• 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.

Gtx 470 vs hd5850

HW91 said:
yeah sure, that's why I'm running a stock cooled 470 with a 33% overclock with max fan speed at 75% and max temp at 78C no matter what game I play. Need proof? I'll get it if you ask

Smoother game play is a well known fact of the 470 vs 5870 battle, ask any one who compared them. And any respectable review site put the 470 with minimums around 5870 level, rarely going below and usually above
Click to expand...

So HW91, you actually own a GTX 470?

Would you recommend it? I game at 1920x1200 resolution.

Heat is fine, I dont care how hot it gets, as long as it doesnt melt, but noise, is it particulalry loud?

Do you think it has more distnace in the tank than the 5870, for future gaming titles?

Is this review belivable: http://www.xbitlabs.com/articles/video/display/geforce-gtx470-sli_6.html#sect1

Man I dont know any more, all the reviews say something different!
 
^ yes I own one, yes I'd recommend it

sound is bearable at idle with no music,etc. and you don't really notice it when playing music, games, etc.

assuming those titles use tessellation then yes

can't really say if that review is believable because my card has been overclocked from pretty much day one but I usually get high 50's and above in BC2 1920x1200 as a reference
 
So the GTX 470 is good enough for 1920x1200 gaming.

Yea I think with the 400 series from Nvidia, its a case of the tech is ahead of where the games are, but when the games are using more DX11 features, then they will come in to their own.

Of course by then, ATI will have their own new series of gpu's out.

Is it possible for you to list what kind of frames you get with your GTX 470 for some games, such as, AVP, Metro, Crysis and warhead, at 1920x1200.

I said somewher else, Im not bothered about saving £40 going for the 470 over the 5870, I want to get the btter ofthe two.
 
TheRealDeal said:
It will also burn your house down and the fan at 100 percent trying to cool it will create a twister in your room. On a serious note though gtx470/480 need watercooled to get the best of both worlds where as ati 5 series can happily be overclocked with none of the drawbacks that fermi has. This smoother gameplay you talk of is also rubbish and i am not sure where you get this from. the last review i read used fraps to show that in games not game benchmarks as most reviews use a 5850 had better minimum frames when comparing to a gtx470 which imo would relate to smooth gameplay.

that review is around a month old or more i think so things may have moved on.
Click to expand...

People over exaggerate with the heat and noise which is really pathetic.
 
opethdisciple said:
So the GTX 470 is good enough for 1920x1200 gaming.

Yea I think with the 400 series from Nvidia, its a case of the tech is ahead of where the games are, but when the games are using more DX11 features, then they will come in to their own.

Of course by then, ATI will have their own new series of gpu's out.

Is it possible for you to list what kind of frames you get with your GTX 470 for some games, such as, AVP, Metro, Crysis and warhead, at 1920x1200.

I said somewher else, Im not bothered about saving £40 going for the 470 over the 5870, I want to get the btter ofthe two.
Click to expand...

overclocked or stock?

at 820MHz Metro never goes below 20fps with all settings on max except AA which is set to AAA, Crysis never goes below 27fps using RE4PT0X's MaximumVisuals config mod and 16xQ AA, said about BC2 before, COD4 gets over 100fps always, COD:WAW usually gets 90fps, JC2 Concrete Jungle bench with 16xQ AA got 32fps average although take that with a grain of salt because I was running a cpu based video conversion at the time
 
HW91 said:
overclocked or stock?

at 820MHz Metro never goes below 20fps with all settings on max except AA which is set to AAA, Crysis never goes below 27fps using RE4PT0X's MaximumVisuals config mod and 16xQ AA, said about BC2 before, COD4 gets over 100fps always, COD:WAW usually gets 90fps, JC2 Concrete Jungle bench with 16xQ AA got 32fps average although take that with a grain of salt because I was running a cpu based video conversion at the time
Click to expand...

Stock performance please, I havent overclcoked before, so for at least a good while I wouldnt tamper with the card.

I currently have a 3870x2. To put that in to perspective for you, it performs like a 4850 in DX10 and a 4870 in DX9.

I currently play BFBC2 at medium setting at 1920x1200 resolution. A GTX 470 is still gona be a huge leap right?
 
HW91 said:
yeah sure, that's why I'm running a stock cooled 470 with a 33% overclock with max fan speed at 75% and max temp at 78C no matter what game I play. Need proof? I'll get it if you ask

Smoother game play is a well known fact of the 470 vs 5870 battle, ask any one who compared them. And any respectable review site put the 470 with minimums around 5870 level, rarely going below and usually above
Click to expand...

I think you missed the main point though. Most sites use game benchmarks. This site used fraps while playing through levels in the actual games to compare with the earlier reviews using the game benchmarks. They were surprised to find that ati cards in game were much closer in actual gaming not benchmarks to there nvidia counter parts.

the minimum frames showed that a 5850 had better minimum frames than a gtx470 in a good few games and that it was usually always close. This review was actually done because of people always spewing out rubbish about ati cards always having lower minimum fps. We all know that benchmarks matter little and its gameplay that matters. there is no denying that in game using fraps shows what people will be seeing when playing games. the only problem is levels can play different each time you play but surely with a good few games tested this would equal itself out.
 
opethdisciple said:
Stock performance please, I havent overclcoked before, so for at least a good while I wouldnt tamper with the card.

I currently have a 3870x2. To put that in to perspective for you, it performs like a 4850 in DX10 and a 4870 in DX9.

I currently play BFBC2 at medium setting at 1920x1200 resolution. A GTX 470 is still gona be a huge leap right?
Click to expand...

it's not risky or hard, just DL msui afterburner and slide voltage to 1050mv and core clock to 800MHz and test for 5 mins with kombustor then go up from there testing each time until you get a greyish blue screen (at least that's what I get) then go back to the previous option where it didn't crash and play a game (Metro, Crysis, I find COD4 particularly stressful, at least it crashed at lower clocks than Crysis)for as long as possible (ideally an hour but if it isn't ok it will usually crash with 30 mins) If it crashes, restart, lower the clocks 10 or 15MHz and try again.

@ TheRealDeal, ok so you're saying that a most likely completely random test which is less demanding and using a program which uses apparent fps not actual fps is more accurate than a benchmark?

apparent fps can easily sway over 10fps from actual fps, enough evidence for you to discard those reviews?
 
Last edited:
HW91 said:
it's not risky or hard, just DL msui afterburner and slide voltage to 1050mv and core clock to 800MHz and test for 5 mins with kombustor then go up from there testing each time until you get a greyish blue screen (at least that's what I get) then go back to the previous option where it didn't crash and play a game (Metro, Crysis, I find COD4 particularly stressful, at least it crashed at lower clocks than Crysis)for as long as possible (ideally an hour but if it isn't ok it will usually crash with 30 mins) If it crashes, restart, lower the clocks 10 or 15MHz and try again.
Click to expand...

Hmm, well thats sounds easy. I could just leave it at that right, no need to go beyound 800mhz on the core.

Just leave the memory clocks as they are?

If I was to just do this, an nothing else to the card, does it increase the cards performance signficantly?

How significantly will this affect the heat/noise tho?
 
opethdisciple said:
This is very true, but so too has the 5870 dropped in price to £275, which Im also considering.
Click to expand...

I would say also that the fact the gtx470 has dropped so low tells you it can't really compete to well with the 5870 as its not in nv's history to do this sort of thing.


@HW91 this 10fps swing works both ways. They tested quite a lot of games and apart from games that worked better on each company minimum frames were always close. The averages were much closer also. Fraps imo is accurate but could be wrong. Like i said above this site in there initial review using benchmarks what most other sites found but in actual games play they found ati cards to be closer than in benchmarks. I am on holiday using my phone so can't really be bothered finding it but can in a few days. I am pretty sure raven posted it.
 
Last edited:
yes leave the mem clocks as is

and yes it increases performance a lot

well max temps rose from 72C at stock with fan at 72% to 78C and 75%, also I'm using a custom fan profile with the two points at 40,30 and 80,80
 
^ well you might not get 800MHz, but you'll almost definitely get 775MHz

also make sure afterburner opens with windows
 
You must log in or register to reply here.
Back
Top Bottom