Sign in with
Sign up | Sign in

Experiment: Comparing Four Quad-Core Architectures At 2.8 GHz

Experiment: Comparing Four Quad-Core Architectures At 2.8 GHz
By

AMD and Intel are relentless when it comes to diversifying their respective CPU portfolios across every possible corner of the processor market where someone might want to spend money. The good news is that these efforts give us lots of technology options across the entire price spectrum.

But buyers who don’t follow the daily cadence of processor development couldn’t possibly know whether Core i7 or Core 2 Quad is the newer product, or how these compare to AMD's own line of obscurely-named models. In some ways, it doesn’t matter which chips were launched most-recently. The more important consideration might be which processor offers the best total performance relative to its peers, and one of the best ways to judge this is with a shoot-out at a given clock rate.

The Issue with Variety

Ten years ago, it was really easy to stay up to date on the latest processor offerings and know what to buy. You had Intel's various product offerings at their different frequencies, and AMD's own counterparts. Today, the game is much more complex. Performance is no longer defined just by clock speed; core count and performance per clock are equally important. In addition, specifics such as cache capacities, as well as bus and memory speed, vary the parameters and hence complicate direct comparisons. Let’s not forget that it’s also important to take features such as virtualization technology and power efficiency into consideration. Intel, in particular, is guilty of selectively removing value-adds like VT-x from some models, while leaving it in others, without making the distinction clear.

Brand Wildfires

Things were relatively easy when there were only three or four brand families to track. Pentium, Celeron, Athlon, Sempron--easy! Today, though, the chip families have expanded and sprouted multiple lines within each brand. It practically requires a CPU workshop to get familiar with all of the names, features, and platform specifics currently available (Ed.: I'm glad Patrick is bringing this up, too; it's a point I harped on in our Clarkdale coverage and bears repeating).

AMD still offers Semprons for the entry level. Turions and Athlons, available as X2s or Neos, are for mobile platforms. Athlon, Athlon II, Phenom, and Phenom II power desktop PCs, but I’ll skip the details at this point, because you need to look at various specifics, including core count, features, and clock speed to properly order all models according to your own priorities. Consider perusing AMD’s Find and Compare feature list of nearly 250 processors.

Intel doesn’t make it easier, as its portfolio is even larger. Celerons and Core 2 processors power notebooks (as do Mobile Core i7s, i5s, and i3s now). Atom is there, too, as a lowest-cost option for both desktops and portables. Core 2, Core i3, Core i5, and Core i7 form the desktop CPU lineup, backed by Pentiums and Celerons at the low-end. Intel’s ARK (Automated Relational Knowledgebase) helps to investigate and compare the company’s processors.

Resetting the Game to 2.8 GHz

We decided to grab some of the latest quad-core mainstream and high-end processor offerings and do a toe-to-toe comparison. This time we didn’t look at market segment as defined by AMD and Intel. Instead, we selected a clock speed that all contenders can run at—2.8 GHz—and we performed benchmark runs at that speed.

Display all 46 comments.
  • 0 Hide
    Anonymous , 15 January 2010 13:32
    It would be interesting to see compared systems at same price (probably would make sense if price included motherboard+processor).
  • 0 Hide
    mactronix , 15 January 2010 14:31
    As far as Gaming goes the results dont represent real life at all. At a resolution that someone with a CPU like the ones used there is really very little differance in performance at all. In fact the 965 would win some benchmarks at a sensable resolution.

    Mactronix
  • 0 Hide
    snipe0876 , 15 January 2010 15:10
    what was wrong with the resolution shown i dont think to many people play at those res anymore with the current line up of monitors everyone has higher
  • 0 Hide
    13thmonkey , 15 January 2010 16:48
    doesn't the low resolutions push the game to be cpu bound and hence remove any chance of the gpu bottlenecking the result? hence the use of very low resolutions.
  • -2 Hide
    Anonymous , 15 January 2010 19:05
    THG constantly manage to make AMD cpu's look a lot worse than they are don't they?

    HINT : AMD already has a 2.8ghz cpu, USE THAT ONE INSTEAD OF DOWNCLOCKING THE FLAGSHIP YOU TARDS.

    How much better a slant would this entire article have on AMD cpu's if they'd simply used a X4 925 instead of the X4 965?

    SHEESH is it any wonder THG gets accusations of bias on every bloody article???
  • -3 Hide
    Anonymous , 15 January 2010 19:10
    http://www.lostcircuits.com/mambo//index.php?option=com_content&task=view&id=44&Itemid=42

    Please read this article before doing any more benchmarks on intel power consumption.
  • 0 Hide
    Chad Boga , 15 January 2010 19:23
    ohreallyTHG constantly manage to make AMD cpu's look a lot worse than they are don't they?HINT : AMD already has a 2.8ghz cpu, USE THAT ONE INSTEAD OF DOWNCLOCKING THE FLAGSHIP YOU TARDS.How much better a slant would this entire article have on AMD cpu's if they'd simply used a X4 925 instead of the X4 965?SHEESH is it any wonder THG gets accusations of bias on every bloody article???

    THG and every website which shows the true state of CPU performance(i.e. that AMD are behind Intel) will always cop accusations of bias from AMD fanboys who are addicted to denying reality.
  • -2 Hide
    Anonymous , 15 January 2010 19:41
    Chad BogaTHG and every website which shows the true state of CPU performance(i.e. that AMD are behind Intel) will always cop accusations of bias from AMD fanboys who are addicted to denying reality.


    Funny how even a massively underclocked Phenom II can still beat an i7 in most of the gaming benchmarks.
  • -2 Hide
    Chad Boga , 15 January 2010 19:46
    ohreallyFunny how even a massively underclocked Phenom II can still beat an i7 in most of the gaming benchmarks.

    In a GPU limited situation, it appears that the PhII can compete with the i7, but once that bottleneck is removed, then the disparity in performance between each processor becomes apparent.
  • -1 Hide
    Anonymous , 15 January 2010 19:51
    Chad BogaIn a GPU limited situation, it appears that the PhII can compete with the i7, but once that bottleneck is removed, then the disparity in performance between each processor becomes apparent.


    A gtx260 gpu limited at 1280x800 on medium settings? I don't think so, no.

    The i5 manages ok, it's just further proof of the i7's gaming failings - and this time you can't even use gpu limited as the excuse.
  • 0 Hide
    mactronix , 15 January 2010 22:14
    Chad BogaTHG and every website which shows the true state of CPU performance(i.e. that AMD are behind Intel) will always cop accusations of bias from AMD fanboys who are addicted to denying reality.


    Yes AMD are behind Intel when all around performance is taken into account but the thing Intel is up on is encoding etc. For everyday usage in the real world, that's gaming and the odd bit of encoding where 18 seconds difference doesn't matter to anyone AMD are totally in the race.
    Low power consumption is nice but it isnt a reason to buy one over the other. The differance is exactly the same thing as buying a diesel car over a petrol one because the running costs are lower, that differance being nothing. The milage you need to do before the fuel costs start to out way the extra you paid is huge. Same as you would get to the point where you are buying a new PC before the low power consunption of the intel would make a differance over teh extra you paid for the hardware.
    In fact im strongly looking at AMD for my next build.

    Mactronix
  • 0 Hide
    codefuapprentice , 15 January 2010 22:17
    i just find it funny how people consider the GTX 260 a weak GPU, but in terms of performance it's holding its own in the top 10 cards from nVidia(from what i can make out)
    Also, the Phenom ii 965 isn't exactly miles behind the i7s, it's clearly slower than the I7s in many benchmarks which involve sheer number crunching muscle, but in gaming benchmarks, AMDs flagship cpu is on par with the I7 regardless of only using a Dual channel memory configuration.

    I personally can't go above 1366x768 in resolution(using a 720p HDTV)
    so its nice to see benchmarks done using 1280x1024
  • 1 Hide
    mactronix , 15 January 2010 22:27
    Exactly if you want to crunch numbers and worry about power usage then get an i5/i7
    If you want a general purpose/gaming PC get an AMD CPU. At a decent resolution with a top end card gaming wise the 965 actually beats an i7 in some games.

    Mactronix
  • 0 Hide
    codefuapprentice , 15 January 2010 22:41
    Imagine what AMD could do with that Triple channel controller...

    But more to the point, in the 10 years i've been using AMD CPUs
    (From K6-III 400 > Duron 750 > Athlon xp 1800 > Athlon 64 2800(754) > AMD Athlon 64 3700(939) then to Phenom II x4 955) and i've noticed in benchmarking AMD have favoured 3D application performance since the end of the Super 7 Era, intel couldn't keep up with the athlon Slot A.
    And as for pricing at the moment, you can't get a more ferocious CPU than the PII x4 965 for £135 (Retail boxed CPU)
  • 0 Hide
    Chad Boga , 15 January 2010 22:41
    ohreallyA gtx260 gpu limited at 1280x800 on medium settings? I don't think so, no.The i5 manages ok, it's just further proof of the i7's gaming failings - and this time you can't even use gpu limited as the excuse.

    The result for the i7 in that particular test seems a bit odd as it doesn't correspond to the results that Tom's got when they tested all four processors previously at 2.8Ghz with 4870x2's as the GPU

    http://www.tomshardware.co.uk/core-i5-lynnfield,review-31672-9.html

    So it does appear that the lack of GPU power distorted the results in AMD's favour for that one result(even though the i5-750 handily beat it), but with plenty of GPU grunt, of course the better CPU shines through.
  • -1 Hide
    Anonymous , 15 January 2010 22:46
    Quote:
    The result for the i7 in that particular test seems a bit odd as it doesn't correspond to the results that Tom's got when they tested all four processors previously at 2.8Ghz with 4870x2's as the GPU

    http://www.tomshardware.co.uk/core-i5-lynnfield,review-31672-9.html

    So it does appear that the lack of GPU power distorted the results in AMD's favour for that one result(even though the i5-750 handily beat it), but with plenty of GPU grunt, of course the better CPU shines through.


    Ye sure...what was it the i5 had over the i7 again? On one hand you claim the better cpu shines through with plenty of gpu grunt, but you can't explain why the i5 is scoring so much higher?

    Face it, the i7 is a lemon of a gaming cpu.
  • 1 Hide
    Chad Boga , 15 January 2010 22:52
    Quote:
    Ye sure...what was it the i5 had over the i7 again? On one hand you claim the better cpu shines through with plenty of gpu grunt, but you can't explain why the i5 is scoring so much higher?

    Face it, the i7 is a lemon of a gaming cpu.

    There appears to be something wrong with that particular benchmark as you don't see those results anywhere else.

    Besides the other Tom's link I showed you(which you obviously couldn't understand), you don't see that result on other sites.

    Here The Tech Report using a GTX260 shows the i7 at stock speeds beating a 3.4Ghz Ph II in Left for Dead on averagey settings

    http://techreport.com/articles.x/17545/7

  • 1 Hide
    mactronix , 15 January 2010 23:21
    No actually the lack of GPU power distorts things in Intels favour.
    If this test was done with a 5850 the gaming results would be level across the board.

    Mactronix
  • 1 Hide
    Chad Boga , 15 January 2010 23:37
    mactronixNo actually the lack of GPU power distorts things in Intels favour.If this test was done with a 5850 the gaming results would be level across the board. Mactronix

    Check out this review with a 4870x2, it does not produce the results you suggest.
  • 1 Hide
    Chad Boga , 15 January 2010 23:38
    mactronixNo actually the lack of GPU power distorts things in Intels favour.If this test was done with a 5850 the gaming results would be level across the board. Mactronix

    http://www.tomshardware.co.uk/core-i5-lynnfield,review-31672.html

    Check out this review with a 4870x2, it does not produce the results you suggest.
Display more comments
React To This Article