Sign in with
Sign up | Sign in

Conclusion

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

Admittedly, this shootout highlights an artificial condition. The direct comparison of various processors at a given clock speed doesn’t consider other important facts, such as market segments, platform constraints, available clock speeds, product cost, or overclocking margins. However, looking at the results of AMD and Intel processors running at the same core clock speed allows us to strictly define their differences at a set baseline.

Regular readers are aware of the situation on the processor market. Intel has the performance and the efficiency crown beyond any dispute, and this has a direct impact on processor pricing. Intel Core processors are always rather expensive. AMD has concentrated on providing maximum value based on its current portfolio, making Athlon II and Phenom II attractive options through aggressive pricing. A specific recommendation in favor or against certain options must only be given after consideration of platform aspects and total cost. Power savings on an Intel system are not worth a lot if the additional hardware cost exceeds the saved amount on the electricity bill (which it sometimes does).

As expected, Intel’s Core i7 on LGA 1366 is the undisputed performance winner at 2.8 GHz. There are a few benchmarks that run slightly better on Core i5 and LGA 1156, but these aren’t too significant in the big scheme of things. The Core 2 Quad is the direct competitor to AMD’s Phenom II X4 but it lags noticeably behind.

The efficiency battle is dominated by the Core i5 due to its 45nm fabrication process and feature cuts that do not have much impact in the real world. Core i5 lacks both the i7’s triple-channel memory controller and its Hyper-Threading, neither of which are necessary for excellent performance. Our Core 2 Quad processor also did very well, but we have to remind you that we used an S-model, which is the power efficient variant. Regular Core 2 Quad processors show significantly higher peak power.

Finally, we encourage you to take the results in stride. AMD’s Phenom II X4 isn’t as poor as the performance results at 2.8 GHz might suggest. You get faster clock speeds with AMD than on the Intel side for very little money, and once again we’d like to point at the fact that it’s the result that counts. Intel is in a comfortable position, but if AMD manages to deliver higher clock speeds and comparable performance at equal or lower cost, I’d say there is nothing wrong with that.

Ask a Category Expert

Create a new thread in the Reviews comments forum about this subject

Example: Notebook, Android, SSD hard drive

Display all 156 comments.
This thread is closed for comments
Top Comments
  • 23 Hide
    chainsaw667 , January 15, 2010 7:15 AM
    How about a price per dollar test.
    There is no doubt that Intel has the faster chips, but we're in a recession and dollar value is a key factor. Example:
    cheapest Quad core i5 199 newegg
    cheapest Athlon II X4 99.00
    cheapest phenom II x4 140.00
    Sure you can get a core 2 quad for 149.00 but that is a dead socket which is why it was left out of the lineup, socket stability adds great value, hear me Intel. Even the newest Phenom II's can plug into any am2+ or better board, Intel has what 2, no 3 sockets for desktops, and they are not compatible. 775 had a great run and were an awesome value, but they went back to multiples again. Virtualization on almost, if not all AM2+ or better CPU's at no charge.

    Even a Sempron II will run most basic software just fine, gamers are a different story and already know which chip to buy for max OC and have no concern for this type of comparison the results of which were obvious to any regular reader before testing. AMD has the extra cores for the lucky adding even more value and nice overclocking.

    Not a fanboi of anything but current and future value, right now it remains to be seen how long the 2 current sockets from Intel will coexist nor how long they will last. But someone with a 3 year old AMD system can add more memory, and go to a new quad for the price of the newer budget i3. Upgrading a three year old intel system to a quad core is even more than the i3, and ends with that motherboard.

    Competition is needed, that and the value offered by AMD systems (virtualization, socket stability, unlockable cores, overclocking) which keeps me coming back. my $.02.

    Try this test take the cheapest intel i5 system, build an amd based system with the same price and then test the systems.
    i5 quad core+2x2GB ddr3 1333+ECS H55H-m=377.00
    phenomII940 quad+MSI770 g45 Mobo+6GB ddr3 1333=380 the money saved on the processor gives a motherboard with higher memory capacity, dual PCI x16 slots and 2 GB more memory.

    The best processor is not always the fastest.
    Value fanboi
  • 23 Hide
    shubham1401 , January 15, 2010 5:45 AM
    Nice article.
    I conclude that All quads at their respective prices perform good.

    Even the aging Core 2 Quad put impressive results.



  • 21 Hide
    gsacks , January 15, 2010 6:57 AM
    Quote:
    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.


    I read this statement and stopped right there. I knew immediately that Intel would "win". I haven't even read the conclusion, but I'm sure it pretty much says that intel wins in all categories. Sorry, but that is a quite stupid methodology. Comparing processors at a similar price point had real meaning because everyone can relate to the 'bang for the buck' argument. Comparing processors at a similar heat dissipation has meaning because most people can relate to the value of a quiet and not overheating system. Comparing processors at a similar power consumption has obvious value for mobile computing. Comparing processes based solely on benchmark results has value if performance is all you care about. But clock rate comparisons between disparate products is pointless at best and at worst it reinforces the false notion that a faster running chip is
Other Comments
  • -3 Hide
    burnley14 , January 15, 2010 5:36 AM
    It looks like Intel pretty well ran the table on AMD.

    But before I jump to conclusions, I'd better wait to hear the outcries from the AMD fanboy camp about how unfair the testing was, how illegitimate the results are, etc.

    Everybody knows AMD sells cheaper chips, but this testing shows that Intel sells faster chips. They are aiming at two different market sectors, end of story.
  • 23 Hide
    shubham1401 , January 15, 2010 5:45 AM
    Nice article.
    I conclude that All quads at their respective prices perform good.

    Even the aging Core 2 Quad put impressive results.



  • 10 Hide
    nzprogamer , January 15, 2010 6:37 AM
    burnley14It looks like Intel pretty well ran the table on AMD.But before I jump to conclusions, I'd better wait to hear the outcries from the AMD fanboy camp about how unfair the testing was, how illegitimate the results are, etc.Everybody knows AMD sells cheaper chips, but this testing shows that Intel sells faster chips. They are aiming at two different market sectors, end of story.


    agree if im short on $ but need a 4 core to play with i will still go with the AMD.
  • -5 Hide
    zdzichu , January 15, 2010 6:38 AM
    AES encryption test looks like AES-NI wasn't used. Could you confirm?
  • -6 Hide
    nzprogamer , January 15, 2010 6:39 AM
    anyway hope intel does not end up like the AMD K7-745/939 socket.
  • 10 Hide
    amgsoft , January 15, 2010 6:49 AM
    Nice article, seems to support my own idea that Intel and AMD are targeting different marked segments. Intel is about horsepower in a box and AMD is about value for money, which is usually enought for the majority of the users. If you need more out of AMD, you can always OC it; at additional cost, i.e. the electricity bill.
  • 21 Hide
    gsacks , January 15, 2010 6:57 AM
    Quote:
    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.


    I read this statement and stopped right there. I knew immediately that Intel would "win". I haven't even read the conclusion, but I'm sure it pretty much says that intel wins in all categories. Sorry, but that is a quite stupid methodology. Comparing processors at a similar price point had real meaning because everyone can relate to the 'bang for the buck' argument. Comparing processors at a similar heat dissipation has meaning because most people can relate to the value of a quiet and not overheating system. Comparing processors at a similar power consumption has obvious value for mobile computing. Comparing processes based solely on benchmark results has value if performance is all you care about. But clock rate comparisons between disparate products is pointless at best and at worst it reinforces the false notion that a faster running chip is
  • -7 Hide
    andy5174 , January 15, 2010 7:12 AM
    It will be much more reasonable to have i7-860 as the representative of LGA1156 platform when comparing with LGA1366 platform, considering the 860 is on the same price level as the 920.
  • 23 Hide
    chainsaw667 , January 15, 2010 7:15 AM
    How about a price per dollar test.
    There is no doubt that Intel has the faster chips, but we're in a recession and dollar value is a key factor. Example:
    cheapest Quad core i5 199 newegg
    cheapest Athlon II X4 99.00
    cheapest phenom II x4 140.00
    Sure you can get a core 2 quad for 149.00 but that is a dead socket which is why it was left out of the lineup, socket stability adds great value, hear me Intel. Even the newest Phenom II's can plug into any am2+ or better board, Intel has what 2, no 3 sockets for desktops, and they are not compatible. 775 had a great run and were an awesome value, but they went back to multiples again. Virtualization on almost, if not all AM2+ or better CPU's at no charge.

    Even a Sempron II will run most basic software just fine, gamers are a different story and already know which chip to buy for max OC and have no concern for this type of comparison the results of which were obvious to any regular reader before testing. AMD has the extra cores for the lucky adding even more value and nice overclocking.

    Not a fanboi of anything but current and future value, right now it remains to be seen how long the 2 current sockets from Intel will coexist nor how long they will last. But someone with a 3 year old AMD system can add more memory, and go to a new quad for the price of the newer budget i3. Upgrading a three year old intel system to a quad core is even more than the i3, and ends with that motherboard.

    Competition is needed, that and the value offered by AMD systems (virtualization, socket stability, unlockable cores, overclocking) which keeps me coming back. my $.02.

    Try this test take the cheapest intel i5 system, build an amd based system with the same price and then test the systems.
    i5 quad core+2x2GB ddr3 1333+ECS H55H-m=377.00
    phenomII940 quad+MSI770 g45 Mobo+6GB ddr3 1333=380 the money saved on the processor gives a motherboard with higher memory capacity, dual PCI x16 slots and 2 GB more memory.

    The best processor is not always the fastest.
    Value fanboi
  • -7 Hide
    chainsaw667 , January 15, 2010 7:20 AM
    Forgot the raid on the AMD mobo too.
  • 9 Hide
    blacksins , January 15, 2010 7:28 AM
    burnley14It looks like Intel pretty well ran the table on AMD.But before I jump to conclusions, I'd better wait to hear the outcries from the AMD fanboy camp about how unfair the testing was, how illegitimate the results are, etc.Everybody knows AMD sells cheaper chips, but this testing shows that Intel sells faster chips. They are aiming at two different market sectors, end of story.

    and everybody knows that AMDs Phenom has a higher clock speed, thats how it manage to compete with the i5 and sometime the mighty i7, and look at the Core 2 Quad clock speed its only dropped by 0.03 GHz!! not like the Phenom which is dropped from 3.4 to 2.8, actually you knew that thats why you said that some fanboy will come and say that the article was unfair.. who's the fanboy now :) 
  • 5 Hide
    randomizer , January 15, 2010 8:31 AM
    andy5174It will be much more reasonable to have i7-860 as the representative of LGA1156 platform when comparing with LGA1366 platform, considering the 860 is on the same price level as the 920.

    But then you'll be comparing it to AM3 where the Phenom II 965 is $70 cheaper. Is that also "reasonable"? You can't compare all platforms here equally on price, that's just the fact of the matter.
  • -9 Hide
    JeanLuc , January 15, 2010 8:39 AM
    chainsaw667How about a price per dollar test.There is no doubt that Intel has the faster chips, but we're in a recession and dollar value is a key factor. Example:cheapest Quad core i5 199 neweggcheapest Athlon II X4 99.00cheapest phenom II x4 140.00.


    Well you can easily do that for yourself. Look at the performane scores on the last page.

    Core i5 = 1072 points @ $199 = $5.39 per point
    Phenom II = 929 points @ $140 $6.64 per point

    Based on that (and if you agree with the aggregate set out on toms scoring) AMD perceived 'value for money' is actually a false economy as although the AMD chip cheaper you get less performance per $ then you do with the Intel Core i5 chip.
  • 6 Hide
    randomizer , January 15, 2010 8:47 AM
    JeanLucBased on that (and if you agree with the aggregate set out on toms scoring) AMD perceived 'value for money' is actually a false economy as although the AMD chip cheaper you get less performance per $ then you do with the Intel Core i5 chip.

    Personally I think synthetics should have been left out of the aggregate "score" system.

    Another thing to remember is that only the i7 920 and i5 750 will be running at 2.8GHz when you buy them, so the "value for money" argument is still valid, and probably shouldn't be discussed much in an article where some chips have been hamstrung to fit some criteria.

    Yes, you get less performance/$ when running the exact same configuration and applications as were used in this review. But if you buy a Phenom II and then downclock it to 2.8GHz you must be puffing the good stuff
  • 18 Hide
    Anonymous , January 15, 2010 8:51 AM
    Actually your math is all wrong..is 199/1072=0.186 $/point and 140/929=0.15 $/point, so it is not a false economy
  • 7 Hide
    amgsoft , January 15, 2010 8:58 AM
    gsacksI read this statement and stopped right there. I knew immediately that Intel would "win". I haven't even read the conclusion, but I'm sure it pretty much says that intel wins in all categories. Sorry, but that is a quite stupid methodology. Comparing processors at a similar price point had real meaning because everyone can relate to the 'bang for the buck' argument. Comparing processors at a similar heat dissipation has meaning because most people can relate to the value of a quiet and not overheating system. Comparing processors at a similar power consumption has obvious value for mobile computing. Comparing processes based solely on benchmark results has value if performance is all you care about. But clock rate comparisons between disparate products is pointless at best and at worst it reinforces the false notion that a faster running chip is


    Well, I don't completelly agree with this view. Remember that for few years ago it was actually AMD claiming that the have faster/more effective CPU's per clock cycle and put a lot of effort to sell their CPU's on that feature (Sempron 3000+ was a 2GHz CPU). It seems to be the history now, Intel has regained the throne back in efficiency. Apart of an ideological battle/discussions for and agains a certain vendor, reminding AMD of their current position in the field now can actually result in more competition between Intel and AMD and thereby better and faster CPU's for us.

    I still think that AMD has a lot to offer to the majority of the customers, Intel is somehow in front and lets see, how AMD responds. At least at the graphical marked they was able to show what they can with the 58x0 series.
  • 6 Hide
    memeroot , January 15, 2010 9:08 AM
    very interesting article, there are enough bang per buck reviews in the world, good to have a different take.
  • 9 Hide
    gti88 , January 15, 2010 9:15 AM
    i7 860 chip + mobo = 440$ (in Moscow)
    Phenom II x4 925 + AMD 785 mobo = 250$
    i7 is about 25% faster, but it's 75% more expensive.
  • 2 Hide
    chainsaw667 , January 15, 2010 9:33 AM
    gti that 190 could be spent on dual vga, or drives for a raid array or an ssd. Which sounds like a value to me.
  • 11 Hide
    Reynod , January 15, 2010 9:35 AM
    Were the synthetic tests all compiled using Intel's old Compiler ... the pro-Intel / anti-AMD one??

    You can imagine how confident we are with the results from any synthetic tests now ... they aren't worth considering until that issue has been sorted out.

    Would have been interesting to put in a Phenon I cpu ... stretched an old black to 2.8 ... with a few squeals I imagine.

    The Penryn core sure has done well too.

    Well done guys.
Display more comments