[Hardcorp] GALAXY GTX 660 Ti GC OC vs. OC GTX 670 & HD 7950

[Final8ty]

http://www.hardocp.com/article/2012/08/23/galaxy_gtx_660_ti_gc_oc_vs_670_hd_7950

 

[brandon888]

http://www.hardocp.com/article/2012/08/16/galaxy_geforce_gtx_660_ti_gc_3gb_video_card_review/5


there is a only 2-3 fps difference in bf3 ????? is this real ???? i mean minimum fps and average ;/

 

[Silverforce11]

Pretty fail no MSAA at 1600p.

LOL @ conclusion: "It is the GPU overclock that provides the most performance on Radeon HD 7950 and it takes near 1.2GHz for the Radeon HD 7950 to shine in this comparison. Unless you have a high-end custom video card with voltage control and great cooling, you won't be able to get the HD 7950 up to these clock levels most likely. We've had some poor overclocks on HD 7950 cards in the past, so it takes a video card with enthusiast performance in mind to get this level of gameplay. Our HD 7950 overclock represents pretty much the best performance you can get out of a custom built Radeon HD 7950 and we just want everyone to be clear on this."

Farking pure fail there [H].

You didn't even reach 1.2ghz on the core and you think its the "best performance you can get out of a custom built Radeon HD 7950"? Clue, 1.2ghz is the ~average from forum posts at oc.net and even [H] has a thread on the OC potential.

"In the end, we were surprised how well the GALAXY GTX 660 Ti GC overclock holds up to these two custom video cards overclock. More often than not the GALAXY GTX 660 Ti GC overclock performance was within 10% of the performance of the overclocked GTX 670! So despite the 192-bit memory bus, the high near 1.3GHz frequency we got on the GALAXY video card plus the memory overclock of 7.6GHz brought performance up to almost the same as the GTX 670 overclocked."

A huge vram OC on the 660ti obviously negates the lower bus, compared to the 670. And its 10% without AA at 1600p, and with 4x AA at 1080p. With 8xAA at 1080p, the gap is more like 20-25% according to heaps of other reviews.

[H], so fail of late. What happened to you guys?

 

[chimaxi83]

As expected, OC for OC, AMD still leads in almost every test they did, even the former Nvidia slaughter of BF3. No surprise.

Yeah it was funny how Hardocp basically says "AMD wins, but you won't be able to get this overclock". Uh huh, sure :thumbsup:

 

[blackened23]

I guess Brent Justice is totally oblivious to Kepler throttle. Odd how he completely ignores this, or maybe his testbed never hits 70C.

 

[exar333]

The overclocks are very strange all around. The 660Ti has a MONSTER memory overclock, and the 670 memory overclock was anemic. Weird...


Sent from my iPad using Tapatalk HD

 

[Ferzerp]

I've noticed that the OP of this thread 90% of the time just does drive-by link posting.

 

[tviceman]

Pretty fail no MSAA at 1600p.

Weren't you trying to convince everyone on here that MSAA in BF3 is pointless? Yeah....

Anyways gtx670 has no memory overclock and since the hd7950 didn't hit 1.3ghz Hardocp must have no idea what they are doing since Silverforce swears every hd7950 can hit 1.3ghz on air.

 

[tviceman]

I guess Brent Justice is totally oblivious to Kepler throttle. Odd how he completely ignores this, or maybe his testbed never hits 70C.

Even if it hits 74 degrees, the throttle is very minimal.

 

[Ferzerp]

Even if it hits 74 degrees, the throttle is very minimal.

This. it's only 8 Mhz or 12Mhz or whatever, and then another for each 5C. It's not like proc throttles that throttle more and more to keep it under a specific temp.

 

[Silverforce11]

Weren't you trying to convince everyone on here that MSAA in BF3 is pointless? Yeah....

Anyways gtx670 has no memory overclock and since the hd7950 didn't hit 1.3ghz Hardocp must have no idea what they are doing since Silverforce swears every hd7950 can hit 1.3ghz on air.

It is pointless in BF3.. but you must have missed the other games being tested?? [H] represents the enthusiast segment, and they run with no AA? Are they being retarded on purpose? I can't see another reason.

Nope, i don't swear "the shit you just claimed". I take my numbers from reading the massive OC threads on tech forums when i decided to buy the 670 or 7950, and picked the best brand for OC potential and thermals/noise. If the average is 1.2ghz, then the <1.2ghz [H] did is below average and to claim it as the "best" is pretty damn ignorant and misleading.. not to mention, XFX are one of the worser cards for OC, their Black Ed had no sinks on the vrms and the fans are noisy.

 

[Keysplayr]

Pretty fail no MSAA at 1600p.

Do you think it is realistic to utilize 1600p resolutions? I mean, I don't know too many people who were willing or had the scratch to lay out about a grand to buy a 30" monitor. I can understand 1440p. Those Catleap are getting pretty popular. 1600p just isn't mainstream an wont be anytime soon.

Anyone else?

 

[toyota]

This. it's only 8 Mhz or 12Mhz or whatever, and then another for each 5C. It's not like proc throttles that throttle more and more to keep it under a specific temp.

its 12mhz and its at every 10 C starting at 70 C.

 

[Silverforce11]

Do you think it is realistic to utilize 1600p resolutions? I mean, I don't know too many people who were willing or had the scratch to lay out about a grand to buy a 30" monitor. I can understand 1440p. Those Catleap are getting pretty popular. 1600p just isn't mainstream an wont be anytime soon.

Anyone else?

I don't have an opinion on catleap vs 1600p, as catleaps are still relatively new, correct?

But it's [H], they have 30 inch 1600p and been benching it for a long time. Honestly, i don't care either way but to not include AA in the test at high res is swallowing NV's load without question, and pretty damn fail for an enthusiast review site which Kyle claims to represent.

 

[blackened23]

its 12mhz and its at every 10 C starting at 70 C.

Its 13MHz and 10mV starting at 70C, adding up if temps go higher or persist. Its not really set in stone either, i've seen the throttle go more than 13mhz, clock oscillation happens quite a lot in some games. I'm convinced the driver has some sort of profiles relating to how certain games throttle and others don't as much - although kepler will always do a minimum of 13mhz at 70C. Maybe someone can shed a light on that (driver profiles for throttling)

I have mixed feelings on this. As a power user I hate it. I like having control, I like overclocking, and I don't like throttling telling me that I can't use my OC fully. For the average user who is clueless maybe its a good thing since they get overclocked performance without having to actually do it. Shrug. Anyway, some people can avoid this with aftermarket (such as the DC2, MSI PE, etc) cards since they cool substantially better than reference, so it doesn't affect everyone.

I'm using MSI lightnings now though so I avoid that entirely, fortunately. (They're on sale for 550$ now BTW!)

 

[toyota]

Its 13MHz and 10mV starting at 70C, adding up if temps go higher or persist. Its not really set in stone either, i've seen the throttle go more than 13mhz quite often, clock oscillation happens quite a lot in some games.

I'm using MSI lightnings now though so I avoid that crap entirely, fortunately.

its normally 12 OR 13.

 

[cmdrdredd]

Do you think it is realistic to utilize 1600p resolutions? I mean, I don't know too many people who were willing or had the scratch to lay out about a grand to buy a 30" monitor. I can understand 1440p. Those Catleap are getting pretty popular. 1600p just isn't mainstream an wont be anytime soon.

Anyone else?

1440p performance won't be that far off in reality. I don't see much wrong with it.

The problem is HardOCP needs to overclock that 670 up a bit. Get the memory up more. 6.34Ghz is pretty low for that card.

 

[Rvenger]

They might as well come out and say they are biased towards Nvidia with that review. I knew Brent over at SG forums back in the day when they were painting the Pentium 4's as the best CPU in the world. Anyone remember the P4 @ 4.4ghz flop?

 

[SolMiester]

Who runs a high end 30"LCD with med range GPU?.....that's the only fail I see.....6600Ti isn't supposed to be winning at 1600p, let alone with any AA....

oops, already stated..anyway.

 

[blackened23]

Who runs a high end 30"LCD with med range GPU?.....that's the only fail I see.....6600Ti isn't supposed to be winning at 1600p, let alone with any AA....

oops, already stated..anyway.

I agree with the point...yet....By the same token why do people pair high end GPUs such as 680 sli with crappy 1600x1200 monitors or 1080p screens?

I will never understand

 

[Piano Man]

AMD 7950 all the way. My question is, which one is the quietest?

 

[VisceralM]

I don't know. I have a 670 and love it, but, it's hard to take away anything from that article other than the 7950 is clearly the better performer. He tries to sugar coat it with platitudes and explanations to the contrary, but I dunno.

 

[SirPauly]

It's nice to see if one is an enthusiast over-clocker and receives a nice core ---- HD 7950 offers a nice strength there with performance and performance/value.

 

[cmdrdredd]

AMD 7950 all the way. My question is, which one is the quietest?

Probably Gigabyte, Asus or MSI. Not in that order.

 

[AnandThenMan]

[H], so fail of late. What happened to you guys?

It's mainly Brent, I don't know what happened but he seems to be going out of his way to make sure his results line up with his expectations. Disappointing, IMO he appears to have lost his objectivity.