Guru3D 970 SLI review

[SirPauly]

It's new and hopefully matures -- nVidia states they are committed to smoothness and created FCAT -- curious to see more scaling and smoothness moving forward.

Going to wait for maturity -- DSR for Sli and really looking forward to MFAA for efficiency.

AT 329 -- GTX 970 offers impressive current value and potentially more as it matures, imho.

 

[Skurge]

And then we take a vacation from the illusion that we live in a perfect world, and come back to reality. New architecture. Normally, that would be all a person needs to hear. Normally. By your logic, no GPU should ever have an issue again. Do agree that will not happen?
I'm all for FCAT testing in all GPU reviews. I think its a fantastic and extremely useful testing method.

I don't agree with not bringing the issue up. You seem to be fine with sitting back and letting your $700 to $1100 purchase perform sub par and not say anything about it. This is an old issue that both vendors have eliminated. It should not be happening. If it were something we haven't seen before, then sure, understandable.

 

[SirPauly]

There are differing point-of-views:

For me, appreciate the awareness so the important players may address and improve their products for their customers.

 

[3DVagabond]

There are differing point-of-views:

For me, appreciate the awareness so the important players may address and improve their products for their customers.

I'm curious if nVidia will own up and admit it.

 

[Keysplayr]

I don't agree with not bringing the issue up. You seem to be fine with sitting back and letting your $700 to $1100 purchase perform sub par and not say anything about it. This is an old issue that both vendors have eliminated. It should not be happening. If it were something we haven't seen before, then sure, understandable.

Completely unreasonable. Are you trying to tell me that you believe that an issue fixed with one driver was never re-broken with a subsequent driver? And yes even on the same architecture? If you're sticking to your guns on this falsehood then we don't have anything further to talk about. I don't do unreasonable. Come up with a reasonable argument and I'm all ears.

 

[raghu78]

AMD got bad press with an already long in the tooth arch when TR unleashed it's frame latency/pacing review methods article. Difference is, Nvidia will improve this on their own. You'd have to hammer AMD for months and maybe years for that to happen.

AMD got bad press and deservedly so. But sometimes the solution to such complex problems is not something you conjure up in a few days or weeks. In fact it could be months and maybe even a year before you have a truly consistent solution like the XDMA based frame pacing in Hawaii. btw its without a doubt that XDMA CF is superior to Nvidia SLI .

http://hardocp.com/article/2014/07/...r9_290x_4gb_overclocked_at_4k/11#.VClUfMlh71U

"Today we can confidently state that AMD CrossFire is going to give you the better gameplay experience than SLI at 4K. "

And then we take a vacation from the illusion that we live in a perfect world, and come back to reality. New architecture. Normally, that would be all a person needs to hear. Normally. By your logic, no GPU should ever have an issue again. Do agree that will not happen ? I'm all for FCAT testing in all GPU reviews. I think its a fantastic and extremely useful testing method.

Nobody is saying that. What people don't want is a launch driver having significant latency issues. This latency issue was an issue both Nvidia and AMD had resolved to a great extent. Nvidia seems to have regressed here. There was no pressure on Nvidia to rush the GM204 launch.

 

[Keysplayr]

I'm curious if nVidia will own up and admit it.

Yes I think you'll see them turning themselves in at the Santa Clara police precinct with arms outstretched and ready to be cuffed.

 

[Keysplayr]

Nobody is saying that. What people don't want is a launch driver having significant latency issues. This latency issue was an issue both Nvidia and AMD had resolved to a great extent. Nvidia seems to have regressed here. There was no pressure on Nvidia to rush the GM204 launch.

There are newer drivers. Do you think they should be tested? And by more than one review site? Especially TR?

 

[davie jambo]

I would quite like to see some benchmarks for these cards with newer games. Is Hitman relevant these days ? I don't think so. Having said that at least they did not bench skyrim and starcraft 2

I would like to see how these cards do in Watchdogs , Dead Rising 3 and Wolfenstein all at 4K instead of some of the older games

 

[sxr7171]

Early drivers, fully expect this to be fixed.

Not sure about sli being tweaked to the point where frame delivery is BETTER than single-gpu, though.

I'll believe that when I see it. Ha ha! Smoother than single card? That's a huge assertion.

 

[guskline]

I own 2 R9 290s in CF in the rig below. I previously had a GTX780 Classified, and due to the fabulous prices on the R9 290s in the used market and the unbelievably high prices on the GTX 780 Classified, I opted to go CF. This was shortly before the release of the GTX 970. Had I known the price (incredible at @$350), I may have sold the single R9 290 I had. Why?

Drivers. The Nvidia drivers are much easier for me to work with. I have opted to stay with the 290s since I just bought waterblocks for them and it's hard to recoup the $$.

I'm learning that ULPS should be disabled in CF (had a few bluescreens). I also learned that Nvidia works much better in SLI than AMD in CF when playing COD Ghosts.

In most games the 290CF power really shows. I have a 2560x1440p monitor so the 290 CF might give me a bit of protection if I eventually opt for a 4k monitor.

I bring up this info to help keep perspective. The GTX970 JUST came out and is really a GREAT card for the price.

The 4k issue ( or lack thereof) is most likely due to the 334 drivers. The 341 drivers need to be tested in 4k.

I have a 3770k rig with 2 GTX670 FTWs in SLI and I keep the latest drivers on it. I also run the GTX 780 Classified in a FX8350 rig.

I'm appreciative of the brute power of the 290s in CF but the Nvidia cards, and especially their drivers feel more refined. I suspect the GTX 970 SLI will follow suit.

 

[Gikaseixas]

The Nvidia drivers are much easier for me to work with

Gus, try using SLI and report back. I'm sure a single R9 290 gives you very a stable experience.
I had CF before, used SLI once too and i'm not looking forward to use one or the other soon.

I'd say it's early but I see nothing wrong with people addressing it. It clearly pushes a few guys button's but they need to leave the green shades at home and stop defending their team at any cost.

 

[sxr7171]

I would quite like to see some benchmarks for these cards with newer games. Is Hitman relevant these days ? I don't think so. Having said that at least they did not bench skyrim and starcraft 2

I would like to see how these cards do in Watchdogs , Dead Rising 3 and Wolfenstein all at 4K instead of some of the older games

Seriously. Why do they cling to irrelevant games? And I like Hitman Absolution.

 

[notty22]

Some people should actually read the reviews before commenting. It's not difficult.

In closing

I love the 970 in 2-way SLI at 2560x1440, I really do. Low noise, low heat and low power consumption are the icing on top of this cake. I mean two GeForce GTX 970 cards in SLI consume roughly 326 Watts, add to that the rest of your system (processor, chipset, peripherals) at say 200 Watts and you'll find yourself in the 500 Watt power consumption region. Have you ever noticed a powerful SLI system to consume that little ?
.....


The framedrops in FCAT we are still investigating. Rest assured you can not see/detect these framedrops yourself. Hence we think it could be an issue with our FCAT system.
So yeah, time to wrap thing up I say. Two GeForce GTX 970 cards set up in SLI will still cost you roughly 650 EURO, for that money you'll gain a incredible performance kicking your configuration into the highest gear performance wise. We'll admit it though, if you are on a single monitor, there's just no need just yet to get a second card, but at this price a SLI configuration hits a very nice sweet spot. With Maxwell the thermals and power consumption are no problem either. So at 2560x1440 if you want to build something bad-ass... why not eh ?

The experience was obviously good, they even question whether the fcat results, that seems strange, are from faulty equipment.

 

[Keysplayr]

Gus, try using SLI and report back. I'm sure a single R9 290 gives you very a stable experience.
I had CF before, used SLI once too and i'm not looking forward to use one or the other soon.

I'd say it's early but I see nothing wrong with people addressing it. It clearly pushes a few guys button's but they need to leave the green shades at home and stop defending their team at any cost.

That's not what's happening here at all. Of course it should be addressed as any problems should, but to feign outrage over it is a different story and that is exactly what's going on.

 

[Keysplayr]

Gus, try using SLI and report back. I'm sure a single R9 290 gives you very a stable experience.
I had CF before, used SLI once too and i'm not looking forward to use one or the other soon.

I'd say it's early but I see nothing wrong with people addressing it. It clearly pushes a few guys button's but they need to leave the green shades at home and stop defending their team at any cost.

That's not what's happening here at all. Of course it should be addressed as any problems should, but to feign outrage over it is a different story and that is exactly what's going on.

And: From article:
"Hence we think it could be an issue with our FCAT system."

Evidently the performance was so smooth that Guru is questioning their equipment.

Any comments?

I've also emailed Tech Report to see if they can do a follow up with this.

 

[Gikaseixas]

Evidently the performance was so smooth that Guru is questioning their equipment.

Any comments?

Multi GPU end user experience for both manufacturers is lacking overall but in the end CF provides a better experience. We have several reviews now showing that.

Guru final comments won't fix broken (immature) drivers, there's work to do

 

[guskline]

Gus, try using SLI and report back. I'm sure a single R9 290 gives you very a stable experience.
I had CF before, used SLI once too and i'm not looking forward to use one or the other soon.

I'd say it's early but I see nothing wrong with people addressing it. It clearly pushes a few guys button's but they need to leave the green shades at home and stop defending their team at any cost.

I have used SLI for quite sometime with my 670s, and it has been very good. So far I've had problems in Crossfire with ULPS but I think that has been solved. I have problems with COD Ghost in CF but that appears to be COD Ghost!

Other games work well with the CrossFire.

I also agree with Keysplayer that time should address the driver issues for the GTX970 in SLI.

As for the R9 290s, they have been out for quite awhile.

Gikaseixas, no doubt that the R9 290 is stable in single setups. I have it in CF and hope to address any problems. So far it runs well in Crysis3 and BF4. It isn't supported yet in Il2 BOS but it doesn't have either SLI or CF support as of yet.

 

[Attic]

I have used SLI for quite sometime with my 670s, and it has been very good. So far I've had problems in Crossfire with ULPS but I think that has been solved. I have problems with COD Ghost in CF but that appears to be COD Ghost!

Other games work well with the CrossFire.

I also agree with Keysplayer that time should address the driver issues for the GTX970 in SLI.

As for the R9 290s, they have been out for quite awhile.

Gikaseixas, no doubt that the R9 290 is stable in single setups. I have it in CF and hope to address any problems. So far it runs well in Crysis3 and BF4. It isn't supported yet in Il2 BOS but it doesn't have either SLI or CF support as of yet.

I never needed to turn of ULPS in my crossfire setup. Heard folks on both sides of the issue. I had blue screen issues that I tied back to a power setting in BIOS that now eludes me. These occurred outside of games.

You may consider turning off HPET in bios. Or rather looking into it as a consideration for issues with mGPU setups. I came across that numerous times while looking into mGPU performance/smoothness considerations.

 

[iiiankiii]

That's not what's happening here at all. Of course it should be addressed as any problems should, but to feign outrage over it is a different story and that is exactly what's going on.

And: From article:
"Hence we think it could be an issue with our FCAT system."

Evidently the performance was so smooth that Guru is questioning their equipment.

Any comments?

I've also emailed Tech Report to see if they can do a follow up with this.

I thought this was addressed earlier. Some people can detect the frame drops. Some can't. So what do we trust; the FCAT system or the subjective opinion of the reviewer? These were the exact same argument raised with AMD's frame pacing situation. Some people couldn't tell if there were stutters even when the data suggested otherwise. So I guess those were equipment failures, too?

 

[Abwx]

I thought this was addressed earlier. Some people can detect the frame drops. Some can't. So what do we trust; the FCAT system or the subjective opinion of the reviewer? These were the exact same argument raised with AMD's frame pacing situation. Some people couldn't tell if there were stutters even when the data suggested otherwise. So I guess those were equipment failures, too?

He could have checked by repeating the test or using other cards to see if his FCAT system was working accordingly, wich was surely done but since the results were likely the same he didnt find anything else than questionning the very set up he used to trash AMD, as you point it at the time he didnt express the slightest doubt about his measurements being right or wrongs...

 

[Keysplayr]

I thought this was addressed earlier. Some people can detect the frame drops. Some can't. So what do we trust; the FCAT system or the subjective opinion of the reviewer? These were the exact same argument raised with AMD's frame pacing situation. Some people couldn't tell if there were stutters even when the data suggested otherwise. So I guess those were equipment failures, too?

You guess those were equipment failures to?
Why? What logic presented itself to you in this way?

Reasonable doubt has been introduced by the very entity that did the test in the first place.
What's next? I'm glad you asked. We contact other sites with FCAT testing capability (I've already asked TR) to run the same tests and get even more extensive with it.

I think that is more reasonable than people here begging for this to be true and accurate, which makes even less sense in light of Guru's conclusions.

 

[Erenhardt]

This is like mirror reality from a year ago. Arguments stayed the same, but sides changed.

I remember there was the problem of drawing the line between good enough and stutter states back in CFX framepacing issues era . I guess we reached good enough point right about now, right?

It's new and hopefully matures -- nVidia states they are committed to smoothness and created FCAT .

I though so. It was quite obvious that PCPer=nvidia ninja marketing ^^ Thanks for clearing that up.

 

[Keysplayr]

This is like mirror reality from a year ago. Arguments stayed the same, but sides changed.

I remember there was the problem of drawing the line between good enough and stutter states back in CFX framepacing issues era . I guess we reached good enough point right about now, right?

Not exactly. Scott Wasson never suggested that his testing equipment might have a problem. Huge difference.

 

[Subyman]

Some people should actually read the reviews before commenting. It's not difficult.



The experience was obviously good, they even question whether the fcat results, that seems strange, are from faulty equipment.

They'd rather argue haha. If they said the experience was great, then who cares about a test?