  3. Seems to be a symptom of the power bug inherent in Maxwell 2. Not all cards experience the issue. It could be something to do with the tighter power tolerances in both the hardware and drivers.
  4. On my TITAN X the difference between 1750 and 2000 MHz (7000 and 8000 effective, respectively) on the memory alone was worth 8-12 FPS in Dying Light at 2560x1440. That increase in memory speed increases maximum bandwidth by 15% (336 GB/s -> 384 GB/s).
  5. Not all silicon is made equally. Minor variances within manufacturing tolerances are going to result in variations across all chips produced. A review is simply going to give a ballpark idea of what to expect.
1) "What percentage of "real world" players are gaming at 4K?"? What percentage of people even buys 650$ worth of GPUs? And you can use downsampling, via DSR or VSR on both parties to simulate 4K conditions.

2) Nvidia is known to dodge the CPU limitation better than AMD. That doesn't mean either party isn't CPU bottlenecked one way or another. Eurogamer's review of the Intel 6600K and 6700K Skylake CPUs showed difference in performance at 4.4 GHz across the last 4 Intel generations using a Titan X at 1440p. So yes, even the 980 Ti is CPU limited at 1440p. Skylake at 4.4 GHz is some 25% faster than Sandy Bridge at 4.4 GHz. At 1440p and using those CPU frequencies, we should only be GPU limited. But the results showed we were still CPU bound.


3) At 4K, the difference between 980 Ti and Fury X is between 2 to 5 fps. So "not even close" is a stretch.
  1. The people gaming at 4K most likely have several of those $650 video cards. Personally, gaming at an average of 40 FPS is unacceptable to me, so as a result the only games I'm downsampling on my single GM200 are those where I'm approaching the refresh rate limit of my display with acceptable visual settings already.
  2. The additional CPU overhead with AMD's drivers is relevant, in my opinion. It really shows what to expect when buying into a video card. These are real-world tests, so it is valuable information in my opinion as a consumer of product reviews on this website. The fact that the spread in average framerate going from 1440p to 4K drops from 20-40% to 10-15% with the Fury X is not lost on me, however.
  3. The difference of 5 FPS when you're talking about averaging 35 FPS is around 15%. Will you notice it while actually playing games? Probably not. That is why it's important to look at the progression graphs [H] provides and FCAT elsewhere.
 
<---- Newbie here

I noticed some Newegg reviews say it has bad coil whine. Is it just luck of the draw when it comes to any video card these days, if it has coil whine or not?
 
<---- Newbie here

I noticed some Newegg reviews say it has bad coil whine. Is it just luck of the draw when it comes to any video card these days, if it has coil whine or not?

I don't have any issues with mine, I often think power supply contributes to coil whine
 
Be wary of what is "marketed" as a 4K card, and what the actual physical results are.

I can call this cup a duck, but it doesn't make it a duck.

We have performed a specific 4K review on the Fury X, as well as included ap2ap results in reviews comparing the Fury X and 980 Ti at 4K. The Fury X is no way close to being a capable 4K card, the results prove this. The 980 Ti is much faster than the Fury X at 4K, but even it is not really a perfect 4K card either.

Quite frankly, we are adiment that single-GPU video cards are not yet powerful for a good 4K gaming experience. That is possible with multi-GPU for now. There will be a time that single-GPU video cards are good 4K cards, but that time is not today.

At 1440p a more enjoyable experience can be had with both video cards, you are able to maximize most game settings at 1440p with a 980 Ti. For single-GPU gaming, 1440p seems to be the sweet spot according to my testing, and I've done a lot of it.

THIS. Also every gamer has different tastes. Nothing will be a 4k gaming card/setup until I can play the most recent games on maxed settings (minus 234683x AA) while floating around 100-150fps. Hence my 2x MSI GTX 980ti gaming @ 1080p still. I need that awesome input latency or I quickly stop playing since the game isn't doing what I want fast enough.
 
I ordered my STRIX GTX 980 Ti from Amazon about 2 weeks ago and it arrived last Friday. As I understand it, there was an issue with the cooler on the first production run and it took them a little while to get it fixed. Kitguru and Guru3d both reviewed the card a couple of months ago and, IIRC, one observed higher temps than the other so presumably one of them got a faulty cooler. I seem to recall there was a gap between the heat pipes where they contact the GPU. Availability was close to nil until about a month ago but they've been trickling in at Newegg and Amazon since then. So presumably, if there really was an issue with the cooler, it's been corrected now. FWIW, my card idles @ 30c and under a full load ~74c.

I bought my card from Amazon for $641. If you keep an eye on it (via nowinstock.net), they've had it listed at that price on and off over the past couple of weeks. Right now, it looks like it's currently back @ $669. On a side note, Amazon doesn't list the STRIX GTX 980 Ti as being eligible for the Metal Gear Solid V promotion that Nvidia is running through September 8. The Asus website clearly states that the card is eligible for the promotion. Amazon, however, refused to provide me with a code for the game. They did, however, give me a 10% refund on the price of the card so I actually ended up paying ~$575. Just an FYI if you're interested in MGS 5 and planning to order this card from Amazon. Newegg seems to be getting them in a little more frequently than Amazon but I prefer the better return policy at Amazon so for me, it was worth paying sales tax and the slightly longer wait to get the card.

My card doesn't have any coil whine and for those of you who put stock in the ASIC, mine is 68.4. The card is pretty long and heavy. Once I got it installed, there was definitely some sagging noticeable. I was worried enough about it that I ran some fishing line through the end of the cooler and attached it to the top of my case. My card boosts to 1425 out of the box in OC mode. I haven't tried manually overclocking it yet but I expect it to be able to boost somewhere in the 1500 range.
 
Wants to spend 650 on a card and worries about 20 bucks lol.
Also I think there are 20+ reviews every where showing the 980 ti reference performance and telling everyone that card runs at 84 C at full load with fans that are noisy. Putting it here again is pointless.

Also Johnny is in my pants and is hard. You can hug him if you are into that sort of a thing.

I have to agree with the johnny-hugger here.....
I like to see direct comparisons side by side as well. I would like to see not only the plain but also the results of the Gigabyte OC data there too.:D

This ASUS card should be better than the plain 980 Ti, that would be obvious by the branding and the introduction write-up.
But side by side, will it be worth buying?

It's not so much the 20 dollars, as it is there are few to none to be found.

Would it be worth waiting for? Or should the buyer just pick up a standard or perhaps a Gigabyte OC card instead?

PS......I will not go near any johnnys today or in the future.:eek:
 
Any update on this? Thanks.

Yes, they sent me pictures to go along with this, i'll have to upload them somewhere.

Here is the info straight from ASUS

DirectCU is more efficient than copper plate.
Heat Transfer Coefficient (K) of heatpipe is 3000-4000, and copper is only 600, (higher is better)
Although we only contact 3 heatpipe, but the heat transfer is much faster from heatpipe to heatpipe than with Copper Plate.

Optimized in 3 primary heatpipes.
We optimized 3 main heatpipes to deliver 70% of heat away, others 30% of heat will be delivered from secondary heatpipe.

In the pictures the Solder Paste and Copper Plate extend to all heatpipes.

The order looks like this, from the GPU upwards:

GPU (the length of 3 heatpipes)
Thermal Grease (the length of 3 heatpipes)
Copper Plate (the length of all heatpipes)
Solder Paste (the length of all heatpipes)
 
Last edited:
My only comment is that you should include more games in your testing especially when the card is not time critical in terms of its release like the asus strix.

Compared to other websites you have slim number of cards and slim number of games. I would add at least a handful more to the bench.
 
is an additional 5 games even feasible considering this site's methodology? another 5 games is doubling the work load. this site has rarely used more than 6 games in its test suite for as long as i've been visiting. if the editors choose to include mgsv, the patched arkham knight, and mad max, likely two if not three current titles will drop out of the suite. i've done actual play tests of games on the few occasions i post results and its a drag. considering how much attention to detail is expressed here i can't imagine having to do this process with 10 games.
 
It does bemuse me why asus get away with totally bs claims of the cooler on this card, "30% cooler 3x quieter". Since when was 79c even remotely close to 30% cooler than the 83c or thereabouts that the reference card hits? I've seen my card at 83c in extended gaming sessions and the fans up at around 52% or thereabouts which I found a bit irritating.

The 980ti matrix platinum seems to pretty much meet those "30% cooler 3x quieter" claims with reviews indicating it to be in the mid 60's at load overclocked and very quiet while gaming, but the strix to be fair isn't much better than the reference card as far as cooling is concerned. :(
 
My only comment is that you should include more games in your testing especially when the card is not time critical in terms of its release like the asus strix.

Compared to other websites you have slim number of cards and slim number of games. I would add at least a handful more to the bench.


If we did nothing but run canned benchmarks or 30 second in-game run-throughs, we could do that, but we don't and we are not going to do that.

I understand your complaint, but your wishes do not coincide with content production quality standards that we think are paramount for building solid GPU reviews and the timelines we are bound by.
 
Back
Top