• 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.

Benchmarks for AT HD 5870

60hz does not equal 60fps max on an LCD, where did that come from? or have I missed a convo somewhere...

Explain, because unless I'm extremely confused, the monitor can only totally refresh the screen 60 times a second (i.e. 60 'cycles' per second - 60Hz where the cycle is a totally drawn picture)). Otherwise you get screen tearing where the monitor attempts to draw the next frame over the frame it was previously drawing. Sure the computer can render over 60FPS internally, but the monitor can't display them that fast, it can only display the frames up to its maximum refresh rate without artifacting. Am I wrong in this understanding?
 
Explain, because unless I'm extremely confused, the monitor can only totally refresh the screen 60 times a second (i.e. 60 'cycles' per second - 60Hz where the cycle is a totally drawn picture)). Otherwise you get screen tearing where the monitor attempts to draw the next frame over the frame it was previously drawing. Sure the computer can render over 60FPS internally, but the monitor can't display them that fast, it can only display the frames up to its maximum refresh rate without artifacting. Am I wrong in this understanding?

It depends on the refresh rate of the device the card is outputting to, if you're getting stupidly high FPS in a game you can have for example 4 (on a 60hz monitor you'd need to be hitting at least 240fps for this example) different sections of different frames frames being displayed at a time (at different positions on your monitors picture) and you get a kind of shearing effect. The effect is less noticible at higher resolutions.

[edit] and more noticible in games where you're moving quickly.
 
Last edited:
well in the resolution that i'm playing the 295 scored 10 wins and the 5870 only 3. i wont be upgrading my gpu for a while it seems.

the 5870 is still a good single card and a nice step up from the 48xx series but its nowhere close to a 8800gtx kind of thing.

What are you talking about, when either card "loses" you literally wouldn't be able to tell the difference, they are the same speed card and its that simple.

As for the 5870 not being a 8800gtx kind of thing, thats nonsense, sorry but the 8800gtx DID NOT destroy 2x7900gtx's, not even sure it was much faster at all, it was significantly faster than a single 7900gtx. The 5870 is seeming to be a standard increase in power over a 4870 that we've seen in other big generation jumps. We've NEVER seen a new gen single card trash a dual gpu or 2 of the old generation cards in sli/crossfire.
 
188 watts full load per card so under 400watts total, the rest of your system may be pulling another 200watts full load but it's unlikely you'll be full loading all components in a gaming environment - only stress benching but either way would be fine...
 
What are you talking about, when either card "loses" you literally wouldn't be able to tell the difference, they are the same speed card and its that simple.

As for the 5870 not being a 8800gtx kind of thing, thats nonsense, sorry but the 8800gtx DID NOT destroy 2x7900gtx's, not even sure it was much faster at all, it was significantly faster than a single 7900gtx. The 5870 is seeming to be a standard increase in power over a 4870 that we've seen in other big generation jumps. We've NEVER seen a new gen single card trash a dual gpu or 2 of the old generation cards in sli/crossfire.

Yes we have, the 4850 and 4870 were both quicker than a 3870X2, and the X2 was slightly higher clocked than a single 3870.
 
Would it have killed them to use the same scale for each graph and start the scale at 0 not at random points above it. Hugely varies how much one bar appears to be over the other if you're not focusing on the numbers.

Its a fine graph, its not fps, its showing the default speed of a 285/295 and the percentage increase or lag for the new card in each game, it can't start at 0% really, well its not worth it as the cards never less than 20% slower, it would never have 0% of the performance of a GTX.

the graph is very misleading.. quite a few games only show 20% increase or less over a GTX 285 for the 5850

they look like good cards from those results but not brilliant imo

5870 looks pretty good but with a graph so misleading how do you know those results are even reliable?

Ah, but you're forgetting cpu limitations, the card without a single game showing otherwise(from what I can see) is hugely fast when 4xaa is used, and huge gain when using 8xaa. What this says to me is obviously not that the card gets better fps with more AA, its just losing basically no performance at all under 4 or even 8xaa, while the other cards are getting hammered.

What you can draw from this is, with AA, the new ATi cards are without equal, they utterly destroy the current gen with current games at max settings with AA being used. lets be honest, who pays for a £150+ gpu and doesn't want to use higher levels of AA and keep performance.


But the one things these graphs don't show, is cpu limitations which may or may not be limiting the non AA results. Really you can only see CPU limitations by seeing multiple res/aa shots of the SAME game, so you can see for instance, lets say it gets 60fps at 1280x1024, 60fps, at 1680x1050, 60fps at 1920x1200, and 45fps at 2560x1600. You could see it was cpu limited up to that point, and only under a harder load(higher aa/af or higher res) does the gpu become the limit.


But obviously you can't round up so many games in a graph like that, you'd need a graph for every game, those ARE the ONLY way to show a massive number of games and it has its limitations. Will be interesting when we get the Anand and other reviews to see, are these cards just limited at high res and so have bags of spare power over the Nvidia cards so AA/AF doesn't phase them. The other option is they aren't cpu limited, offer similar or better performance with standard settings but have such monumentally good AA hardware onboard that they it just doesn't cost them any performance to turn it on, honestly, either way, they are still way faster with high IQ options turned on which is great for us. Just interesting to see which it is.

To be honest, my gut instinct would be that its simply massive cpu limited in most games without AA/AF and hell, it could even still be cpu limited with 8xAA
 
Yes we have, the 4850 and 4870 were both quicker than a 3870X2, and the X2 was slightly higher clocked than a single 3870.

I said TRASH, not ever so slightly quicker. There ARE games that are faster on a 3870x2 than a 4870, still. The main upgrade was microstutter, or lack thereof, and in several games it is faster, but its not a huge margin. However a single far lower power, cooler and quieter card beats a 3870x2 hands down, not to mention in price.

http://www.bit-tech.net/hardware/graphics/2008/09/25/bfg-tech-geforce-gtx-260-ocx-maxcore/5

http://www.bit-tech.net/hardware/graphics/2008/09/25/bfg-tech-geforce-gtx-260-ocx-maxcore/6

Yes, you'll see it there, faster than a 4850 almost always, close to a 4870 and beating a stock 260gtx.


New generations will hopefully offer double the performance of the previous generation, but expecting it to beat two of the previous generation cores by anything more than a slim margin is simply expecting too much.

Expecting a 5870x2 to be double the speed of a 4870x2/295gtx isn't silly, but expecting a 5870 SINGLE GPU to beat either of those cards in a significant way that makes it worth an upgrade, is simply nuts.

http://www.bit-tech.net/hardware/graphics/2008/09/25/bfg-tech-geforce-gtx-260-ocx-maxcore/9

and yes, theres the 3870x2 beating the 280GTX, the 4850 and everything below.
 
Last edited:
What are you talking about, when either card "loses" you literally wouldn't be able to tell the difference, they are the same speed card and its that simple.

As for the 5870 not being a 8800gtx kind of thing, thats nonsense, sorry but the 8800gtx DID NOT destroy 2x7900gtx's, not even sure it was much faster at all, it was significantly faster than a single 7900gtx. The 5870 is seeming to be a standard increase in power over a 4870 that we've seen in other big generation jumps. We've NEVER seen a new gen single card trash a dual gpu or 2 of the old generation cards in sli/crossfire.

sorry mate but i didnt buy a gtx 295 because my gtx 280 couldnt play games fine. i got it because i wanted to and i found it in a really good price (£270).

i knew from the beginning that the gtx 295 provides me with more than enough fps but if the 5870 was about 20% faster across the board i would consider buying it.

not only its not faster but it loses from the gtx 295 in the majority of games. i dont care about 2560 resolution because thats not what i'm using.
 
sorry mate but i didnt buy a gtx 295 because my gtx 280 couldnt play games fine. i got it because i wanted to and i found it in a really good price (£270).

i knew from the beginning that the gtx 295 provides me with more than enough fps but if the 5870 was about 20% faster across the board i would consider buying it.

not only its not faster but it loses from the gtx 295 in the majority of games. i dont care about 2560 resolution because thats not what i'm using.

Thats simply incorrect IF theses graphs are true and if thats what you're basing your opinion on. These graphs show with AA the 5870 is faster, its really as simple as that. Not only that but the trend is new games are all faster on the 5870, and most of them aren't dx10.1. Its already the faster card, even though not by much but with dx10.1 games and upcoming dx11 games it will in most likelyhood get a much bigger boost. Not to mention losing all problems associated with sli/crossfire and getting a cheaper to run card that finally(I pray its not talk and actually works) Idle's at a sane power consumption.

But thats negating the main point you made and the main point I was talking about, you claiming the performance increase from 4870 to 5870 is nothing like the jump to the 8800gtx, which I claimed was untrue, is untrue and its infact looking to be a larger performance jump.

I have no idea why people believe the 8800gtx was some massive ridiculous leap never before or since achieved.

http://www.anandtech.com/video/showdoc.aspx?i=2870&p=23

x1950 crossfire, faster than a 8800gtx, 7900gtx sli with almost identical performance to the 8800gtx. Oblivion got a much better boost on the 8800gtx, though that was quite a bit down to the memory increase. But again both Battlefield and prey it couldn't beat previous ATi gen in crossfire. The games it offered a massive boost in were few and far between, there were more games that it couldn't beat crossfired x1950's than it could. IT was also slower than the x1950 in crossfire in HL:2 which was probably the best game around at the time.
 
Last edited:
Thats simply incorrect IF theses graphs are true and if thats what you're basing your opinion on. These graphs show with AA the 5870 is faster, its really as simple as that.

well take off your fanboy glasses and have a look again:

GPU: Gtx295 5870

Battleforge 1920 4xAA 8xAF 50.8 47

COH 1920 8xAA 16AF 112.7 103.4

Cod WAW 1920 8x aa 16af 120.8 100.7

Crysis Warhead 1920 4xaa 8x af 48.8 43.6

DMC4 1920 4x aa 8xaf 139.3 99.2

Farcry 2 1920 8xaa 63.9 51.8

Left4Dead 1920 8xaa 16af 157.6 152.3

Prey 1920 4xaa 8xaf 225.9 217.4

Stalker CS 1920(unplayable with AA) 40.5 38.4

Stormrise 1920 4xaa 8xaf 49.2 60.2

Riddick AODA 1920 8xaa 16af 63.9 71.3

Unigine Tropics 1920 4xaa 8xaf 34.7 53.8 (sli profile missing)

UT3 1920 4xaa 8x af 143.6 106.6

Wolfenstein 1920 4xaa 8xaf 126.9 104.9

thats 11 wins for the 295 and 3 wins for the 5870. not only you are so biased that is impossible to reason with you but you even fail to see whats in front of you. i'm done with you.
 
I simply wont be burned by AMD/ATI again, Dont get me wrong given the choice I'm Ati all the way but after taking the plunge on the ill fated 2900XT I simply cant bring myself to buy one of these till at least some of you chaps have come back with some real world results.
 
well take off your fanboy glasses and have a look again:

GPU: Gtx295 5870

Battleforge 1920 4xAA 8xAF 50.8 47

COH 1920 8xAA 16AF 112.7 103.4

Cod WAW 1920 8x aa 16af 120.8 100.7

Crysis Warhead 1920 4xaa 8x af 48.8 43.6

DMC4 1920 4x aa 8xaf 139.3 99.2

Farcry 2 1920 8xaa 63.9 51.8

Left4Dead 1920 8xaa 16af 157.6 152.3

Prey 1920 4xaa 8xaf 225.9 217.4

Stalker CS 1920(unplayable with AA) 40.5 38.4

Stormrise 1920 4xaa 8xaf 49.2 60.2

Riddick AODA 1920 8xaa 16af 63.9 71.3

Unigine Tropics 1920 4xaa 8xaf 34.7 53.8 (sli profile missing)

UT3 1920 4xaa 8x af 143.6 106.6

Wolfenstein 1920 4xaa 8xaf 126.9 104.9

thats 11 wins for the 295 and 3 wins for the 5870. not only you are so biased that is impossible to reason with you but you even fail to see whats in front of you. i'm done with you.

Conversely I can choose resolutions and settings where the 5870 wins every test.

/waves hands and mumbles something about glass houses.
 
Back
Top Bottom