• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Best nvidia card for £200?

Not really...the so-called "faster on average", according to anandtech, the 7870 is like 4fps faster in Metro2033, 1fps faster in Skyrim, and in Shogun 2 with the same frame rate and in Civ5 with 10fps ahead, however in Batman the 660 is 4fps ahead, Starcraft II 24fps ahead, BF3 8fps ahead . So it is the case or win some lose some on both cards. However if BF3 is the main game, and not playing older games like Metro2033 and Civ5, the GTX660 is no-brainer really for the newer titles.
http://www.anandtech.com/show/6276/nvidia-geforce-gtx-660-review-gk106-rounds-out-the-kepler-family

What about both the 660/7870 overclocked?. I'm only asking as this forum is usually filled with people who would do so.
 
The other question that people always forget to ask is 'are you going to overclock the new card'.
If your not then the 660 is a good card and slightly better than the 7850 and very similar to the 7870. But if you do intend to overclock then the 7850 is very hard to beat in the under £200 price range.

^bugger lol :)
 
What about both the 660/7870 overclocked?. I'm only asking as this forum is usually filled with people who would do so.
Both cards clock to around 1200MHz (may be 1300MHz with a push or luck on landing a good chip)...and considering the stock clock are simliar on both cards, so there's not much difference in it in terms of overclocking headroom.
 
Last edited:
The other question that people always forget to ask is 'are you going to overclock the new card'.
If your not then the 660 is a good card and slightly better than the 7850 and very similar to the 7870. But if you do intend to overclock then the 7850 is very hard to beat in the under £200 price range.

^bugger lol :)

Thank you, I thought that was the case but needed verification as I can't be bothered checking a plethora of articles :).

Both cards clock to around 1200MHz...and considering the stock clock are simliar on both cards, so there's not much in it on overclocking side.

That's really unfair though. The 660 will boost to a higher speed in 3D and then even more when needed (boost & max boost). I just sent back a GTX 670 recently and have experienced this, so the 7870 when overclocked will be vastly different.
 
Thank you, I thought that was the case but needed verification as I can't be bothered checking a plethora of articles :).



That's really unfair though. The 660 will boost to a higher speed in 3D and then even more when needed (boost & max boost). I just sent back a GTX 670 recently and have experienced this, so the 7870 when overclocked will be vastly different.
Have a look on the reference 660 spec...when I said simliar "stock clock", it was taking into the account of GTX660 980MHz/with boost 1033MHz/max boost 1084MHz vs 7870 at 1000MHz...even if both cards were at 1084MHz, the GTX660 would still be faster in BF3, Batman and Starcraft II etc...I don't see the reasoning in getting 7870 for the same price, unless I am a big fan of Metro2033 or Civ5...hell, even Shogan 2, which is a game that is known for AMD walking all over Nvidia with it being a AMD game, the 660 is actually matching the performance of 7870 there.
 
Last edited:
Do we know actual boost clocks of the 660 though? If you look at reviews they seem to scale pretty poorly with over clocking the core.
 
Have a look on the spec...when I said simliar "stock clock", it is taking into the account of GTX660 980MHz/with boost 1033MHz vs 7870 at 1000MHz...even if tossing the boost aside, when both cards are overclocked, the GTX660 would still be faster in BF3, Batman and Starcraft II etc...I don't see the reasoning in getting 7870 for the same price, unless I am a big fan of Metro2033 or Civ5...

Fair enough, I see your point but I think there's something still not right about it. That is the boost clock, not the max boost/Keplar boost which will push the card even further at a higher Mhz than 1033Mhz. The general consensus I've gathered so far, has been when both cards are overclocked, the 7870 is the better performing card. Yes, the 660 is in it's infancy but if most agree with the general consensus it confuses me when I see post like yours that go against it. If the clocks are not side by side for your estimate then maybe you have it wrong or is the general consensus wrong and the cards are a lot closer when both pushed to their maximum potential?.

I know this sounds like I'm debating your point but I assure you, I'm not. I just want the truth so when asked, I can advise better to help others out and it seems there are mixed views about this due to what you're posting.
 
Fair enough, I see your point but I think there's something still not right about it. That is the boost clock, not the max boost/Keplar boost which will push the card even further at a higher Mhz than 1033Mhz. The general consensus I've gathered so far, has been when both cards are overclocked, the 7870 is the better performing card. Yes, the 660 is in it's infancy but if most agree with the general consensus it confuses me when I see post like yours that go against it. If the clocks are not side by side for your estimate then maybe you have it wrong or is the general consensus wrong and the cards are a lot closer when both pushed to their maximum potential?.

I know this sounds like I'm debating your point but I assure you, I'm not. I just want the truth so when asked, I can advise better to help others out and it seems there are mixed views about this due to what you're posting.
Yea I did more looking up and have already added the max boost clock of 1084MHz. But the point remains that even if the 7870 was at 1084MHz instead of 1000MHz, it is still not enough gain to catch up with the 660. Take BF3 for example...let's pretend that the 660 on anandtech's review that at 1084MHz is 48.2fps vs 7870 at 1000MHz is 40.3...7870 overclocking from 1000 to 1084MHz is 8.4% overclock...and even IF the peroformance gain is as much as the overclock %, 40.3fps+8.4% is still only 43.7fps (vs 660 at 48.2fps)...and this is based on the best case scenerion of overclock %=performance gain %.

Also don't forget the 660 is only on its initial driver release, whereas 7870 already got half a years of toning their drivers. The way I see it is if I'm playing BF3 now, and playing Starcraft II now or will be playing Starcraft II Heart of the Swarm, the 660 just seem to be a better bet, and on the other hand there's isn't a game in the horizon that people waiting for would guarantee to run better on the 7870. Also even for existing (old) games that 7870 do better in, the amount is relatively small that it's not even worth mentioning.

I used to recommend the 7870 a lot when it was only £20 more than the 7850...but now that the 660 is here and at the same price, I just can't bring myself to recommend the 7870 over the 660...
 
Last edited:
The problem with looking at Skyrim performance in Anandtech's Bench is that they only test plain vanilla Skyrim. Like previous Bethesday games, I don't see the point of playing without mods. Far more realistic is computerbase's reviews where they run with texture mods:

http://www.computerbase.de/artikel/grafikkarten/2012/test-nvidia-geforce-gtx-660-ti/24/

And there the Kepler's don't do that well.

Also, GCN tends to get more out of overclocking. Can't think of any group reviews which focus on overclocking, but techPowerUp do overclock cards and run BF3 (which tends to run better on Nvidia cards) in all their reviews. Here are their three most recent reviews of GTX660, 660TI and HD7950:
http://www.techpowerup.com/reviews/MSI/GTX_660_Twin_Frozr_III/30.html
http://www.techpowerup.com/reviews/Palit/GeForce_GTX_660_Ti_Jet_Stream/31.html
http://www.techpowerup.com/reviews/Sapphire/HD_7950_Vapor-X/31.html

GTX660:___59.7 FPS
660TI:_____69.0 FPS
HD7950:___74.8 FPS
Also of interest is what they say in terms of % overclock vs % performance gain.
 
Last edited:
Yea I did more looking up and have already added the max boost clock of 1084MHz. But the point remains that even if the 7870 was at 1084MHz instead of 1000MHz, it is still not enough gain to catch up with the 660.

Brilliant and thank you. I missed your edit in my quote so didn't know this.

I'm still not swayed completely until I see some user results with the card versus 7870 owners. You've made me doubt the general feel around the 7870 when comparing it to the 660 though so it looks like I'm going to have to do a lot of reading before opening my mouth on this subject :).
 
The problem with looking at Skyrim performance in Anandtech's Bench is that they only test plain vanilla Skyrim. Like previous Bethesday games, I don't see the point of playing without mods. Far more realistic is computerbase's reviews where they run with texture mods:

http://www.computerbase.de/artikel/grafikkarten/2012/test-nvidia-geforce-gtx-660-ti/24/

And there the Kepler's don't do that well.

Also, GCN tends to get more out of overclocking. Can't think of any group reviews which focus on overclocking, but techPowerUp do overclock cards and run BF3 (which tends to run better on Nvidia cards) in all their reviews. Here are their three most recent reviews of GTX660, 660TI and HD7950:

GTX660:___59.7 FPS
660TI:_____69.0 FPS
HD7950:___74.8 FPS
Also of interest is what they say in terms of % overclock vs % performance gain.

I agree regarding review sites. It's never quite right. It is why user reviews and benchmarks are so important to me.

I knew the 7950 is the better card than the 660ti though. Just confused with the 7870 and 660 at max overclock.

Always nice to see AMD doing well in BF3 though after a shaky start with their drivers.
 
The problem with looking at Skyrim performance in Anandtech's Bench is that they only test plain vanilla Skyrim. Like previous Bethesday games, I don't see the point of playing without mods. Far more realistic is computerbase's reviews where they run with texture mods:

http://www.computerbase.de/artikel/grafikkarten/2012/test-nvidia-geforce-gtx-660-ti/24/

And there the Kepler's don't do that well.

Also, GCN tends to get more out of overclocking. Can't think of any group reviews which focus on overclocking, but techPowerUp do overclock cards and run BF3 (which tends to run better on Nvidia cards) in all their reviews. Here are their three most recent reviews of GTX660, 660TI and HD7950:
http://www.techpowerup.com/reviews/MSI/GTX_660_Twin_Frozr_III/30.html
http://www.techpowerup.com/reviews/Palit/GeForce_GTX_660_Ti_Jet_Stream/31.html
http://www.techpowerup.com/reviews/Sapphire/HD_7950_Vapor-X/31.html

GTX660:___59.7 FPS
660TI:_____69.0 FPS
HD7950:___74.8 FPS
Also of interest is what they say in terms of % overclock vs % performance gain.
I have no idea why you are comparing the 660 to the 7950, especially considering a custom 7950 cost £50-£60 more than a custom 660. The 660 compete against the 7870, not the 7950...

If people have a budget of £250 instead of £180...then yea...sure...

And as many people already pointed out...Skyrim mods is by far the worst to be use for representing how demanding future games' could potentially be. Mod is something that people added to a finished product, not part of the game development process...

I'm just saying hardware wise, 660 and 7870 is shoulder to shoulder to each other that getting either wouldn't make much different, but considering future game releases, the 660 just seem like a better choice, plus the room for improvement over the current initial driver.

I didn't want to say this (and people are probably gonna burn me like a witch for saying this), but from my own experience of currently using a 5850 on this PC, and 9800GTX+ in another at the moment, despite what the frame rate tell me, smoothness wise my 9800GTX+ at 15fps feel as smooth as as my 5850 at 25fps, and 30fps feel as smooth as 40fps on my 5850. I don't know if this still applies for the current gen of cards.
 
Last edited:
I'm just saying hardware wise, 660 and 7870 is shoulder to shoulder to each other that getting either wouldn't make much different, but considering future game releases, the 660 just seem like a better choice, plus the room for improvement over the current initial driver.

I dont know how you can predict future game release performance one way or the other. Alan wake is more recent than BF3 and performs better on AMD hardware. You can also make a good gues that Metro:last light will perform better on AMD also, and probably crysis 3 as crysis games love memory bandwidth. Some games will perform better on AMD, some better on nvidia.

660 drivers are relatively mature, the keplar architecture is almost 6 months old now. It may be the first driver specifically for the 660, but the drivers are the same for all keplar cards, save for an entry in the .ini.
 
I have no idea why you are comparing the 660 to the 7950, especially considering a custom 7950 cost £50-£60 more than a custom 660. The 660 compete against the 7870, not the 7950...

Okay, didn't look up the 7870 reviews from techPowerUp since I had the reviews sorted by time the last one was a good while ago:
http://www.techpowerup.com/reviews/Sapphire/HD_7870_Flex/31.html

For comparison that gets 56.6FPS in BF3.

And as many people already pointed out...Skyrim mods is by far the worst to be use for representing how demanding future games' could potentially be. Mod is something that people added to a finished product, not part of the game development process...

I have to disagree there: PC gaming is all about adding mods otherwise we might as well game on a console.

I didn't want to say this (and people are probably gonna burn me like a witch for saying this), but from my own experience of currently using a 5850 on this PC, and 9800GTX+ in another at the moment, despite what the frame rate tell me, smoothness wise my 9800GTX+ at 15fps feel as smooth as as my 5850 at 25fps, and 30fps feel as smooth as 40fps on my 5850. I don't know if this still applies for the current gen of cards.

I can understand that. These kind of perceptions don't come out of nowhere. So let me add my own caveat: My BFG 8800GT and other Nvidia stuff 'bumbgated' on me so for me to buy or recommend Nvidia they'd have to have something with very compelling value (since they're generally more expensive I doubt that will happen any time soon).

Having said that, it's a pity more sites don't post the min frames in their benchmarks. That Tom's Hardware article on memory bandwidth of 660TI's (although not very scientific: scientific would be to use the same card and under/overclock its memory) did list the min frames and there wasn't much difference between NV and AMD.
 
@ KompuKare What's this bumbgated think you've mentioned?

To Marine-RX179, human perception is incredibly fickle.

I'm sure you know really, but your conclusions are very unreliable in the way they've been presented.
 
I dont know how you can predict future game release performance one way or the other. Alan wake is more recent than BF3 and performs better on AMD hardware. You can also make a good gues that Metro:last light will perform better on AMD also, and probably crysis 3 as crysis games love memory bandwidth. Some games will perform better on AMD, some better on nvidia.

660 drivers are relatively mature, the keplar architecture is almost 6 months old now. It may be the first driver specifically for the 660, but the drivers are the same for all keplar cards, save for an entry in the .ini.
Well, I guess each to their own. But I think generally speaking, there are far more people and spending far more hours on playing BF3 and Starcraft II than Alan Wake or Crysis...and quite frankly speaking, by the time Crysis 3 hit, people that got 7870 OR 660 would need better graphic card to run at close to the max settings anyway...so the slightly higher memory bandwidth on the 7870 ain't gonna be saving grace when the cards don't have enough GPU grunt...

It's all about realistic expectation...if people want to future prove (if there's such a thing), then they should forget both 660 and 7870, and go for at least either a 7950 or GTX670.
 
Last edited:
@ KompuKare What's this bumbgated think you've mentioned?

Though this was general knowledge. A few years ago (8800GT / 8600M / possible nForce 7150) Nvidia chose the wrong type of solder (think it was when leaded solder was being phased out). That solder was too brittle and would eventually (after so and so many on-off cycles) fail. Affected mainly laptop parts but some desktops too and eventually Nvidia were forced to put $250 million aside to cover 'some' of the damages.

'Some' because plenty of people were not eligible (like me and my brother's 8800GT :( ). In my case the fact that BFG went bankrupt did not help either.
 
@Marine-RX179

While it's true that tons of people play Starcraft II even a 560Ti gets 88.5 FPS at 1600P according to techPowerUp:

http://www.techpowerup.com/reviews/MSI/GTX_660_Twin_Frozr_III/20.html

So I wonder how relevant SC2 performance is(computerbase, xbitlabs for instance don't even bother to bench SC2)? Though that game is more CPU bound anyhow (and TPU are using an [email protected])?

Anyway, yes Nvidia cards do really well in SC2 with the 660TI almost as fast as a stock 7970 at 1600P. Guess that's enough frames for 120Hz 3D although I think a RTS in 3D would give me a headache pretty quickly...
 
@ KompuKare What's this bumbgated think you've mentioned?

To Marine-RX179, human perception is incredibly fickle.

I'm sure you know really, but your conclusions are very unreliable in the way they've been presented.
I'm not the only one that pointed this out you know...there are others that laos mentioned this. But if it was really a case of perception being fickle, then it doesn't explain why it is never the other way round. Unless you are using both Nvidia and ATI/AMD at the same time or at least used both in the pass rather than just reading reviews, I don't think you are in the position to say I'm wrong on my perception. I recall even hardocp mentioned in one of their review that the Nvidia card feel smoother overall despite frame rate is similiar to the AMD's. I take no enjoyment on claiming my 5850 isn't as smooth as my 9800GTX+ when on the same frame rate you know.
 
Last edited:
Though this was general knowledge. A few years ago (8800GT / 8600M / possible nForce 7150) Nvidia chose the wrong type of solder (think it was when leaded solder was being phased out). That solder was too brittle and would eventually (after so and so many on-off cycles) fail. Affected mainly laptop parts but some desktops too and eventually Nvidia were forced to put $250 million aside to cover 'some' of the damages.

'Some' because plenty of people were not eligible (like me and my brother's 8800GT :( ). In my case the fact that BFG went bankrupt did not help either.

Ohh, you're quite right it is general knowledge. I've had two laptops go due to this, I was however not aware that it had been named "bumbgate", I know of it as "The nVidia defect".

Yeah, it was pretty bad, wasn't the failure rate for the laptop parts over 90% or something?
 
Back
Top Bottom