Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

i5 2500k memory controller question!

drbaltazardrbaltazar Member UncommonPosts: 7,856
That processor as two memory channel right? So can user set bios so 100% of cpu access to one memory controller and/or the other

(A LA ddr4) I would say single channel or unganged but I ain't sure ,also I believe the e3 1230 v2 is this way .I ain't sure ,does intel supply this optionally?this would still be dual channel but unganged ,but not really unganged either because in Unhanged not 100% of cpu have access to channel one and/or two.

Comments

  • stragen001stragen001 Member UncommonPosts: 1,720
    I dont know the answer to your question, but why in the name of all that is holy would you want to intentionally gimp your system like that?

    Cluck Cluck, Gibber Gibber, My Old Mans A Mushroom

  • RidelynnRidelynn Member EpicPosts: 7,383


    Originally posted by drbaltazar
    That processor as two memory channel right? So can user set bios so 100% of cpu access to one memory controller and/or the other (A LA ddr4) I would say single channel or unganged but I ain't sure ,also I believe the e3 1230 v2 is this way .I ain't sure ,does intel supply this optionally?this would still be dual channel but unganged ,but not really unganged either because in Unhanged not 100% of cpu have access to channel one and/or two.

    image

    Your CPU has 100% access to both channels. They are ganged for better throughput when you have memory installed in compatible channels.

    They are not core specific. Nor would you want them to be because you would pretty much destroy the purpose for having a shared L3 cache. You already have a ring-style interconnect, so pretty much each core gets full access when it needs it (barring a traffic jam, which is going to be rare at the low core counts Intel has available in the consumer space)

  • QuizzicalQuizzical Member LegendaryPosts: 25,531
    If you want a processor to only use one of the memory channels, that's pretty easy to do:  don't put any memory in the other channel.  It's also pretty stupid to do.  Which is why some OEMs will do it for you, especially in laptops.
  • RidelynnRidelynn Member EpicPosts: 7,383

    I also thought of a nother way you could restrict your memory to a specific CPU core.

    The BIOS usually has the option to enable or disable per core on the CPU. If you disable all the cores you don't want to have access to the memory controller, only the ones you leave enabled will be able to access your memory channels.

  • ToxiaToxia Member UncommonPosts: 1,308

    I know its really off topic, but ive been curious how the I5-2500k's are stacking up these days to newer cpu's?

    My build is from 2010 i believe, so i was wondering if i should replace my 2500k first or my xfired 6870's.

     

    The Deep Web is sca-ry.

  • RaunuRaunu Member UncommonPosts: 480
    Originally posted by Toxia

    I know its really off topic, but ive been curious how the I5-2500k's are stacking up these days to newer cpu's?

    My build is from 2010 i believe, so i was wondering if i should replace my 2500k first or my xfired 6870's.

     

    xfire your 6870s. Of course your 2500k will end up being weak point in the chain, but you can always upgrade that later.

    - - "What if the hokey pokey really is what it's all about?" - -

  • RidelynnRidelynn Member EpicPosts: 7,383


    Originally posted by Toxia
    I know its really off topic, but ive been curious how the I5-2500k's are stacking up these days to newer cpu's?

    My build is from 2010 i believe, so i was wondering if i should replace my 2500k first or my xfired 6870's.

     


    Keep your 2500 in my opinion. Drop the XFire and go to a single nice card, you'll see a lot more improvement from that across the board than anything else.

    New CPUs are more power efficient, and tend to be a bit faster at stock speeds, but Sandy is still the overall overclocking champ (as in more of those CPUs can more reliably overclock to mostly higher levels safely), and if you are pushing your 2500k versus an Ivy or Haswell via overclocking, they all tend to pretty much equal out at that point. The big improvements have mostly been laptop-oriented... more power efficient with more powerful integrated video (neither of which really matter in a desktop to a large degree).

    I still can't see a point in upgrading a quad-core 920 Bloomfield, and that's the generation before your 2500k and over 600Mhz slower on stock with an anemic turbo frequency... but it's a nice overclocker, and when you do that, it changes the equation and all of a sudden it's still a pretty nice chip and a capable performer.

  • drbaltazardrbaltazar Member UncommonPosts: 7,856
    Better then a stock i5 2500k ? E3 1230 v2
    Ok ,so I can have my i5 2500k unganged ? How's I do this? I would end up with 2 non interleaved Unhanged set up (just like ddr4)
  • RidelynnRidelynn Member EpicPosts: 7,383


    Originally posted by drbaltazar
    Better then a stock i5 2500k ? E3 1230 v2Ok ,so I can have my i5 2500k unganged ? How's I do this? I would end up with 2 non interleaved Unhanged set up (just like ddr4)

    The DIMM sockets on the motherboard are paired - like pairs are a matched memory channel.

    A typcal motherboard setup would look like:
    {A}{A}{B}{B}
    Where A is channel 0, and B is channel 1, and all the DIMMS in channel 0 are ganged together, and the DIMMS in channel 1 are ganged together (provided they are matched pairs with a common size and clock speed).

    To ungang them, you just don't use the second ganged DIMM slot of each pair:

    Most sane people, when installing 2 DIMMS, would put them both into the A slots. This gangs them and allows them to get roughly double the throughput. If you put one in an A slot, and one in a B slot, they can't gang, because they are on physically different channels, and you end up with the same amount of memory, each on different channels, but with half the throughput.

    I don't think motherboards will allow for non-matched pairs in a common channel - so the only way you could prevent channels from ganging would be via leaving DIMM slots vacant.

    So Yes, you can do it. No, it wouldn't be like going from DDR3 to DDR4. It would be like going from DDR3 to DDR2.

  • QuizzicalQuizzical Member LegendaryPosts: 25,531
    Let's ignore the computer side of things and make an economics argument instead.  If you could increase your performance on existing hardware with no drawbacks, don't you think that every memory, motherboard, and CPU manufacturer in the world would eagerly explain how you can get more performance out of their hardware?
  • RidelynnRidelynn Member EpicPosts: 7,383


    Originally posted by Quizzical
    Let's ignore the computer side of things and make an economics argument instead.  If you could increase your performance on existing hardware with no drawbacks, don't you think that every memory, motherboard, and CPU manufacturer in the world would eagerly explain how you can get more performance out of their hardware?

    Of course not - they would want to suppress that information via an elaborate cover up to compel you to go out and buy new hardware; even though it's all a hoax and your current hardware from several years ago obviously has the same capability (after all, they invented all of this based on alien technology recovered at Rosewell back in the 1950's, the CIA has been using it for decades in their mind control machines). It's a conspiracy in order to get sheep-like consumers to participate in an endless cycle of needless hardware upgrades in order to perpetuate an economic machine that serves no master except an elite aristocracy, namely the Bilderburg Group, but also several world governments who use it as a pacification tool in order to keep their populations under control.

  • GruntyGrunty Member EpicPosts: 8,657
    Originally posted by Ridelynn

     


    Originally posted by Quizzical
    Let's ignore the computer side of things and make an economics argument instead.  If you could increase your performance on existing hardware with no drawbacks, don't you think that every memory, motherboard, and CPU manufacturer in the world would eagerly explain how you can get more performance out of their hardware?

     

    Of course not - they would want to suppress that information via an elaborate cover up to compel you to go out and buy new hardware; even though it's all a hoax and your current hardware from several years ago obviously has the same capability (after all, they invented all of this based on alien technology recovered at Rosewell back in the 1950's, the CIA has been using it for decades in their mind control machines). It's a conspiracy in order to get sheep-like consumers to participate in an endless cycle of needless hardware upgrades in order to perpetuate an economic machine that serves no master except an elite aristocracy, namely the Bilderburg Group, but also several world governments who use it as a pacification tool in order to keep their populations under control.

    "I used to think the worst thing in life was to be all alone.  It's not.  The worst thing in life is to end up with people who make you feel all alone."  Robin Williams
  • drbaltazardrbaltazar Member UncommonPosts: 7,856
    I use bing I search YouTube or Google and get 20 billion result instead of 18 billion .why ? (Changed to flex channel )this is a 10% increase in search but still .that's a huge difference .was surprised I expected same result as tomhardware(margin of error difference)
  • RidelynnRidelynn Member EpicPosts: 7,383

    I know when I'm searching for stuff, the problem usually isn't a lack of results, it's being able to find the one result I really need.

    How did you get there from memory controllers anyway

  • miagisanmiagisan Member Posts: 5,156

    every time i read the OPs posts, i want to break something.

     

    Its like he finds some random tidbit of knowledge on the web and thinks he finds the holy grail, no matter how bad of an idea it is or how unrelated it is to what he is trying to accomplish.

    image

  • HulluckHulluck Member UncommonPosts: 839
    Originally posted by Ridelynn

    I know when I'm searching for stuff, the problem usually isn't a lack of results, it's being able to find the one result I really need.

    How did you get there from memory controllers anyway

    I like derails. image  That statement is so true. 

  • RidelynnRidelynn Member EpicPosts: 7,383

    http://www.intel.com/support/motherboards/desktop/sb/CS-011965.htm

    So ok I found the link between memory controllers and his search post. Either this is the lamest troll post ever, or it's the most hilarious thing I've ever seen in my life, and the funny part isn't the post, the funny part is that I, for the life of me, can't figure out which it is.

  • drbaltazardrbaltazar Member UncommonPosts: 7,856
    OK forget the start .found the answer ,but I brought more question.like,why when I bing say YouTube or Google in flex channel I get more result (10% more)
  • RidelynnRidelynn Member EpicPosts: 7,383

    Dude, your computer isn't out there searching the internet. Your RAM speed, and in fact your entire computer, has zilch to do with search results from the internet.

  • jdnewelljdnewell Member UncommonPosts: 2,237
    Originally posted by drbaltazar
    OK forget the start .found the answer ,but I brought more question.like,why when I bing say YouTube or Google in flex channel I get more result (10% more)

    Does it matter?

    Did you really tinker with the memory channels so you could get 20 billion instead of 18 billion results on a google search!?

    Especially when out of the 20 billion 19.9999 billion are useless results.!!

    You seem to enjoy dicking around with your PC to try and get performance gains that are inconsequential at best. And thats all good as its probably your hobby and it is your money and PC.

    If nothing else it is entertaining. =)

Sign In or Register to comment.