GTX 660 Ti vs. HD 7950 - An Interesting Comparison

Serephucus

No, this isn't what you're thinking!

http://www.youtube.com/watch?feature=player_embedded&v=ZM043GqzRuk#!

Something I came across, found it quite interesting. There seems to be a few sites that support this as well. Especially this one from Tech Report. Now, I haven't used any of the recent AMD cards (last one I got a play around with was a 5870) but I think this is quite useful information, and personally, I'd prefer a smoother experience to an outright slightly faster one.

Thoughts?

challengemaster

I now have a headache

Eboggles

Serephucus wrote: »

No, this isn't what you're thinking!

http://www.youtube.com/watch?feature=player_embedded&v=ZM043GqzRuk#!

Something I came across, found it quite interesting. There seems to be a few sites that support this as well. Especially this one from Tech Report. Now, I haven't used any of the recent AMD cards (last one I got a play around with was a 5870) but I think this is quite useful information, and personally, I'd prefer a smoother experience to an outright slightly faster one.

Thoughts?

Certainly looks like the 660ti is smoother, but would that be noticeable?

Of course when it's slowed down the comparisons can be conclusively drawn, but what about a blind test? If there was a noticeable difference, the 660ti beats the HD7950 (stock v stock), however at 1200mhz (for example) the HD7950 would probably beat it?

Interesting stuff anyway, thanks for the link .

EDIT: After having another look, I'm slightly confused; it looks like the GTX 660ti is at a moderate choppiness, and then the HD7950 is not choppy at all, but then takes large stutters?

It's actually splitting hairs at this stage... Go with which one you think is better I guess?

Dcully

@ Eboggles , i linked you this the other day, didnt you read it then ?

Im no expert but all i know is my MSI GTX660 TI TwinFrozr is a little rocket.
Performance is phenomenal for the price of entry.
Im running BF3,Planetside2 etc @ ultra settings 8XAA and get fantastic framerate yet the temperature never goes over 45C,as i say im no expert but thats the coolest card ive ever used, its also the quietest.
Will it be as good when i change from 1920X1200 to 2560X1440?
Time will tell but benchmarks ive seen are very positive.
Both cards offer amazing performance/price bang for your buck.
Im not into this nvidia V ati stuff, my last 2 cards before the 660ti were ATI for the simple reason that they offered what seemed to me and reviewers to offer the best bang per buck,those cards being HD4870 and HD6950 ,both cracking cards and more than done the job for me at the time.

Eboggles

Dcully wrote: »

@ Eboggles , i linked you this the other day, didnt you read it then ?

Im no expert but all i know is my MSI GTX660 TI TwinFrozr is a little rocket.
Performance is phenomenal for the price of entry.
Im running BF3,Planetside2 etc @ ultra settings 8XAA and get fantastic framerate yet the temperature never goes over 45C,as i say im no expert but thats the coolest card ive ever used, its also the quietest.
Will it be as good when i change from 1920X1200 to 2560X1440?
Time will tell but benchmarks ive seen are very positive.
Both cards offer amazing performance/price bang for your buck.
Im not into this nvidia V ati stuff, my last 2 cards before the 660ti were ATI for the simple reason that they offered what seemed to me and reviewers to offer the best bang per buck,those cards being HD4870 and HD6950 ,both cracking cards and more than done the job for me at the time.

I did read the TechReport review if that's what you're talking about, I haven't seen the video before though.

The 660ti should be great for 1920x1200, but at 2560x1440 with AA the performance will start to diminish pretty rapidly. This can be attributed to the 660ti's 192-bit bus, and less so to the 2GB of vRAM.

Tom's aren't the most thorough of reviewers, but these charts demonstrate the point I'm trying to make very well.

Without AA:

What I would call an expected result, all cards line up according to price.


With 4x AA:


With 8x AA:


When 8x AA is applied, the HD7950 (at stock) has a framerate that is nearly double that of the 3GB 660ti, which oddly enough finished below the 2GB version. So yeah, if you want AA, or high-res textures, or higher-than-HD gaming, you should probably pick up a 7870, or a 79xx card.

Serephucus

I still think you'd be fine. 25x14 is only about 50% more pixels than 19x10, which is much less than even 2xAA.

Dcully

I doubt ill need AA at all tbh @ 2560X1440.
What swings it for me with these cards is the Nvidia has:

PhysX [simply mindboggling in Borderlands2 in particular]
TXAA [which is awesome from what ive seen in blops2 atleast with no performance hit at all]
and Adaptive Vsync for when i use Vsync

marco_polo

Serephucus wrote: »

No, this isn't what you're thinking!

http://www.youtube.com/watch?feature=player_embedded&v=ZM043GqzRuk#!

Something I came across, found it quite interesting. There seems to be a few sites that support this as well. Especially this one from Tech Report. Now, I haven't used any of the recent AMD cards (last one I got a play around with was a 5870) but I think this is quite useful information, and personally, I'd prefer a smoother experience to an outright slightly faster one.

Thoughts?

My main thought is it is quite annoying that they have changed every single game in their benchmark suite (they even changed the tested section of the only common game Skyrims) so you can't really compare with the previous release time reviews .

It hard to know what to make of the results TBH, but I do have alot of faith in tech report as a excellent review site, and they seem to have covered off all the variables (driver version, different card, even windows 7/8) It will be interesting to see what AMD discover, as there definitely appears to be something up.

Still it is exceptionally puzzling as to why it appears so far out of whack with previous results (on completely different suite of games albeit), however nearly all of the games in the latest one are very new, released within the last few months, what does does spring to my mind that perhaps the AMD driver team have a good bit of work to do on those new releases yet?

http://techreport.com/review/23419/nvidia-geforce-gtx-660-ti-graphics-card-reviewed/11
http://techreport.com/review/23150/amd-radeon-hd-7970-ghz-edition/11
http://techreport.com/review/22384/amd-radeon-hd-7950-graphics-processor/12

Anyway as a HD 7850 owner I'd kill to have either of those cards. These first word PC gaming problems hold little concern for me

Eboggles

This thread is kinda old, but there's been some new developments in the story thanks to Cats 13.2 beta: http://techreport.com/review/24218/a-driver-update-to-reduce-radeon-frame-times

Basically, it cuts down 12.11s latencies to make it more in line with Nvidia's.

Here's Skyrim with 12.11 betas:


With Nvidia:


With 13.2:


So, it's a nice improvement but it seems to be a tiny bit off the 660ti. Hard to say if you'd notice it though.

Dcully

I found it funny that plenty on this forum was peddaling the 7950 as destroying the 660ti , how wrong they were

Thanks for the update Eboggles,ati just had to do something.

Fluffy88

I think it was in the last Tech Report podcast they mentioned that they had been talking to AMD about these issues and AMD knew what was causing these latencies and were committed to delivering new drivers that would fix the problem.
But it will take time to fix all the problems as there is a few different reasons.

So we should see a few new driver releases coming over the next few weeks that should bring AMD even closer to nVidia in terms of frame latencies.

BloodBath

Thee is no comparison.

The 7950 is a vastly superior card.

It has more ram, a far better memory interface meaning it can handle higher resolutions far better and will also be able to deal with future graphics memory hungry games better. It also overclocks to far greater levels than the 660ti. An overclocked 7950 can outperform a factory overclocked 680gtx. You will never get that performance from an overclocked 660ti.

The videos and benchmarks being linked are also from older drivers. The most recent drivers imporved the 7000 series performance a lot.

BloodBath

Serephucus wrote: »

I still think you'd be fine. 25x14 is only about 50% more pixels than 19x10, which is much less than even 2xAA.

It's actually almost double.

1920x1080 = 2,073,600

2560x1440 = 3,686,400

Dcully wrote: »

I doubt ill need AA at all tbh @ 2560X1440.
What swings it for me with these cards is the Nvidia has:

PhysX [simply mindboggling in Borderlands2 in particular]
TXAA [which is awesome from what ive seen in blops2 atleast with no performance hit at all]
and Adaptive Vsync for when i use Vsync

The 7950 handles that res far better with the better memory interface and 3gb ram.

PhysX is a marketing gimmick. If you want it that badly it can be run through the cpu on borderlands on amd cards perfectly.

Amd has a feature identical to TXAA.

Adaptive Vsync can be done on amd cards as well.

Eboggles

BloodBath wrote: »

The most recent drivers improved the 7000 series performance a lot.

They're using the latest drivers .

Dcully

BloodBath wrote: »

Thee is no comparison.

The 7950 is a vastly superior card.

It has more ram, a far better memory interface meaning it can handle higher resolutions far better and will also be able to deal with future graphics memory hungry games better. It also overclocks to far greater levels than the 660ti. An overclocked 7950 can outperform a factory overclocked 680gtx. You will never get that performance from an overclocked 660ti.

The videos and benchmarks being linked are also from older drivers. The most recent drivers imporved the 7000 series performance a lot.

Vastly superior card?
Do you understand what vastly means?
You couldnt be more wrong mate, they are very very close to eachother especially @ 1080 or 1200.

As for overclocking ,Ive overlocked my MSI twin frozr 660ti well past 670 speeds and almost to 680 speeds with stock cooling, im pretty sure i could get to 680 speeds if i wished.

marco_polo

Incidentially this driver has been leaked on 3d guru if anyone can't wait till until the official release next week and fancies doing some QA for AMD.

http://www.guru3d.com/files_details/amd_catalyst_13_2_beta_download.html

BloodBath

Dcully wrote: »

Vastly superior card?
Do you understand what vastly means?
You couldnt be more wrong mate, they are very very close to eachother especially @ 1080 or 1200.

As for overclocking ,Ive overlocked my MSI twin frozr 660ti well past 670 speeds and almost to 680 speeds with stock cooling, im pretty sure i could get to 680 speeds if i wished.

I beg to differ.

Speed is one thing. Performance is another.

This is comparing it to the 3gb 660ti to make things more fair. Most models are 2gb.

http://www.hardocp.com/article/2012/08/23/galaxy_gtx_660_ti_gc_oc_vs_670_hd_7950/2

These benchmarks were before the latest amd drivers which boost performance even more.

Dcully

http://www.techpowerup.com/reviews/MSI/GTX_660_Ti_Power_Edition/28.html

I could post links all night long mate

Both cards are very very close, neither is vastly superior.

BloodBath

Dcully wrote: »

http://www.techpowerup.com/reviews/MSI/GTX_660_Ti_Power_Edition/28.html

I could post links all night long mate

Both cards are very very close, neither is vastly superior.

What you just linked shows the stock 7950 outperforming the stock 660ti at both 1920x1080 and even more so at 2560x1600.

Taking into account that the 7950 overclocks better and will always handle AA and higher resolutions better because of the better memory interface and more memory + the newer drivers boosting performance by around 10% then there is clearly only 1 winner.

At least my link was a fair comparison of 3gb 660ti model and a 670gtx all overclocked to around the max overclocks most people will get and you can see the real differences there. Add another 10% onto those amd frames and it's an even bigger gap.

The 1920x1080 results from the same article.

http://www.hardocp.com/article/2012/08/23/galaxy_gtx_660_ti_gc_oc_vs_670_hd_7950/3

Let's not forget the 7000 series supports dx11.1 and the nvidias don't.

Dcully

stock 7950 outperforming 660ti? i think you should take another lok, here ill make it easy for you.

3GB 7950 neck and neck with stock 2GB 660ti
and beaten by the 660ti power editions.




and 1 % difference @ 1600




Now how can you say with a straight face the 7950 is a vastly superior card?
fanboy much ? :rolleyes:

Just face facts mate, there is sweet feck all between the cards,even your link shows this.

Ill quote the summary for you

"At that price the card easily beats AMD's HD 7950 in all important criteria: performance, power, noise, heat, performance per Dollar, performance per Watt"

Serephucus

It's actually a 7% difference. MSI's PE 660 had a ~100MHz OC.

BloodBath

I can't believe you're still trying to flog this one.

I'm not saying the 660ti is a bad card. If someone came here asking which to get though then it's pretty obvious the 7950 is the better card by quite a reasonable margin. Maybe not at stock levels but overclocked is a different story. The simple fact is the 7950 is still better even at stock and supports DX11.1 while the whole Nvidia line only supports 11.0.

A fairer comparison in performance would be the 670gtx. This is very close to the 7950 but costs more so the 7950 is still a better choice.

Pro's over the 660ti;

3gb vs 2gb memory
384 bit memory interface vs 192 bit
DX 11.1 support vs DX 11.0
Larger % overclocks over stock frequency's
Beats the 660ti in most benchmarks especially at higher than 1080p
Overclocked performance is closer to factory overclocked 680gtx performance
Better multi monitor support

Cons (according to Nvidia Fans)

Worse 3d vision support (99% won't use this feature)
No gpu Physx support (Marketing ploy that can be run through cpu for most of the few titles that support it )
Slower frame latency ( has been largely fixed in the latest amd drivers )


I'm by no means a fanboy. I'll go for whatever cards offer the best bang for my money and at the moment that is simply amd. If I could get a 680gtx for similar money to a 7970 then I would but you can't.

That all resolution graph means nothing. If you are buying either of these cards then 1080 should be the minimum resolution you are playing at.

As sereph has pointed out as well you are comparing a pre overclocked 660ti to a stock 7950 when the stock 660ti is also in the graph.

Look at my benchmark links. All of the cards tested are at max overclocks most people will get. The 7950 is 10-35% faster in the small sample of benchmarks there. Even in games that were considered to favor Nvidia cards like BF3. The newer AMD drivers improved performance in bf3 by another 10%.

The frame latency issue was the only concern and the last bastion of any argument that the Nvidia cards were better. The latest drivers make this a non issue.

The summary you are quoting was also from how long ago? Well before AMD reduced the prices on the 7950 and 7970 no doubt.

This is the same blind commitment I see from Nvidia fans everywhere. If things were the other way around I'd be tooting Nvidia's horn just the same. I don't have loyalty to either company. Present them with facts on why the amd cards are better and you are labelled a fanboy.

Serephucus

BloodBath wrote: »

The simple fact is the 7950 is still better even at stock and supports DX11.1 while the whole Nvidia line only supports 11.0.

Means nothing. Virtually no devs out there are using 11.1, very few are even using DX11, so it's just a useless spec.

BloodBath wrote: »

3gb vs 2gb memory
384 bit memory interface vs 192 bit
DX 11.1 support vs DX 11.0

Again, specs. Means nothing if the overall performance doesn't stack up. It's like AMD's processors; 8 cores, 4GHz, a billion megs of cache, and they still get beaten by dual cores.

BloodBath wrote: »

Larger % overclocks over stock frequency's
Beats the 660ti in most benchmarks especially at higher than 1080p
Overclocked performance is closer to factory overclocked 680gtx performance
Better multi monitor support

Links to back this up?

BloodBath wrote: »

Worse 3d vision support (99% won't use this feature)
No gpu Physx support (Marketing ploy that can be run through cpu for most of the few titles that support it )
Slower frame latency ( has been largely fixed in the latest amd drivers )

3D I don't use, and I think it's a gimmick so I'll leave that one alone. PhysX is no a gimmick. It's not used a lot, but I think when it is used, it adds a lot to a game. Agreed on frame latency.

I should also point out that that TPU review I linked to is using older NVIDIA drivers. The 310.70 drivers added performance increases on par with AMD's 12.11 driver, they just didn't publicise it. Performance was bumped up again with 310.90.

Dcully

The fact remains and the only reason i challenged you is you said there is no comparison, the 7950 is a vastly superior card.
A few frames here and there between the cards and differing games does not equate to vastly superior.

Ill say again, theres is sweet feck all between these cards.
Certainly not as much as you wildly stated.

BloodBath

Means nothing. Virtually no devs out there are using 11.1, very few are even using DX11, so it's just a useless spec.

Currently no. Does that mean it won't be in the future? I don't plan on replacing my 7970 in the next 2-3 years.

Again, specs. Means nothing if the overall performance doesn't stack up. It's like AMD's processors; 8 cores, 4GHz, a billion megs of cache, and they still get beaten by dual cores.

It does stack up though. Look at any benchmarks as the resolutions go up the gap between the 7950 and the 660ti increases. Future games are also more likely to make use of that extra memory and bandwidth. The 660ti in those bench's is also a 3gb model.

Links to back this up?

I already provided links to back it up.

1920x1080

http://www.hardocp.com/article/2012/..._670_hd_7950/3

2560x1600

http://www.hardocp.com/article/2012/08/23/galaxy_gtx_660_ti_gc_oc_vs_670_hd_7950/2

There's plenty more out there that show the gap increasing even more if you go to higher multi monitor resolutions.

3D I don't use, and I think it's a gimmick so I'll leave that one alone. PhysX is no a gimmick. It's not used a lot, but I think when it is used, it adds a lot to a game. Agreed on frame latency.

PhysX is a low grade tactic used by Nvidia to try and improve sales. They basically pay companies to implement features that could be run through the cpu or on any graphics card. In most cases it can still be run through the cpu with amd cards.

This is similar to the low grade tactics used by Nvidia when they had superior tessellation to the amd cards. It was shown that certain things like square blocks and underwater areas were being tessellated in games to artificially improve frame results for them in benchmarks. Always in "The way its meant to be played titles".

I suspect that the 7950's poor showing in batman Arkham city has something to do with a similar underhand method.

These kind of tactics are damaging to everyone involved and only hold back progress in the games market.

I should also point out that that TPU review I linked to is using older NVIDIA drivers. The 310.70 drivers added performance increases on par with AMD's 12.11 driver, they just didn't publicise it. Performance was bumped up again with 310.90.

It's still behind the amd card.

The fact remains and the only reason i challenged you is you said there is no comparison, the 7950 is a vastly superior card.
A few frames here and there between the cards and differing games does not equate to vastly superior.

Ill say again, theres is sweet feck all between these cards.
Certainly not as much as you wildly stated.

Up to 35% better performance for a card that is priced the same and supposed to be it's direct competitor is sweet **** all?

Fluffy88

I'll just point out that the whole point of those Tech Report articles is that they are using a new way to measure GPU performance as they feel that the traditional FPS measurement doesn't show enough detail and can sometimes be misleading.

FPS averages the frame rate of all the frames produced in one second. As we know in terms of PC's one second is a very long time and averaging performance over a full second can really blur the results.
For Example, if your GPU produces 60 frames in one second, it could produce 1 frame every 16 milliseconds or it might have taken 500 milliseconds to produce one frame and the other 59 frames were produced at 1 frame per 8 milliseconds.
Both give an FPS of 60.

The method being used by The Tech Report in the articles Serephucus linked too in the OP (and apparently is used by game devs) is to measure how long the GPU takes to produce each frame. Therefore giving a more detailed look into the actual performance of the GPU.
What they found using this method to compare the 7950 against the 660ti is that the 660ti is a lot smoother. The 7950 has a lot more latency spikes compared to the 660ti, i.e. one frame taking a long time but the rest coming fast.
This latency doesn't reduce the FPS of the game by much since FPS is taking an average over the full second, therefore removing the outliers (slow frames) but these slow frames are still there and are perceptible by the user when playing the game.

So the conclusion was that the 660ti actually gives a better gaming experience as it is smoother, therefore it's the better card. But they did point out, just like Dcully, that even though there is a difference in the cards it is only slight.

BloodBath

Fluffy88 wrote: »

I'll just point out that the whole point of those Tech Report articles is that they are using a new way to measure GPU performance as they feel that the traditional FPS measurement doesn't show enough detail and can sometimes be misleading.

FPS averages the frame rate of all the frames produced in one second. As we know in terms of PC's one second is a very long time and averaging performance over a full second can really blur the results.
For Example, if your GPU produces 60 frames in one second, it could produce 1 frame every 16 milliseconds or it might have taken 500 milliseconds to produce one frame and the other 59 frames were produced at 1 frame per 8 milliseconds.
Both give an FPS of 60.

The method being used by The Tech Report in the articles Serephucus linked too in the OP (and apparently is used by game devs) is to measure how long the GPU takes to produce each frame. Therefore giving a more detailed look into the actual performance of the GPU.
What they found using this method to compare the 7950 against the 660ti is that the 660ti is a lot smoother. The 7950 has a lot more latency spikes compared to the 660ti, i.e. one frame taking a long time but the rest coming fast.
This latency doesn't reduce the FPS of the game by much since FPS is taking an average over the full second, therefore removing the outliers (slow frames) but these slow frames are still there and are perceptible by the user when playing the game.

So the conclusion was that the 660ti actually gives a better gaming experience as it is smoother, therefore it's the better card. But they did point out, just like Dcully, that even though there is a difference in the cards it is only slight.

The latency issue has already been addressed by amd and future updates will improve it further. That leaves raw fps as the deciding factor between the 2 and the 7950 is simply the better card when it comes down to that.

BloodBath

Most recent benchmarks backing up what I was saying about the gap changes at higher resolutions.

http://www.techspot.com/review/603-best-graphics-cards/

They are very close at stock speeds yes but the 7950 is still ahead and goes further ahead when overclocking.

Dcully

From your link

"In this three-way shootout, the GeForce GTX 660 Ti was 9% faster than the Radeon HD 7870 at 1920x1200 and just 2% slower than the 7950. This makes the Radeon HD 7870 a better value option, while the GeForce GTX 660 Ti is the best buy if you can do without the $280."

2% hardly makes it "a vastly superior card" or "no comparison between them" LOL

That right there is why im even bothering to debate with you, you are so wrong to say the 7950 is the "vastly superior card" and ive yet to see anything to back up your claims.


Ive read countless benchmarks,reviews,comparisons etc.
Some favour the 7950 some the 660ti,we are always going to get differing opinions but my point all along is these cards are very close and certainly not anywhere as big a difference as you suggest.
To 99.9% of users both cards will offer almost identical performance give or take a few frames.
Both great cards - end of story!

beno619

Can we agree that the 7950 is a better value proposition when you include the never settle bundle, and the likelihood that it will last a bit longer in a system without needing an upgrade?
Its not vastly superior, but it's overclock potential vs the 660ti defiantly makes is a better card.

BloodBath

Dcully wrote: »

From your link

"In this three-way shootout, the GeForce GTX 660 Ti was 9% faster than the Radeon HD 7870 at 1920x1200 and just 2% slower than the 7950. This makes the Radeon HD 7870 a better value option, while the GeForce GTX 660 Ti is the best buy if you can do without the $280."

2% hardly makes it "a vastly superior card" or "no comparison between them" LOL

That right there is why im even bothering to debate with you, you are so wrong to say the 7950 is the "vastly superior card" and ive yet to see anything to back up your claims.


Ive read countless benchmarks,reviews,comparisons etc.
Some favour the 7950 some the 660ti,we are always going to get differing opinions but my point all along is these cards are very close and certainly not anywhere as big a difference as you suggest.
To 99.9% of users both cards will offer almost identical performance give or take a few frames.
Both great cards - end of story!

Sigh I've provided plenty of evidence as to why. You just chose to ignore it every time. If someone came here asking which to chose would you honestly recommend the 660ti over the 7950?

I've said all along that the huge gap doesn't appear until you start overclocking and that they are close at stock speeds. I've provided the links backing this up showing an overclocked 7950 beating a 660ti by as much as 35%.

Even if someone wasn't overclocking they should still chose the 7950. Better memory, similar frames, DX 11.1 support. There is no contest.

Serephucus

BloodBath wrote: »

Even if someone wasn't overclocking they should still chose the 7950. Better memory, similar frames, DX 11.1 support. There is no contest.

This is the part I disagree with. Better memory doesn't matter. Even with surround, most games won't use 3GB. Nothing using DX11.1, and it'll probably be skipped all together, like DX10.1 was. The 660 Ti uses slightly less power, and has PhysX, as well as CUDA, for supported applications. That's not even mentioning SHIELD. There are plenty of reasons why I'd choose a 660.

BloodBath

Serephucus wrote: »

This is the part I disagree with. Better memory doesn't matter. Even with surround, most games won't use 3GB. Nothing using DX11.1, and it'll probably be skipped all together, like DX10.1 was. The 660 Ti uses slightly less power, and has PhysX, as well as CUDA, for supported applications. That's not even mentioning SHIELD. There are plenty of reasons why I'd choose a 660.

Good god. Can you agree that the trend in graphics memory is an upwards one? At the moment games aren't pushing over 2gb's but they will soon. BF3 and crysis 2 with high res texture pack at high res is already pushing 2gb's.

Unless you plan on upgrading the card again within a year then having better memory can only be a good thing. The next gen of consoles is around the corner and will be native 1080p meaning most games will be using high res textures and more graphics memory.

DX 11.1 has some great features. If it's held back it will be held back by Nvidia as it adds support for things like native stereoscopic 3D support. It's rumored both microsofts and sonys next consoles will support proper stereoscopic 3d as well and use amd cards so i'm sure they will be running dx 11.1 as well, meaning all pc ports should also make use of it. I agree it's still minor but it will be supported.

CUDA matters not to most people. AMD has openCL anyway.

PhysX is the only argument that can be made if you agree with Nvidias tactics on the matter. It's a damn shame these features aren't being implemented as standard by developers when most of us are sitting on quad core's that aren't being fully utilised.

As I've pointed out before though you can get most of these working through your cpu, granted at lower framerates. Since cpu Physx is single threaded, things like fluid physics calculations really slow things down thanks to single thread calculations. The likes on Borderlands 2 is still fully playable at medium Physx settings though.