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

[The Alias]

Lower power draw, 3 monitors without adapter, adaptive Vsync, GPU Boost for lazy overclockers, CUDA/PhysX, better gamedev relations (and also TWIMTBP), arguably better multi-GPU drivers... NV has more than just Physx, let's be fair here. They are all minor things but they add up and I can see why most people choose NV when price/perf is comparable.

it's not when they're oc'ed didn't you see the beatdown that ensued when both cards were oc'ed ?

 

[blastingcap]

How about that Unigine based game? I forget what it's called. But I'm sure that is DX11 and chock full of tesselation.

Then there is always the Heaven bench if you want to just check for pure tesselation performance.

Oh you mean this one?

http://www.legitreviews.com/article/1925/12/

Even at Extreme Tessellation, Unigine is not a pure tessellation perf benchmark btw; if it were the 7870@stock > 7970@stock. They also have things like high textures, AA/AF, etc.

it's not when they're oc'ed didn't you see the beatdown that ensued when both cards were oc'ed ?

Which cards are we talking about here? And yes I advocate for oc vs oc results all the time. I am not "most people," I go for price/perf with a few other considerations like power draw, noise, heat, etc. and don't personally care about CUDA/PhysX/etc. much. But I know that others do. Plus NV GeForce is a stronger brand. AMD is a tainted brand (due to getting beaten up by Intel so often), and AMD was stupid to throw away the ATI brand which is not as tainted.

 

[blastingcap]

sorry for double post

 

[Silverforce11]

It was a response to your friend AtenRA WHO BROUGHT UP THE USELESS BENCH IRRELEVANT FOR ACTUAL GAMING and to your USELESS REMARK IRRELEVANT FOR ACTUAL GAMING:

Not sure if you noticed, recent dx11 games have actually used direct compute Forward+.. if you think AAA titles are irrelevant and tessmark is, no point discussing it further.

 

[Riek]

Lower power draw, 3 monitors without adapter, adaptive Vsync, GPU Boost for lazy overclockers, CUDA/PhysX, better gamedev relations (and also TWIMTBP), arguably better multi-GPU drivers... NV has more than just Physx, let's be fair here. They are all minor things but they add up and I can see why most people choose NV when price/perf is comparable.

- i wont contest CUDA/Physx that much since they have some areas where they have an advantage. (although for the consumer space CUDA is more wortheless than openCL at this point).

- Boost is hardly an advantage since its already incorporated in the peformance. Wether its boosts is irrelevant (except if you watch benchmarks of a higher boosted gpu).

- power draw is very card specific. the 660 consumes more power than the 7870 while being slower. The 670-680 have the power advantage at stock vs their opponents. So not sure if power draw can be seen as a general thing here...?

- why is having no adapter an advantage?

- game developers, TWIMTBP... would have a stronger deffence if there weren't 3 new games in which AMD developers did a much better job

 

[Silverforce11]

- contest CUDA/Physx that much since they have some areas where they have an advantage. (although for the consumer space CUDA is more wortheless than openCL at this point).

- Boost is hardly an advantage since its already incorporated in the peformance. Wether its boosts is irrelevant (except if you watch benchmarks of a higher boosted gpu).

- power draw is very card specific. the 660 consumes more power than the 7870 while being slower. The 670-680 have the power advantage at stock vs their opponents. So not sure if power draw can be seen as a general thing here...?

- why is the adapter an issue? you get it in the package?

Not to mention Gaming E has been doing the rounds of late producing many awesome games. Minus one advantage blasting?

 

[AtenRa]

It was a response to your friend AtenRA WHO BROUGHT UP THE USELESS BENCH IRRELEVANT FOR ACTUAL GAMING and to your USELESS REMARK IRRELEVANT FOR ACTUAL GAMING:

You do remember this,

http://www.anandtech.com/show/4008/nvidias-geforce-gtx-580/16

Now that AMDs cards have more tessellation performance it is a useless bench once again. ??

It is funny that two of last gen NVIDIA benchmarks featuring Tessellation are performing faster in AMDs GCN cards than Kepler.

http://www.techpowerup.com/reviews/ASUS/GeForce_GTX_660_Direct_Cu_II/12.html

 

[Keysplayr]

Oh you mean this one?

http://www.legitreviews.com/article/1925/12/

Even at Extreme Tessellation, Unigine is not a pure tessellation perf benchmark btw; if it were the 7870@stock > 7970@stock. They also have things like high textures, AA/AF, etc.

Right, cause Pitcairn tesselates better than Tahiti. You're right. But a good point of reference can be made by the performance hit any given card takes.

 

[SirPauly]

You do remember this,

http://www.anandtech.com/show/4008/nvidias-geforce-gtx-580/16

Now that AMDs cards have more tessellation performance it is a useless bench once again. ??

It is funny that two of last gen NVIDIA benchmarks featuring Tessellation are performing faster in AMDs GCN cards than Kepler.

http://www.techpowerup.com/reviews/ASUS/GeForce_GTX_660_Direct_Cu_II/12.html

Indeed! Isn't that great to see? With both offering impressive tessellation may translate into even more tessellation in titles!

 

[Termie]

I will. As they turn up.

New GTX650 review added:

http://www.guru3d.com/article/geforce-gtx-650-msi-power-edition-review/1

 

[The Alias]

Which cards are we talking about here? And yes I advocate for oc vs oc results all the time. I am not "most people," I go for price/perf with a few other considerations like power draw, noise, heat, etc. and don't personally care about CUDA/PhysX/etc. much. But I know that others do. Plus NV GeForce is a stronger brand. AMD is a tainted brand (due to getting beaten up by Intel so often), and AMD was stupid to throw away the ATI brand which is not as tainted.

7870 vs 660 http://www.xbitlabs.com/picture/?src=/images/graphics/geforce-gtx-660/zfulltable.png
7950 vs 660ti/670 (and by extension 7970 vs 670) http://www.hardocp.com/article/2012/08/23/galaxy_gtx_660_ti_gc_oc_vs_670_hd_7950/2

 

[blastingcap]

- i wont contest CUDA/Physx that much since they have some areas where they have an advantage. (although for the consumer space CUDA is more wortheless than openCL at this point).

- Boost is hardly an advantage since its already incorporated in the peformance. Wether its boosts is irrelevant (except if you watch benchmarks of a higher boosted gpu).

- power draw is very card specific. the 660 consumes more power than the 7870 while being slower. The 670-680 have the power advantage at stock vs their opponents. So not sure if power draw can be seen as a general thing here...?

- why is having no adapter an advantage?

- game developers, TWIMTBP... would have a stronger deffence if there weren't 3 new games in which AMD developers did a much better job

To repeat, I am not "most people," and I am not saying these are killer features or anything, more like NV tiebreakers. But I can understand where a lot of people are coming from, though for the truly inept builders who don't even know how to overclock, I wonder if they would be better off just gaming on a console instead.

As an experienced tri-monitor user, I can tell you that not having to use an adapter to get that third monitor up and running for Eyefinity/Surround is a blessing, not a curse.

GPU Boost is for people who don't want to overclock. Better than AMD's less elegant alternative.

Ditto CUDA/Physx--even if it's of limited value, it is better than nothing.

As for gamedevs, it is not just TWIMTBP vs Gaming Evolved. NV has historically had far, far better gamedev relations than AMD has had. It's not just for AAA-title developers, NV tries to help little guys too. Ask some real-life gamedevs what their take on this is.

Not to mention Gaming E has been doing the rounds of late producing many awesome games. Minus one advantage blasting?

See above. Talk to some gamedevs and ask them what they think.

 

[AtenRa]

You dont need an adapter for triple Eyefinity if one monitor has a DP port.

 

[AtenRa]

Indeed! Isn't that great to see? With both offering impressive tessellation may translate into even more tessellation in titles!

I sure hope so

 

[blastingcap]

You dont need an adapter for triple Eyefinity if one monitor has a DP port.

Most cheaper monitors don't have DP ports. Nor do the cheaper Korean monitors, for that matter. Over time this will change, probably, but it's a hassle right now, and I would say it's a Kepler advantage over Tahiti. AMD knows this, which is why on some 79xx models, they bundle miniDP to DVI adapters for free.

 

[AtenRa]

Yes if you dont have a DP port in your monitor.

 

[Elfear]

GPU Boost is for people who don't want to overclock. Better than AMD's less elegant alternative.

You make some good points blastingcap but I see the other poster's point with GPU Boost. If Boost is already taken into consideration in benchmarks, is it really an advantage for a non-overclocker? They're just getting the performance they've seen in reviews.

And for the overclocker isn't the ability to increase voltage an advantage for AMD cards (besides the few Kepler cards with voltage control) vs GPU Boost that locks you out?

 

[blastingcap]

You make some good points blastingcap but I see the other poster's point with GPU Boost. If Boost is already taken into consideration in benchmarks, is it really an advantage for a non-overclocker? They're just getting the performance they've seen in reviews.

And for the overclocker isn't the ability to increase voltage an advantage for AMD cards (besides the few Kepler cards with voltage control) vs GPU Boost that locks you out?

Let me restate my original point: I think that if two cards (one by AMD, one by NV) are in the same price/perf ballpark, I can see why most people would go for the NV part due to the advantages I laid out, and also due to NV's brand equity. Even if you do not personally use ANY of those minor advantages, the fact is that all ties go to NV. NV does not need to win outright, it just needs to be in the same price/perf ballpark and its other advantages will help sell their cards. That's not even including gamedev relations.

I only brought this up as a response to Silverforce's claim that NV's "only" advantage was Physx, and that is not true.

Enthusiasts who know how to hot-rod their GPUs and don't care for things like A-Vsync and GPU Boost and PhysX may come to a different conclusion, of course.

 

[Silverforce11]

Lets be real here, even if NV offer terrible perf/$, perf/w and had terrible OC potential.. they would still most likely outsell AMD.

Thus for informed users, especially those who visit enthusiast GPU forums.. NV's "only" advantage is Physx, everything else you mention is fluff.

 

[toyota]

Just a food for thought,



GTX560Ti MSRP price of the time of release 1/25/2011 = $249,00
GTX660 MSRP price of the time of release 9/14/2012 = $229,00

GTX660 vs GTX560Ti = ~27% faster at almost the same MSRP price after 20 months


HD6870 MSRP price of the time of release 10/21/2010 = $239,00
HD7870 MSRP price of the time of release 3/5/2012 = $350,00
HD7870 MSRP price as of today 9/14/2012 = $249,00

HD7870 vs HD6870 = ~47% faster at almost the same MSRP price after 33 23 months

I call that progression D:

Edit: 23 not 33

so you are comparing the 6870 launch price to the now reduced 7870 price? a 6870 can be had for $149 so it offers way more performance per dollar than the 7870. and at launch, the 7870 had zero performance per dollar advantage over the 6870 launch price plus you could buy the 6870 for way less than its launch price. so at no point has the 7870 been as good as the 6870 as for as performance per dollar goes.

 

[Elfear]

Let me restate my original point: I think that if two cards (one by AMD, one by NV) are in the same price/perf ballpark, I can see why most people would go for the NV part due to the advantages I laid out, and also due to NV's brand equity. Even if you do not personally use ANY of those minor advantages, the fact is that all ties go to NV. NV does not need to win outright, it just needs to be in the same price/perf ballpark and its other advantages will help sell their cards. That's not even including gamedev relations.

I only brought this up as a response to Silverforce's claim that NV's "only" advantage was Physx, and that is not true.

Enthusiasts who know how to hot-rod their GPUs and don't care for things like A-Vsync and GPU Boost and PhysX may come to a different conclusion, of course.

I agree with you. Nvidia has advantages over AMD that matter to some consumers and might be the tie-breaker for two comparative cards. I just don't think Boost is really an advantage. Either you're already taking the extra performance into account when comparing benchmarks of two cards or you want to extract the last bit of performance from you're card and don't want Boost.

Either way not really worth arguing over in a review thread. ()

 

[HurleyBird]

Everything I've seen seems to point to Nvidia having better tessellation performance when the mesh is subdivided to extreme levels, while AMD fares better than Nvidia in more sane scenarios. I would say neither company has better tessellation -- they have balanced their tessellators differently and which company comes ahead depends on how it is implemented in a game.

Of course the fanboys have a field day because each side has data that "proves" the other side wrong despite the truth that is right in front of them. You people need to use some common sense for once, and stop with this damned drama. Republicans and Democrats aren't this bad. Have some self respect for once in your lives, all of you.

 

[Silverforce11]

Everything I've seen seems to point to Nvidia having better tessellation performance when the mesh is subdivided to extreme levels, while AMD fares better than Nvidia in more sane scenarios. I would say neither company has better tessellation -- they have balanced their tessellators differently and which company comes ahead depends on how it is implemented in a game.

Of course the fanboys have a field day because each side has data that "proves" the other side wrong despite the truth that is right in front of them. You people need to use some common sense for once, and stop with this damned drama. Republicans and Democrats aren't this bad. Have some self respect for once in your lives, all of you.

True.

But lets not forget on AMD gpus, if you are forced to deal with tessellation being missused, you can force it to sane levels in CCC.

 

[blastingcap]

Everything I've seen seems to point to Nvidia having better tessellation performance when the mesh is subdivided to extreme levels, while AMD fares better than Nvidia in more sane scenarios. I would say neither company has better tessellation -- they have balanced their tessellators differently and which company comes ahead depends on how it is implemented in a game.

Of course the fanboys have a field day because each side has data that "proves" the other side wrong despite the truth that is right in front of them. You people need to use some common sense for once, and stop with this damned drama. Republicans and Democrats aren't this bad. Have some self respect for once in your lives, all of you.

I do not know about who has better tess right now because frankly I turn off tess in games I play, or the games do not support tess in the first place. I do however think John Carmack's words hint that "sane" tessellation is of some, but limited, value and that "extreme" tessellation will be of more value. Right now that kind of tess power is prohibitively expensive, but we'll see what happens in the next 10 years.

"John Carmack: No, we don’t have any tessellation. Tessellation is one of those things that bolting it on after the fact is not going to do anything for anybody, really. It’s a feature that you go up and look at, specifically to look at the feature you saw on the bullet point rather than something that impacts the game experience. But if you take it into account from your very early design, and this means how you create the models, how you process the data, how you decimate to your final distribution form, and where you filter things, all of these very early decisions (which we definitely did not on this generation) I think tessellation has some value now. I think it’s interesting that there is a no-man’s land, and we are right now in polygon density levels at a no-man’s land for tessellation because tessellation is at it’s best when doing a RenderMan like thing going down to micro-polygon levels. Current generation graphics hardware really kind of falls apart at the tiny levels because everything is built around dealing with quads of texels so you can get derivatives for you texture mapping on there. You always deal with four pixels, and it gets worse when you turn on multi-sample anti-aliasing (AA) where in many cases if you do tessellate down to micro-polygon sizes, the fragment processor may be operating at less than 10% of its peak efficiency. When people do tessellation right now, what it gets you is smoother things that approach curves. You can go ahead and have the curve of a skull, or the curve of a sphere. Tessellation will do a great job of that right now. It does not do a good job at the level of detail that we currently capture with normal maps. You know, the tiny little bumps in pores and dimples in pebbles. Tessellation is not very good at doing that right now because that is a pixel level, fragment level, amount of detail, and while you can crank them up (although current tessellation is kind of a pain to use because of the fixed buffer sizes on the input and output [hardware]) it is a significant amount of effort to set an engine up to do that down to an arbitrary level of detail. Current hardware is not really quite fast enough to do that down to the micro-polygon level.

It’s almost like procedural data where we’ve heard tessellation is going to be the “big thing” since the NP patches from the ATI stuff, and there are reasons why it never caught on. Because... in the early days of shells on things, when you say “well we’ve got a Bézier spline, or a NURB, or something like that,” what we would find is that, well, if we are going to have this net of 16 vertexes around here, you can do cooler game art by making that 16 vertexes for triangles. You’ll have cooler protrudes rather than your smooth Gumby shape. Now that we have the ability to go ahead and do texture sampling, and do real bump mapping, it becomes interesting from a content point, but we don’t quite have the power to do the entire world. You can run a character down like that with current generation stuff, and that’s probably useful directions but you can’t yet go ahead and render your 2 million pixel world at sub-pixel micro-polygon triangles (and certainly not at 60fps). That’s the type of performance that we are going to get no matter what, and I think the smart money bet for next generation consoles is that early on they are just going to be hyped up versions of our current technologies, but when people build technologies from scratch for that, the smart money would be on a tessellation based, you know, all the way down to micro-polygon levels with outside bets on voxel, ray tracing, and hybrid engines."

http://www.pcper.com/reviews/Editor...-Graphics-Ray-Tracing-Voxels-and-more/Transcr

Emphasis in bold added by me.

 

[Keysplayr]

True.

But lets not forget on AMD gpus, if you are forced to deal with tessellation being missused, you can force it to sane levels in CCC.

Or used properly at higher levels than AMD gpu's can deal with, you can ALSO force it to lower levels in CCC.

See, it's all in how you word it.