Sign in with
Sign up | Sign in

Turbo Boost And Our Test Platform

Six-Core Analysis, Part 2: Intel Core i7-980X Scaling
By

Both processor flagships from AMD and Intel are equipped with performance-enhancing features that allow the CPUs to increase clock speeds when two requirements are met. First, CPU load has to go through the roof, and secondly, there has to be sufficient thermal headroom for increasing the clock rate. The features, however, are implemented differently.

AMD’s Turbo CORE function only knows one acceleration mode, while Intel implements two (at least on this particular model; other CPUs are more dynamic). The first mode applies when all cores are accelerated (a 133 MHz boost). The second kicks in when only one or two cores are active and can benefit from additional clock speed (up to 266 MHz). The Turbo Boost implementation is more aggressive on Intel’s 32 nm processors, whether in dual- or multi-core models, but it's still notable on the 45 nm Core i5 and i7 processors. Note that Turbo Boost accelerates cores by increasing the CPU’s multiplier within a set range, but the feature can't always take advantage of maximum acceleration if the processor is already operating close to its thermal/power limits.

AMD’s Turbo CORE basically works like a reversed implementation of Cool'n'Quiet, AMD's power saving feature for CPUs. To make a long story short, Turbo CORE exploits thermal headroom if there's sufficient workload demand, and it does it for exactly three cores (unless you alter that through AMD's OverDrive software. In theory, this speaks to Intel’s configuration, since there should be higher clock speeds available if few cores are required. Only limited acceleration is available if all cores are involved, since there would be little headroom left. AMD’s feature, on the other hand, also kicks in when needed, but probably reaches thermal limits quicker because all cores are involved at all times. However, this is just a theory, and we need to put it to the test and directly compare Turbo Boost against Turbo CORE in a different article.

Regarding our test platform, we found that you can't just pick any socket LGA 1366 motherboard and expect to reduce the number of active cores. Fortunately, we found a feature for switching off individual cores on Gigabyte’s EX58-UD4P with the F12 BIOS version. Although this might not be a really important BIOS switch for most users, it's worth exploring, since power consumption does decrease if you switch off Intel cores. This wasn’t the case on our Phenom II X6 1090T test system. Here, idle power remained constant whether one or six cores were used.

Display all 23 comments.
This thread is closed for comments
  • 2 Hide
    The Greater Good , July 7, 2010 6:53 AM
    Nice review. To be honest, I didn't really expect anything less.
  • 2 Hide
    bitplayer , July 7, 2010 6:59 AM
    Well done Tomshardware, the comparison between platforms in respect to core count and memory bandwidth saturation was enlightening; Intel having the upper hand with triple channel memory. When 2-4 cores saturate "global memory" bandwidth, developers should take note and differentiate between tasks which require "global memory" and those best executed in cache.
    Again, well done and thanks.
  • -4 Hide
    chefboyeb , July 7, 2010 7:05 AM
    Nice read Tom... Another reason why I come on here every morning instead of picking up my bible and enriching my soul with some morning devotion... Smdh!
  • 0 Hide
    eddieroolz , July 7, 2010 7:26 AM
    Although I think it was a forgone conclusion that more core count = better, it's nonetheless good to get actual confirmation.

  • 3 Hide
    SchizoFrog , July 7, 2010 7:33 AM
    Can I just ask why in this type of article you never mixed 'professional' apps with more 'home user' apps, namely games? It would be of massive benefit to a variety of consumers to cross reference the pros and cons from both sides in the same article.

    As an example, Bit-Tech just did an article on the benefits of multi-cores in games, but they too didn't mix that article against other more professional apps. So it is difficult to see the best of both worlds.

    Here is the Bit-Tech article for reference.
    http://www.bit-tech.net/hardware/cpus/2010/07/05/how-many-cpu-cores-do-games-need/1
  • 9 Hide
    haplo602 , July 7, 2010 8:19 AM
    I don't get it. Why benching single applications on a multicore system ? You can have many background tasks running and still game on the system. However you are only testing sequential workloads.

    What's the use of having multicore systems when you are NOT using them ? Put a video decode, torrent download, lightroom image exports in the background and then bench a game.

    Or better yet, UPDATE your testing methodology to properly deal with multitask hardware.
  • 7 Hide
    Tamz_msc , July 7, 2010 9:17 AM
    It would have been better if you included the results from the Phenom II X6 test side by side with the results from these tests.
  • -1 Hide
    thejerk , July 7, 2010 11:21 AM
    "Clearly, the difference between one, two, and four cores is significant, while adding cores five and six don't yield the same performance jumps."

    I think you misinterpret your own data. The turning point is actually at three cores, and not four.
    Significantly diminishing returns are seen thereafter.
  • 2 Hide
    angryfingertips , July 7, 2010 12:05 PM
    haplo602I don't get it. Why benching single applications on a multicore system ? You can have many background tasks running and still game on the system. However you are only testing sequential workloads.What's the use of having multicore systems when you are NOT using them ? Put a video decode, torrent download, lightroom image exports in the background and then bench a game.Or better yet, UPDATE your testing methodology to properly deal with multitask hardware.


    I agree with you on this one. Just a few years ago, ok maybe 5 or so, you could not not play a video game and run anti-virus at the same time. Today you can do that with the 4 cores. Multi-core is more than just how fast it gets one thing done, it is how fast it gets multiple things done.
  • 1 Hide
    gfg , July 7, 2010 12:27 PM
    100% bandwiht utilization
    Intel 2 core
    AMD 4 core

    Then ddr3 1333 vs ddr3 2000, is valid for some applications.
    Maybe winrar is a good example:
    Intel good increase up to 2 cores
    AMD good increase up to 4 cores

    Conclusion:
    - memory speed in theory, for the architecture of microprocessors, should provide an increase in performance. Although according to tests is not so, so application developers should take more advantage of the bandwidth provided.

    sorry my english.
  • 2 Hide
    bigbopper66 , July 7, 2010 1:43 PM
    Nice article, and I agree with many of the previous comments on multitasking. You mentioned the lack of multi-threaded applications, but that's just in the windows world. Take a look at most Linux apps, and you will find that they are able to take advantage of the multiple cores, and have been that way for many years, even before multicore processors were mainstream, or even available.
  • 0 Hide
    cablechewer , July 7, 2010 3:03 PM
    I agree with the previous posters about needing to see benchmarks with professional applications. Many of your readers are also in the IT industry and their home gaming rigs often double as lab machines running multiple VMs when they have home work to do (if they are 'playing' in the server world) or they run other professional software for when work, training or schooling follow them home.

    The only question is what professional apps to test and does anyone at Tom's have the expertise with those apps (Compilers? Databases? Mail Servers? CAD? Multiple, large, inter-linked spreadsheets when they recalc?)? Is there another site you could partner with? I just did a casual search and didn't find any other sites (like Tom's, but based on Professional software), but I didn't spend a lot of time on it yet.
  • 0 Hide
    infodan , July 7, 2010 3:22 PM
    subject to how many tabs you have open, title Not work safe!
  • -1 Hide
    lapoki , July 7, 2010 4:22 PM
    infodansubject to how many tabs you have open, title Not work safe!

    LMAO
  • 5 Hide
    ta152h , July 7, 2010 4:35 PM
    The first article on scaling with six-cores had me scratching my head. Then to follow it with this one is even more bizarre. So, we came to the conclusion that if you buy a six-core processor, you should leave them all on?

    Or with multi-threaded apps, six cores can make a difference. With limited threads, it's not as effective?

    Maybe I'm a simpleton, but I kind of figured that without doing all the tests. What's next, a test showing that if you buy a keyboard, it generally makes sense to leave all the keys on there and not remove them, because if you do a lot of typing, you'll probably end up using them?

    Most of the articles on Tom's make sense. These seem to come to an obvious conclusion that are pretty much understood without even testing.
  • -1 Hide
    wolfram23 , July 7, 2010 4:39 PM
    Seems logical, but I would have liked to see results from turbo boost.
  • 0 Hide
    awood28211 , July 7, 2010 4:49 PM
    When are we going to see benchmarks that show two (or more) applications running at once. What if I wanna do a movie conversion at the same time I use winrar to compress files?
  • -1 Hide
    iqvl , July 7, 2010 6:05 PM
    haplo602I don't get it. Why benching single applications on a multicore system ? You can have many background tasks running and still game on the system. However you are only testing sequential workloads.What's the use of having multicore systems when you are NOT using them ? Put a video decode, torrent download, lightroom image exports in the background and then bench a game.Or better yet, UPDATE your testing methodology to properly deal with multitask hardware.

    The difference between 4-core and 6-core when many applications run at the same time will be very similar to when one 6-threaded application is run. For example, when one 6-threaded and one 4-threaded apps run at the same time, the difference of time taken to complete the tasks between 4-core and 6-core will be very close to when one 6-thread apps is run if they are as fast in four or less threaded apps(i.e. quad i7 v.s. hex i7, instead of quad i7 v.s. PII-X6).

    Update:
    This is the reason why PII-X6 loses to quad i7 when most apps run at the same time are four or less threaded(especially after all 6-thread of X6 are occupied) which is the most common cases(90%+) to the majority. Forget about the hex, save the money towards SSD, more RAM or better GFX for better overall performance.
  • -1 Hide
    elbert , July 7, 2010 7:36 PM
    Very good article. Well presented but I have a different conclusion. Seems to me the only real advantage going from 4,5,6 cores is mostly simulation. Simulation being many programs running at once. If running many programs is your thing then more cores the better for you. I mostly run 1~3 programs as most all programs require some human involvement. I cant see adding more CPU cores in the future benefiting the average PC users.

    The very best program multi-threaded seems to produce little benefit beyond 4 cores. I'm looking forward to the APU with hope it solve the overhead. The efficiency test seem to really highlight the drop off beyond the 4th core.
  • -1 Hide
    Anonymous , July 7, 2010 7:44 PM
    Quote:
    WinZip needs a serious update

    Uh, dude,
    Winzip is not used anymore!
    Winrar, and later 7z took over the field!
    The few zip files you find on the internet are done by guys not knowing much of compression, if they lived anywhere after 2002, they would have chosen Winrar, and anywhere after 2008 they would have chosen to compress them with 7Z (which uses Lha compression).
    Lha exists already for a long time, but was dos based. 7z has a good enough graphical interface for the leek to understand, without having to type cryptic codes in a DOS styled window.
Display more comments