1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Hardware ATI Radeon HD 6870 Review

Discussion in 'Article Discussion' started by Guest-16, 22 Oct 2010.

  1. Mentai

    Mentai What's a Dremel?

    Joined:
    11 Nov 2007
    Posts:
    758
    Likes Received:
    1
    Cheers impar.

    @Lizard

    That reasoning is fine with one caveat. I assume that people will be playing Crysis 2 when it comes out, which I'm also assuming will be a good replacement for the first (benchmark wise). Could you really not benchmark Crysis 1 until that point? And if not, could you at least do a performance comparison when Crysis 2 comes out so we can see how to judge it.

    Thanks.
     
  2. Combatus

    Combatus Bit-tech Modding + hardware reviews Lover of bit-tech Super Moderator

    Joined:
    16 Feb 2009
    Posts:
    2,761
    Likes Received:
    89
    It makes perfect sense. High average and maximum frame rates are pointless if a graphics card dips to 10fps on occasions. I want smooth gameplay all the time and I'd choose a card that does a min of 30fps and avg of 40fps vs min of 20fps and average of 50fps any day. This is also a great way of highlighting driver problems and so identifying products you might not want to buy if you play specific games. This is why we filter the graphs by the minimum frame rate.
     
  3. Hustler

    Hustler Minimodder

    Joined:
    8 Aug 2005
    Posts:
    1,039
    Likes Received:
    41
     
  4. impar

    impar Minimodder

    Joined:
    24 Nov 2006
    Posts:
    3,109
    Likes Received:
    44
    Greetings!
    It makes sense to order by minimum FPS if all the cards dips on the same benchmark the same times, the same order of magnitude.
    I ask you both, on my previous example*, which card would be ranked best? And why would it be ranked lower than the worse card?
    Its not a matter of "last time Crysis was played", its a matter of consistency. Players know Crysis, everybody played it, old reviews show Crysis numbers, they dont show Arma II numbers, making comparisons between old cards and new cards impossible.

    *- "A card that has 50 FPS average and one dip to 20 FPS, gets ordered bellow a card that has 40 FPS average and ten dips to 25FPS?"
     
  5. Lizard

    Lizard @ Scan R&D

    Joined:
    17 Feb 2007
    Posts:
    2,890
    Likes Received:
    37
    I'd hesitate to commit bit-tech to testing a game that hasn't even been released yet, so for now the jury is out on Crysis 2.

    That said, we will of course keep an eye on Crysis 2, so if it's popular and hardware intensive (and therefore revelant) we will include it in the benchmarks suite.
     
  6. leveller

    leveller Yeti Sports 2 - 2011 Champion!

    Joined:
    1 Dec 2009
    Posts:
    1,107
    Likes Received:
    24
    A little note: see how confused everyone is in this thread with these ridiculous naming schemes from ATi and Nvidia? Also a quick look around all the other popular sites reveals exactly the same issues with the naming. Nvidia and ATi need to stop it now and just stick to one theme for a long while. It's beyond a joke. And bloody annoying.

    Another note: I use a 4870x2 and after a quick scan of replies just to this thread, others are using: 4870 (crossfired), 4890 (crossfired), 260GTX (SLi'd), X1800XT (Xir ... upgrade?! lol), and then you have other using 460's 470's etc. But in the comparisons you don't include: 4870, 4870x2, 4890, 260GTX and certainly not the X1800XT (!). So how do we know how these new cards compare to our old?

    Can you guys devise a way to include previous results? Obviously the test rigs would need to stay the same between different cards.
     
  7. RonanH

    RonanH mod-envious

    Joined:
    30 Mar 2010
    Posts:
    89
    Likes Received:
    2
    I'm all for not using the "Crysis Test". That game isn't a good indication of performance, not because it's such a graphically intense game but because it's just really badly coded and poorly optimised.

    If modern cards still struggle with a 3yo game, there's something wrong with the game, not the hardware.

    As to the hardware reviewed, for that money they look like an awesome buy. What's especially promising is that AMD seem to have finally gotten their crossfire scaling right, the performance of 2x 6870 is amazing. Bet nVidia are worried, and not just at the high end, the £130 segment should now belong to AMD too.
     
  8. impar

    impar Minimodder

    Joined:
    24 Nov 2006
    Posts:
    3,109
    Likes Received:
    44
  9. Krayzie_B.o.n.e.

    Krayzie_B.o.n.e. What's a Dremel?

    Joined:
    2 Aug 2009
    Posts:
    427
    Likes Received:
    6
    For me I stopped using Crysis as the benchmark of benchmarks a while ago seeing how Metro 2033 with all the eye candy on with full Tessellation can only be run by a GTX 480 SLI set at 33 FPS.

    Granted Crysis did raise the bar but being DX10 it's no longer the bar as the DX11 Metro 2033 takes the crown as of now.

    I just don't understand why so many people are confused about the new AMD naming scheme? As if Bit-tech needs to give these guys a crayon colored chart showing....... awe just forget it here...................

    http://en.wikipedia.org/wiki/Compar...g_units#Northern_Islands_.28HD_6xxx.29_series

    Memorize it as there be a test in the morning
     
    Er-El likes this.
  10. Yoy0YO

    Yoy0YO Lurky Lurker

    Joined:
    20 Mar 2009
    Posts:
    100
    Likes Received:
    4
    Performance was a bit of a let down, price is rather optimistic. Thanks for the article BT. I think AMD should've just dropped the prices of their current 5xxx series with more numbers saving the 6xxx for 32nm fabbing. I do look forward to their 28nm stuff.

    Thanks BT.
     
  11. sp4nky

    sp4nky BF3: Aardfrith WoT: McGubbins

    Joined:
    15 Jul 2009
    Posts:
    1,706
    Likes Received:
    53
    I'm not confused. It seems pretty simple to me that they've gone back to their naming scheme from 2 generations ago, so the 6870 is the equivalent, albeit with a significant performance upgrade, to the 4870.

    If you only look back one generation, you may be confused, but I think most people here have longer memories than that.
     
  12. Senilex

    Senilex Not a Troll. Honest.

    Joined:
    11 Dec 2009
    Posts:
    55
    Likes Received:
    1
    It's a pretty poor decision by Bit-Tech then to continue to call them by a name that is WRONG. It doesn't matter if internal workings of their driver still says ATI the cards are AMD Radeon. Period. Every other major Hardware site refer to them by their correct name of AMD Radeon. By referring to them as their correct name, you wont be confusing your reader you will be educating them.
     
  13. [USRF]Obiwan

    [USRF]Obiwan What's a Dremel?

    Joined:
    9 Apr 2003
    Posts:
    1,721
    Likes Received:
    5
    So why did BT not replace Crysis for Metro as ultimate benchmark for DX11 hardware.

    To me it seems this is another card into the crowded 'midrange' section that competes with 20 other cards for give or take 4 frames per second of game play.

    I hope all other cards that are made before 2010 will drop 30/80 euro in price now or just die..
     
  14. Xir

    Xir Modder

    Joined:
    26 Apr 2006
    Posts:
    5,412
    Likes Received:
    133
    Yah...I've been waiting for a year for the 5870 pricing to move...
    Which it didn't, so a year ago would have been a great time to buy one. :duh:
    With this 6870 "Mainstraem" card and the upcoming enthousiast cards...looks like I'll be waiting for about three months more :D

    Now you Understand why I haven't played Crysis...yet.

    Speaking of which
    it's not good for playing, but great for comparing :D
     
  15. knuck

    knuck Hate your face

    Joined:
    25 Jan 2002
    Posts:
    7,671
    Likes Received:
    310
    Lot's of bitching in this thread. What's going on ? Are the stars aligned in a bad way ? Or are people actually bitching because the reviews they've been reading for years (for free!) are not written the way they want them to be ?
     
    Material likes this.
  16. Stewb

    Stewb What's a Dremel?

    Joined:
    16 Dec 2009
    Posts:
    600
    Likes Received:
    17
    Sorry, I am absolutely idiotic, I typo'd 6870 rather than 6970, sorry for the hassle of having to write that out
     
  17. Fizzban

    Fizzban Man of Many Typos

    Joined:
    10 Mar 2010
    Posts:
    3,691
    Likes Received:
    275
    I'm not confused at all. I got used to daft naming policies when I got interested in PC components. And this makes more sense than some.

    5770 is similar in performance to a 4870. You can work it out from there. Just read the reviews man.
     
  18. Claave

    Claave You Rebel scum

    Joined:
    29 Nov 2008
    Posts:
    691
    Likes Received:
    12
  19. Nikols

    Nikols What's a Dremel?

    Joined:
    13 Oct 2009
    Posts:
    443
    Likes Received:
    16
    I flogged my two 5870's in expectation of a price drop this month and am now waiting impatiently for the 6990 to arrive! Pleasantly surprised in the interim at how well the gtx280 I bought for a €120 s/h for physx is now holding up as my main gpu at decent res'

    6870 - Mid range card that is almost on par with a 5870. That's job done imo. Now give us the high end before I spend the money!
     
  20. Deadpool46

    Deadpool46 What's a Dremel?

    Joined:
    28 Jul 2009
    Posts:
    27
    Likes Received:
    0
Tags: Add Tags

Share This Page