Closed

5970 Review

First one I have seen:

http://www.hardwarezone.com/articles/view.php?cid=3&id=3072

(Certainly not what I call a quality review, but it is the first of what should be many over the next couple of days)
22 answers Last reply
More about 5970 review
  1. Note that the review uses a rather (crippling) small benchmark set. The drivers were 8.66RC4.

    Overclocking on their sample (due to what I'm not sure) was terrible. Likely a driver issue, or AMD's slides about 1ghz are from the future 5990.. lol

    New ati drivers should be out by the time anyone uses these things, though I am sick of having to wait for a re-review of a card :)
  2. There is always the possibility, that the 5970 is using lower binned parts that otherwise wouldn't cut it as a 5870, but better than a 5850, knowing that lower clocks would be expected due to thermal issues. Or perhaps intentionally crippled so as not to cannibilize 5870 sales in the long term (one 5970 is 25% cheaper than 2x 5870s)

    Perhaps we will see a 4gb 5990 surface whenever Duke Nukem Fermi decides to exist, based on higher binned parts and better cooling. Who knows, fun to dream :D
  3. daedalus685 said:
    Note that the review uses a rather (crippling) small benchmark set. The drivers were 8.66RC4.

    Overclocking on their sample (due to what I'm not sure) was terrible. Likely a driver issue, or AMD's slides about 1ghz are from the future 5990.. lol

    New ati drivers should be out by the time anyone uses these things, though I am sick of having to wait for a re-review of a card :)



    Agreed not a good review, i mean they were told there was a vapour chamber so they obviously havent had the card to bits so whats the chances they dont know about the voltage chips at all and were trying to OC it on the stock voltages ?

    Mactronix
  4. Also, consider my previous rantings about cpu limitations in the cpu section.
    We need faster cpus, better more challenging games
  5. mactronix said:
    Agreed not a good review, i mean they were told there was a vapour chamber so they obviously havent had the card to bits so whats the chances they dont know about the voltage chips at all and were trying to OC it on the stock voltages ?

    Mactronix


    That would be priceless if they tried on stock voltage.. lol

    As for CPU limitation, aye we are there.. but I want to see a comparison at equal clocks to 5870 crossfire. Based on this one, the OC gave almost no improvement, indicating either a pcie bottleneck, or they did something funny. Release is supposed to be tomorrow, so real reviews should pop up tonight.

    I can't agrue with how we need better games...
  6. Ive heard 2 things, cpu bottlenecks and possibly memory limitations/timings.
    Also could be drivers as well. Some games it works, others...
    I say its drivers cpu, but time will tell
    Also, very astute about voltage. Here we have world class electronics for the best to push/oc, and if they didnt raise the power, this too could be it.
    The only problem Im having with all this is, if you look at the power usage of the 295, its using 44 watts less, both at stock, leaving alot of headroom, where the 5970 teeters on the pci com pliancy numbers, while the 295 exceeds them
  7. JAYDEEJOHN said:
    Ive heard 2 things, cpu bottlenecks and possibly memory limitations/timings.
    Also could be drivers as well. Some games it works, others...
    I say its drivers cpu, but time will tell
    Also, very astute about voltage. Here we have world class electronics for the best to push/oc, and if they didnt raise the power, this too could be it.
    The only problem Im having with all this is, if you look at the power usage of the 295, its using 44 watts less, both at stock, leaving alot of headroom, where the 5970 teeters on the pci com pliancy numbers, while the 295 exceeds them


    They are estimating through total system power, no real reason this has to be accurate. The real power draw of the 5970 may very well be less than the TDP spec, where as that for teh 295 may be closer. Also, there is lots of possible error in the measurement. Before i read to much into it I'd like to see a test done by anand, hardware canucks, and so on.

    I'm not seeing how a CPU bottleneck could hamper the 5970 any more than 5870's in crossfire though, pcie aside.

    We will surely find out soon.

    By the way, I realised i made this a question thread... oops ;)
  8. daedalus685 said:
    They are estimating through total system power, no real reason this has to be accurate. The real power draw of the 5970 may very well be less than the TDP spec, where as that for teh 295 may be closer. Also, there is lots of possible error in the measurement. Before i read to much into it I'd like to see a test done by anand, hardware canucks, and so on.

    I'm not seeing how a CPU bottleneck could hamper the 5970 any more than 5870's in crossfire though, pcie aside.

    We will surely find out soon.

    By the way, I realised i made this a question thread... oops ;)



    Got to agree with that, we just need some decent benchies to sort it all out.
    My selection though if i was guessing would be drivers. Seen those same type differances in graphs of new products too many times now ;) usually sorted within the first couple of updates.

    Mactronix
  9. Not all the games showed a difference tho.
    And yes, for the most part, tho it was a small sampling, it showed what it should otherwise.
    No mention of cpu limitations, saying it was only 20% faster than the 295 etc. Therell be much better reviews.
    I do know that memory and or cpu and drivers are the suspects currently.
    http://www.xtremesystems.org/forums/showpost.php?p=4113991&postcount=352
    Notice HWZones OC on the ram was only 1100, not 1200, and poor for core as well.
    Im thinking its BW starved a lil, plus drivers and occasional cpu limations.
  10. Slightly off topic JD but let me just run this idea past you its sort of related.
    Do you think its possable that the auto overclocking that the i5 and i7 are capable off could be cocking up the internal timings between the CPU-GPU-Monitor relationship ? I have seen a few guys posting here and other places who are getting what seems to be Micro stutter on a single card. All were running i7's and 5800 series cards.

    Mactronix
  11. That goes against the grain tho.
    If the QPI or whatever, or connect speeds are raised, it should mean better output consistancy, as the timing should be tighter.
    At least it should be less noticeable, not more
  12. Well that's what i was thinking but thought i would ask someone a bit more knowledgeable than me about it. Just something that came up as a possable. My thinking was that it would be kinda like triple buffering, having more not less resource.
    Thanks, anyway guess we wait now for the reviews to flood in, drat got work for the next few days :fou:

    Mactronix
  13. The only thing I can think of, is possibly its not syncing with the monitor speeds, and skipping frames altogether, looking like MS.
    Did any of these complaints try Vsync for corrections?
    Also, with this tho, youd think youd see tearing as well
  14. ok here it comes again....
    it's last gen vs this gen

    but this time it's Nvidia's own damn fault they got raped
  15. also....the overclock was pathetic.....

    "And then there is its massive launch price, a wallet-busting US$599 (S$830). This is US$200 more than a Radeon HD 5870, and is enough to buy two Radeon HD 5850 cards with change to spare."
    why is he comparing 2 5850's to 2 5870's?
    that's not exactly in the same league is it.....
    I don't think he realized the price of the 5850's went up to $300,-
  16. I think itll show they did the pooch on this review
  17. JAYDEEJOHN said:
    I think itll show they did the pooch on this review


    "Oh no, our review is balls! what will we do?

    Well, if you post it a day early maybe you will get hits anyway! First is first.

    Oh hey, yeah. Maybe if we write tomorrows date on it noone will notice that we broke the NDA either!"


    And the sad thing is... being first has gotten them hits as we have no other preliminary numbers to look at.
  18. Another great review: http://www.guru3d.com/article/radeon-hd-5970-review-test/

    Extensive testing and well written IMO
  19. Hilberts good. Wonder when the competitions going to show?
    This just put huge pressure on nVidia
  20. It's about time...I feel that recently nVidia has been sort of sitting back thinking "how can we improve our income without spending too much time on R&D.."

    The recent 210, 220, 240 series is really sort of a joke if you ask me. They are not bad cards, don't get me wrong--but they are overpriced for what they truly are. If they released these for less money, with the current performance, I think they would sell much better.

    Seems any "budget" nVidia GPU can barely keep up with a midrange ATI that is actually cheaper.

    Makes me sad, always used nVidia products--but they just are not a smart choice in terms of price/performance as of late...
  21. The real sad part is, seeing AMDs new cpus with IGPs attached, theyll have more power than those nVidia gpus.
    They arent out yet, but soon, these little crumbs will no longer be useful
  22. Sorry but we already have a 5970 review thread,please follow that:
    http://www.tomshardware.com/forum/274507-33-official-5970-reviews
Ask a new question

Read More

Radeon Graphics