Aftermarket OC 290/290X vs 970 benches?

[crisium]

It's frustrating not finding a true comparison between the 970 and 290 series. Most new reviews still use reference 290 series. No one ever seems to overclock both. You are left with trying to reason that a reference 290 throttles so the aftermarket will be better vs. 970 should overclock higher.

Where's the true comparison of OC aftermarket 290/290X vs OC 970 in games released in late 2014-2015? Does no one compare what really matters?

 

[Eric1987]

4k the 290x wins any lower its a tie/970

 

[crisium]

No, see it's not that simple.

http://gamegpu.ru/images/remote/htt...rds-game_2014-video-test-games__2014_1920.jpg
Source:http://gamegpu.ru/test-video-cards/igry-2014-goda-protiv-sovremennykh-videokart.html

10 games from late 2014 and even the 290X wins at 1080. I assume that's reference too. But surely the 970 OC's higher?

But Tom's still has the 970 cleaning up at 1080:
http://www.tomshardware.com/reviews/nvidia-geforce-gtx-960,4038-12.html

Do you think this one throttles too much?

Even latest TPU has the 290X only 2% slower at 1080, but 2% faster at 1440:
http://tpucdn.com/reviews/ASUS/GTX_980_Matrix/images/perfrel_1920.gif
http://tpucdn.com/reviews/ASUS/GTX_980_Matrix/images/perfrel_2560.gif

Inconsistencies everywhere and I really think it's due to throttling. I want aftermarket battles with OC thrown in for good measure.

 

[notty22]

4k the 290x wins any lower its a tie/970

Hocp has done gpu reviews where they test aftermarket 290x's VS 980 and sometimes 970. You can at least get some data points there.

IMO, you won't throttle with aftermarket 290x's, but there is less o/c room, compared to the 970/980. There are exceptions before people freak, on the AMD side. Like the Lightning or Vapor-x or Tri-x, while any Maxwell does about 1500 mhz. Of course no o/c is guaranteed.
http://www.hardocp.com/reviews/gpu_video_cards/

 

[Jhatfie]

I have not tested my MSI 290X directly against a 970, but I did test it against my 980 that I had before I sold it. The Strix 980 was OC'd to 1480/7800 and my 290X to 1140/6000 (if I remember correctly) which should be average OC's for both. At 2560x1600 across 8 game tests my 980 was only 8.5% faster on average with gamework title games like FC4, AC Unity and Bioshock being about 14% faster at the same settings. Games like Shadow or Mordor and Dragon Age Inq were just about a wash. I think clock for clock a 970 is about 10-15% slower than a 980.

 

[Udgnim]

most aftermarket 290s clock ~1050 out of the box

OC around ~1150 is a realistic expectation without too much of a massive increase to voltage offset

there's a wall at around >=1200 where you need to feed more and more voltage for smaller and smaller amounts of OC increase imo

 

[Headfoot]

http://www.techspot.com/review/898-geforce-gtx-970-sli-4k-gaming/

Pretty good, but still not 100% comprehensive.

This one is even more recent, using a Gigabyte 290/x aftermarket, but doesn't include head-to-head overclocking.

http://www.techspot.com/guides/912-best-graphics-cards-2014/page2.html

Both the above have frame times, which is awesome.

 

[Silverforce11]

Hawaii peaks out at 1.2ghz, and to get there you often need to give it +50mV to even +100mV, that causes about 100W extra in power consumption. It's not worth it to push it via high OCs. I run at 1.1ghz without vcore and power use increase by around 20W. That isn't much of an OC, ~10% for most aftermarket 290.

970s tend to get in-game boost into the 1.3ghz range out of the box, or higher for aftermarket models. It can be regularly pushed to 1.5ghz. On a percentage term, it doesn't actually have an edge. R290/X going from 1ghz to 1.2ghz is a 20% OC. Going from 1.3 with a 20% OC would require 1.56ghz. The problem with Hawaii OC is if you chase that 20% OC, you do so at a massive power increase that isn't worth it. So realistically its a 10% OC vs a 20% OC on 970/980s.

Under than scenario, I would say a 1.56ghz 970 is faster at normal resolution than a 1.1ghz R290X.

 

[Leyawiin]

I don't know why anyone worries about hair-splitting differences between the two. Unless your games are heavily AMD or Nvidia favored or you obsess over electric bills it really doesn't matter much what you get. None of them are stinkers and produce similar results. I know the raison d'être for hardware forums is too discuss minutiae to death, but there's just not much to separate those two performance or price wise.

 

[lehtv]

I agree with Leyawiin here. It's somewhat ironic what the crisium says: "Does no one compare what really matters?" ... That is exactly what sites do, they compare what matters. They're not splitting hairs over minute differences.

 

[BonzaiDuck]

I don't know why anyone worries about hair-splitting differences between the two. Unless your games are heavily AMD or Nvidia favored or you obsess over electric bills it really doesn't matter much what you get. None of them are stinkers and produce similar results. I know the raison d'être for hardware forums is too discuss minutiae to death, but there's just not much to separate those two performance or price wise.

I just haven't had a big graphics-OC and gaming focus until now, or not since about five years ago. These new cards are really big performers compared to my GTX 570 or my GTX 8800 of 7 years ago.

You can see the performance benchmark comparisons. It isn't necessarily a better strategy to get only the top performer, and there are other features or factors of relevance. With the comparisons, at least you have an idea of what you're getting. If it's a matter of 10 to 15% between two single cards, your interest in that 10 to 15% may vary.

 

[ocre]

I don't know why anyone worries about hair-splitting differences between the two. Unless your games are heavily AMD or Nvidia favored or you obsess over electric bills it really doesn't matter much what you get. None of them are stinkers and produce similar results. I know the raison d'être for hardware forums is too discuss minutiae to death, but there's just not much to separate those two performance or price wise.

+9000

The difference between a 980 and a 970 is larger than both the 970 vs 290 and 290x vs 970.

I will tell anyone who will listen, these cards are so close its unlikely anyone would be able to tell them apart while gaming.

Sometimes the 980 can be 20% faster than the 970 but most of the time it is less than that. That 20% is very very small in reality. I can't believe how much we argue over smaller percentages than that. Even 20%, its not easily noticed. Most of the time it takes a fps overlay or benchmark scores to tell it.

The 970 and 290(x) are close enough it just doesn't matter.

 

[Headfoot]

I don't know why anyone worries about hair-splitting differences between the two. Unless your games are heavily AMD or Nvidia favored or you obsess over electric bills it really doesn't matter much what you get. None of them are stinkers and produce similar results. I know the raison d'être for hardware forums is too discuss minutiae to death, but there's just not much to separate those two performance or price wise.

You're right on every point but price. $270 for 290 vs $330 for 970. $60 is a pretty decent amount of cash

 

[Geforce man]

Or if you snag one of the sales and you get a r9 290 for 200$, then its a 130$ difference!

 

[boozzer]

Hocp has done gpu reviews where they test aftermarket 290x's VS 980 and sometimes 970. You can at least get some data points there.

IMO, you won't throttle with aftermarket 290x's, but there is less o/c room, compared to the 970/980. There are exceptions before people freak, on the AMD side. Like the Lightning or Vapor-x or Tri-x, while any Maxwell does about 1500 mhz. Of course no o/c is guaranteed.
http://www.hardocp.com/reviews/gpu_video_cards/

base on the fps numbers with crysis 3, one of the few fair games for benchmarks, 290x looks fan @#$%ING tastic.

 

[Gikaseixas]

The 970 and 290(x) are close enough it just doesn't matter.

This, the deciding factor would be vendor specific features and bugs hahahahaha, now seriously, it's close between these cards, just check what features would you like to have, what games you play and such