CPU scaling in games with dual and quad core processors

Discussion in 'Frontpage news' started by Guru3D News, May 15, 2008.

  1. gmavignier

    gmavignier Ancient Guru

    Messages:
    1,537
    Likes Received:
    22
    GPU:
    RTX 2060
    Very nice comparison. Only thing I would like changed is the color code. Keep it constant trough all the artichle. I got lost after a while when the Q6600 changed color.

    One more thing...
    Well put. Maybe we really should have one of these as it keeps the forums clear from quad X duo threads.
     
  2. roguesn1per

    roguesn1per Ancient Guru

    Messages:
    9,508
    Likes Received:
    0
    GPU:
    GTX580
    Joey you mean GPU....

    These test prove that most if not all games are GPU bound....not CPU. a better GPU will do alot more than a better cpu.
     
  3. tucsonrider

    tucsonrider Master Guru

    Messages:
    302
    Likes Received:
    0
    GPU:
    Sapphire HD6870 1GB
    What of dual vs quad at similar clock speeds?

    Say in Crysis, a Duo 2 @ 3.2ghz vs a Quad @ 3.2 GHz.

    Would the performance stay roughly the same, or would one win out over the other?
     
  4. Denial

    Denial Ancient Guru

    Messages:
    13,565
    Likes Received:
    3,118
    GPU:
    EVGA RTX 3080
    Yea..

    Great article btw, it really shows how processors are becoming less and less important for gaming.
     

  5. mike41

    mike41 Maha Guru

    Messages:
    1,208
    Likes Received:
    0
    GPU:
    Evga 560ti
    I Hope Finch Agrees With This
     
  6. scipio

    scipio Ancient Guru

    Messages:
    4,636
    Likes Received:
    9
    GPU:
    Evga 1080 gaming
    ????
     
  7. Classic Satch

    Classic Satch Banned

    Messages:
    4,777
    Likes Received:
    0
    GPU:
    2xEVGA GTX 260 216 SC
    I've suspected this all along and that's why I haven't bothered to do a total upgrade yet. I agree with a previous poster that Supreme Commander would have been possibly the best choice for testing in that article given it's built-in benchmarking utility.
     
  8. Kollunz

    Kollunz Master Guru

    Messages:
    321
    Likes Received:
    0
    GPU:
    Radeon 5870
    hmm I guess I should save my money for now and just get a 6400+ instead of a new 775 board and a c2d and call it a day (or a year).
     
  9. Cybermancer

    Cybermancer Don Quixote

    Messages:
    13,795
    Likes Received:
    0
    GPU:
    BFG GTX260OC (192 SP)
    That would be the more cost-effective solution. ;)

    ...and maybe get a new graphic card this summer? :D
     
  10. Anaksulnamun

    Anaksulnamun Member

    Messages:
    17
    Likes Received:
    0
    GPU:
    Onboard 64Mb
    Great post :D Now I have more confidence when buying a new compy: just a good big LCD and a 8800GT will do it great for gaming...
    For RAM: 2gb or 4gb, everyone (maybe 4gb for the (not yet to be seen) future ;) )
     

  11. Stukov

    Stukov Ancient Guru

    Messages:
    4,899
    Likes Received:
    0
    GPU:
    6970/4870X2 (both dead)
    I have been saying this forever, unless you are playing games at 1280x1024 or lower there is no point in getting a faster CPU, the extra money you would better off putting in a GPU...unless you win the lottery then get a skull trail so your balls can feel satisfied.

    The only games it matters are the RTS's like Supereme Commander and such. Most likely due to the calculations and AI.
     
  12. NoviceRei

    NoviceRei Ancient Guru

    Messages:
    2,185
    Likes Received:
    0
    GPU:
    Inno3D 460 825/1079/1650
    Company of Heroes is also a bit heavy on the CPU..
     
  13. nutyo

    nutyo Ancient Guru

    Messages:
    4,589
    Likes Received:
    1
    GPU:
    Sapphire Vapor-X HD5870
    Cheers for these awesome comparisons.
     
  14. Drunken Shooter

    Drunken Shooter Ancient Guru

    Messages:
    1,799
    Likes Received:
    4
    GPU:
    EVGA 2070 super FTW
    What I don't understand is why the qx6850 and qx9770 did not match in speed
    given the one processor is 200mhz more but you can clearly OC the qx6850 to that speed easy. so what giving the qx9770 more horse power?
     
  15. Anaksulnamun

    Anaksulnamun Member

    Messages:
    17
    Likes Received:
    0
    GPU:
    Onboard 64Mb
    Cache ;) that is the answer ;)
    Well, for some games (Like the Heroes V) what is b4 ur eyes is based on the GPU's horse power, but when we end turn, la la la, the CPU + RAM will do the job...
    Well, it's hard to afford a better system than Hibert used to test those ten CPUs...
    So well, plz be satisfied with the result, mates :D
     

  16. gx-x

    gx-x Ancient Guru

    Messages:
    1,509
    Likes Received:
    156
    GPU:
    1070Ti Phoenix
    Good point. I was just about to post something similar. The article is good but it is not of much relevance to people that overclock.

    Basically, here we have a comparison of 12mb cache penryn vs 8mb cache core2quad at different speeds and different FSB/strap. Also we have comparison od wolfdale with 6mb cache vs old core2duo with 4mb also at different speeds.
    I see this test more as CPU test rather than quad vs dual...
    I would like to see apples to apples comparison with all processors clocked at same speed.
    C2Q @3.2 will always be faster than any other core2 at 2.66 or 3ghz. But it is not always going to be faster (I think it will rarely be faster in today's games) compared to core2duo at 3.2ghz. I also think (I am 99% sure) that core2duo @ 3GHz will be always faster than core2quad @2.8GHz.
    I am yet to see a game that really benefits from 4 cores.

    Thanks for reading this post :)
     
  17. gmavignier

    gmavignier Ancient Guru

    Messages:
    1,537
    Likes Received:
    22
    GPU:
    RTX 2060
    Agreed. Apples VS oranges here but that does not discredit the comparison. For none overclockers (belive it or not they are a majority) this is just what they need to know: raw power is better than having multiple cores unless you run one of the few programs multi core enabled.

    There is never going to be a real 1X1 comparisson because the new 45nm CPUs are different than the old 65nm. No way even to compare raw core power. The only conclusion here would be the new CPUs are better than the old ones.
     
  18. Pelle V

    Pelle V New Member

    Messages:
    4
    Likes Received:
    0
    GPU:
    Sapphire Radeon X1950 Pro 512 AGP
    When you are doing

    When you are doing those tests why cant one part have the same couler through the whole test .........







     
  19. Sieras

    Sieras Ancient Guru

    Messages:
    3,815
    Likes Received:
    0
    GPU:
    Gigabyte GTX 970 G1 4GB
    Great article :)
     
  20. Norvekh

    Norvekh Ancient Guru

    Messages:
    2,678
    Likes Received:
    12
    GPU:
    EVGA RTX 2080 Ti
    The QX9770 has more cache, but most importantly, it runs on a 1600MHz FSB.

    A lot of factors outside of the GPU can affect performance, processor being most important but even the choice in motherboard, but especially memory, can affect performance. Some games really like higher memory bandwidth while others aren't quite so bottlenecked.

    The only thing I'm disappointed about with the review is a lack of testing of any Source (or really, any CPU-bound) games. I'm rather interested to see a CPU comparison in that engine simply because it is more CPU-limited than most game engines out there and it is used in so many games and mods. Another one I noticed that seems to be multi-threaded is Test Drive Unlimited. When I exit the game and review CPU history it's all over the place with two cores, and the other two aren't idling but aren't full usage either. I guess my point is that there are more CPU-bound games, in which case having a quad can bring a boost (saying apples to apples clocks) due to being able to process the system data as well as the games two or more threads.
     

Share This Page