• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

***Official GeForce GTX660/GTX650 Review Thread***

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
WUT? LOL can you believe this guy?
Thinly veiled? Dude if I wanted to I could just say that AMD cheated with their drivers. I'm not a beat around the bush type of guy.
But, I don't know for sure, nor do I have any proof of this. All I have is a sudden jump in performance on AMD cards after 12.8, and reports of smoother gameplay, even with slower fps, on competing Nvidia cards. As you have just mentioned yourself.

"[T]he GeForce 306.23 WHQL drivers enable high-quality TXAA anti-aliasing on GTX 600 Series GPUs, improve game performance by up to 17% on all GeForce 400, 500, and 600 Series GPUs"

OH MY GOSH, we must investigate Nvidia, they have a sudden performance jump with a new driver release. I wonder what they compromised in video quality to get that kind of performance gain?

/sarcasm
 

This again? Comparing reference cards?

There are only 2 reference cards on Newegg among 20. I know you are trying really hard but HD7870 ~ GTX660. Is an after-market 660 better than an after-market 7870? No.

Considering how off AT's power consumption results have been way off from all the other professional reviews, I am not sure we can just use that 1 review.

Tom's Hardware
Power.png


TechSpot
Power.png


TPU
power_peak.gif


HardwareCanucks
GTX-660-78.jpg


HardOCP
13473897920GOm2rnpF3_12_1.gif


At best you are looking at 10-15W differential on a system that uses 280-300W. Are you serious now? It has come down to 10-15W of power difference on average? but losing by 10%+ in games doesn't count? When are you going to sell your GTX460s and get a PS3? You know they use 400W of power together.

It must be worth it to save 10-15W of power and then have terrible performance in modern titles with MSAA, mods or global illumination:
http://www.xbitlabs.com/articles/graphics/display/geforce-gtx-660_9.html#sect3

Sounds well worth it!
 
Last edited:
Xbitlabs, Tom's Hardware, HT4u, Computerbase, TPU are all showing 7870 to be a faster card than the reference 660.

Add Bit-Tech to this list too:

"Slotting between AMD's 7800 cards, GTX 660 2GB fills a gap in that ever-important sub-£200 market. The HD 7850 2GB is 10-15 per cent slower but 10-15 per cent cheaper, while the HD 7870 2GB is, on average, seven per cent faster but ten per cent dearer at around £200. All three cards come out of the mix competitively then, each offering a healthy amount of additional performance the further up the pricing ladder you go."
http://www.bit-tech.net/hardware/2012/09/13/nvidia-geforce-gtx-660-2gb-review/11

The 6-7% faster for 7870 over the 660 is pretty much consistent for many reviews, but after-market 660s close this gap.

Is GTX660 a good card for $229 against a 7870? Sure. Is it a game changer for the marketplace? Not at all unless you only buy NV cards.

There is still a question mark with how GTX660/660Ti will handle future games with MSAA:

Framerate%201920x1080%20Percent.png


Framerate%202560x1440%20Percent.png


"Nvidia's GeForce GTX 660 Ti isn’t really a premium card. It would need to perform better at higher-end settings to satisfy the folks shopping for a less expensive alternative to the GTX 670. If you're really only looking for a middle-of-the-road card to game at mainstream resolutions and modest settings, the less expensive Radeon HD 7870 is ample." ~ Tom's Hardware
 
Last edited:
Personally, I don't understand the use of high res with mid range cards?.....and $/perf isn't the same the world over!
 
For launch drivers for the 660 also, performance is neck and neck with Pitcairn, while Pitcairn is on it's 6th or so revision, and from the panicked outcry's the AMD cards always need to be tested with today's break-through driver that uncovers never before seen performance/sarcasm. Of course we will now hear how Nvidia won't benefit from driver updates. And that has nothing to do with them having better drivers. It's just biased myths touted for convenience.
 
For launch drivers for the 660 also, performance is neck and neck with Pitcairn, while Pitcairn is on it's 6th or so revision, and from the panicked outcry's the AMD cards always need to be tested with today's break-through driver that uncovers never before seen performance/sarcasm. Of course we will now hear how Nvidia won't benefit from driver updates. And that has nothing to do with them having better drivers. It's just biased myths touted for convenience.

Indeed and it seems some of the guys here can see things on the horizon multiple reputed reviewers fails to see.When AMD themselves compared the 660Ti against 7870 in their own slides people here compares it against 7950.I believe after the rumored price cut on 660Ti this nonsense will stop.
 
Yes, I would definitely not use the term "enthusiast" for someone who buys midrange. Your personal opinion is fine and all, but the fact is Nvidia and AMD are closer than they have been in a long long time regarding performance and perf/W. Perf/$ should not be judged so prematurely. The card launched today, the 7800 series had much more time to settle in. Give it another 2 weeks or so.

IMO, you can be an enthusiast and think that people who spend $400+ on a video card are maladjusted. Just saying. 😀
 
It's okay for the price range, where high AA is almost irrelevant in this segment.

But its pretty moronic to claim its superior on perf/w metric when its either using 10W less or up to 20W more depending on the reviews, and its either winning by 4% or losing by 7-25% *with high AA. It's not even grasping at straws..

This is also the only generation so far where i've seen so many ppl on this forum ignore overclocking as a major factor. At least be consistent.
 
What's the problem? Seem like that gamer didn't rushed out to buy overpriced products. BTW the GTX660 costs $229 and is slighly slower than a 7870 for $349 6 months ago. You should more worry about people who bought a 7870 8 weeks ago. They lost 30% of the value and that not include all the free games like Sleeping Dogs - nice to see that AMD just matches "the price/performance" point of the GTX660 after 6 months. :awe:

I don't recall anyone here recommending the 7870 when it was $350.
 
Have to add, perhaps some of you seem to need this, let's start a hypothetical "what-if" scenario rant: I'll start, what if AMD was 7 months late and only offered parity (or worse) perf/$ and perf/w?

Rant away..
 
Of course we will now hear how Nvidia won't benefit from driver updates.
Sorry, but you're the only one on this board that's been saying this. And you only need to look here at the AT review to see that both Nvidia and AMD increased performance with their newer drivers. I don't get why you're complaining about wanting to see the cards tested with drivers that have enabled performance enhancements.
 
Personally, I don't understand the use of high res with mid range cards?.....and $/perf isn't the same the world over!

True, true. Where we live the cheapest 660 in stock on pricespy is $439NZD while 7870's start at $382NZD (incl. GST). I don't know why anyone with any common sense would consider a 660 at those prices.
 
Xbitlabs, Tom's Hardware, HT4u, Computerbase, TPU are all showing 7870 to be a faster card than the reference 660.

Add Bit-Tech to this list too:

"Slotting between AMD's 7800 cards, GTX 660 2GB fills a gap in that ever-important sub-£200 market. The HD 7850 2GB is 10-15 per cent slower but 10-15 per cent cheaper, while the HD 7870 2GB is, on average, seven per cent faster but ten per cent dearer at around £200. All three cards come out of the mix competitively then, each offering a healthy amount of additional performance the further up the pricing ladder you go."
http://www.bit-tech.net/hardware/2012/09/13/nvidia-geforce-gtx-660-2gb-review/11

The 6-7% faster for 7870 over the 660 is pretty much consistent for many reviews, but after-market 660s close this gap.

Is GTX660 a good card for $229 against a 7870? Sure. Is it a game changer for the marketplace? Not at all unless you only buy NV cards.

There is still a question market with how GTX660/660Ti will handle future games with MSAA:

Framerate%201920x1080%20Percent.png


Framerate%202560x1440%20Percent.png


"Nvidia's GeForce GTX 660 Ti isn’t really a premium card. It would need to perform better at higher-end settings to satisfy the folks shopping for a less expensive alternative to the GTX 670. If you're really only looking for a middle-of-the-road card to game at mainstream resolutions and modest settings, the less expensive Radeon HD 7870 is ample." ~ Tom's Hardware

You base this question on one game?
A midrange card is not for higher-end settings but around 1080p with 4xMSAA. At this setting, the 660 is almost on par with the 7870 and the 660 Ti almost on par with the 7950. Most people only look at ratings but don't consider actual fps. At higher settings one card might speed away a bit, but that is irrelevant as fps on both cards would not be exactly playable.

It is always wise to have as many data points as possible, so I would like to direct your attention to the excellent launch analyses of 3DCenter. They average up to 10 (!) reviews, not only for performance numbers but also for power consumption values (card only, not full system). For a averaged performance index, go to page 2 in both articles and scroll down.

GTX660
http://www.3dcenter.org/artikel/launch-analyse-nvidia-geforce-gtx-660

GTX660 Ti
http://www.3dcenter.org/artikel/launch-analyse-nvidia-geforce-gtx-660-ti
 
Last edited:
You base this question on one game?
A midrange card is not for higher-end settings but around 1080p with 4xMSAA. At this setting, the 660 is almost on par with the 7870 and the 660 Ti almost on par with the 7950.

It is always wise to have as many data points as possible, so I would like to direct your attention to the excellent launch analyses of 3DCenter. They average up to 10 (!) reviews, not only for performance numbers but also for power consumption values (card only, not full system). For a averaged performance index, go to page 2 in both articles and scroll down.

GTX660
http://www.3dcenter.org/artikel/launch-analyse-nvidia-geforce-gtx-660

GTX660 Ti
http://www.3dcenter.org/artikel/launch-analyse-nvidia-geforce-gtx-660-ti

All the midrange cards can do 1920 * 1080 with 4* MSAA, by your logic we shouldn't bother finding out which card can go faster / offer the best value or is more future proof because it's only a midrange card and knowone wants 8* MSAA or higher resolutions.. Get Real.
 
Personally, I don't understand the use of high res with mid range cards?.....and $/perf isn't the same the world over!
I agree and disagree 🙂

I think the reviews should focus on playable settings, sort of like HardOCP to give the buyer a good picture of what the GPU is able to do. And that means to not use more AA than the card can handle. And with that metric the GTX 660 does OK, or about the same as a 7870 or 570.

But I also think the reviewers should try to find out the limits for the cards so that a buyer can get a more informed picture about the cards limits, and in that aspect it's important to use settings the card cannot handle, like AA and tessellation.

And lastly I find it meaningless when they compare the total % differences of the performance when they include older games that has higher minimums than 60fsp (or maybe 120fps if they are using a 3d screen). They should just skip those games in the total as the cards have bigger engines than needed.

Edit: And in Norway the cheapest 660 is about 7-8% cheaper than 7870.
 
Last edited:
All the midrange cards can do 1920 * 1080 with 4* MSAA, by your logic we shouldn't bother finding out which card can go faster / offer the best value or is more future proof because it's only a midrange card and knowone wants 8* MSAA or higher resolutions.. Get Real.

Do you think, future games will run faster on these cards than current games do? I don't.

Look at the following results of the 660 for example.
1080p, 4xMSAA where possible, only actual gameplay (no built-in benchmarks):

Batman AC: 40 fps (Hardwarecanucks)
BF3: 48 fps (Hardwarecanucks)
Crysis 2: 40 fps (Computerbase)
Skyrim+ENB: 38 fps (Computerbase)
Metro 2033: 24 fps (Computerbase)
Max Payne 3: 29 fps (Computerbase)
Shogun 2: 38 fps (Hardwarecanucks)
Dragon Age 2: 38 fps (Computerbase)
The Witcher 2: 43 fps (Hardwarecanucks)

Need I go on?
The 7870 basically ties the 660 at these settings on average, maybe it is a tad faster. These cards can do 1080p with 4xMSAA in current games...barely (at least by my standards). These are average values and don't even include significant dips below that disrupt gameplay. And there are honestly people who want to compare midrange cards with 8xMSAA or 2560x1440 or bring in future games that will be even more demanding? Get real.

Based on these fps values, these cards are best suited for 1080p with no or 2xMSAA or for 1680x1050 with 4xMSAA. Unless you like low fps.
 
Last edited:
I only skimmed a few reviews and such but it seems that the GTX 660 is competitive with the HD 78xx brothers. Exact pricing, game bundles, etc. may break the tie here, as well as the usual power/noise/heat and Physx/CUDA/a-Vsync/GPUBoost/OC headroom factors. I'd be curious to see more oc tests done so we get a better idea of the typical OC headroom.
 
Do you think, future games will run faster on these cards than current games do? I don't.

Look at the following results of the 660 for example.
1080p, 4xMSAA where possible, only actual gameplay (no built-in benchmarks):

Batman AC: 40 fps (Hardwarecanucks)
BF3: 48 fps (Hardwarecanucks)
Crysis 2: 40 fps (Computerbase)
Skyrim+ENB: 38 fps (Computerbase)
Metro 2033: 24 fps (Computerbase)
Max Payne 3: 29 fps (Computerbase)
Shogun 2: 38 fps (Hardwarecanucks)
Dragon Age 2: 38 fps (Computerbase)
The Witcher 2: 43 fps (Hardwarecanucks)

Need I go on?
The 7870 basically ties the 660 at these settings on average, maybe it is a tad faster. These cards can do 1080p with 4xMSAA in current games...barely (at least by my standards). These are average values and don't even include significant dips below that disrupt gameplay. And there are honestly people who want to compare midrange cards with 8xMSAA or 2560x1440 or bring in future games that will be even more demanding? Get real.

Based on these fps values, these cards are best suited for 1080p with no or 2xMSAA or for 1680x1050 with 4xMSAA. Unless you like low fps.

No but I think future games will run better on a 7850 / 7870 than the 660/ ti because the nvidia cards lack memory bandwidth. Future games are going to tax these cpu with high res textures even at 1080p ect.
The point I was trying to make is the 7850/70 may last a whole upgrade cycle longer than the 660/ti.

I swear no alliance to either card company and am trying to decide between a 7950 vs 660 Ti right now I'm favoring the 7950 for reasons stated above.
 
No but I think future games will run better on a 7850 / 7870 than the 660/ ti because the nvidia cards lack memory bandwidth. Future games are going to tax these cpu with high res textures even at 1080p ect.
The point I was trying to make is the 7850/70 may last a whole upgrade cycle longer than the 660/ti.

I swear no alliance to either card company and am trying to decide between a 7950 vs 660 Ti right now I'm favoring the 7950 for reasons stated above.

Run better, maybe. Will that matter? No - as both GPUs will be too slow by then for that setting and you would have to turn down MSAA/texture resolution anyway. You also cannot compare the memory bandwidth demands of MSAA with the demands of a (future) game in general. Just because a card is faster with MSAA today doesn't mean it will speed away in say BF4 or Witcher 3.
So I seriously doubt that the 7800 series will last longer than the 660 series.
 
You can focus all you want (snip)

I know. Thank you.

RussianSensation said:
I don't think Keys can. You can't ask a guy who works for Coca-Cola if Pepsi tastes better and expect an honest public answer? 😀

Congrats on that new low you've reached in such a short period of time. Just reach out there and call me dishonest when all I want to do is explore the possibilities.
I do not know the reason you wish this topic to go away, but my lord it's blatantly obvious.

LOL TR ONLY test 5 five 5 games. What's next? They only tested 7? 8?
5 games is significant don't you think? If it was one, then I'd agree with your assessment.

And for all those claiming conspiracy theories, you could be right. But you could also be wrong. I know you don't want to be wrong, not one inch, but it's perfectly legitimate to discuss the possibilities when observing the data we have.

Call me whatever you wish, from this point on at your own peril. If you keep the discussion civil and respectable, you have no concerns.

Laugh, do whatever. Enjoy yourself.
 
Last edited:
"[T]he GeForce 306.23 WHQL drivers enable high-quality TXAA anti-aliasing on GTX 600 Series GPUs, improve game performance by up to 17% on all GeForce 400, 500, and 600 Series GPUs"

OH MY GOSH, we must investigate Nvidia, they have a sudden performance jump with a new driver release. I wonder what they compromised in video quality to get that kind of performance gain?

/sarcasm

AHahahahaaaa Ohmylord sofunny. Sidehurtszzz111...

Do reviews support this? Is the gameplay not as smooth as before and more like AMD gameplay now?

So petty.
 
Back
Top