No announcement yet.

Is ECC RAM worth it for a desktop PC?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Is ECC RAM worth it for a desktop PC?

    I'm planning a big upgrade for a desktop PC which is 6 years old already.
    The thing I've found it was more limited was in RAM quantity (only 2GB).
    My most memory demanding tasks now require about 8GB of RAM (without the OS, programs, etc).
    Since I want to be future-proof, I've decided to max out whatever is the maximum RAM in the new system (32GB)
    I'm currently considering to buy an Asus Crosshair V Formula Z, and an AMD FX-8350 CPU.
    Since both components support ECC RAM, and my budget seems to be enough to buy it, I'm considering that option.

    But, is it really worth it?
    How common are bit-flip errors in modern RAM?
    Is the performance drop noticeable?

    ECC RAM availability in my zone is pretty rare, but I've found this kit that may work:
    KVR16E11K4/32 (4x8GB DDR3 ECC 1600Mhz CAS 11)

    It isn't listed in the motherboard docs thought, so it's a bit risky. I haven't contacted Asus yet.

  • #2
    I've got 12 GB of RAM, and never seen a problem. I can't give you exact figures, but I don't think that ECC RAM has much point unless you've got some mission-critical setup, in which case you'd go for more expensive server / workstation hardware anyway most likely.

    But as for it not being on the support list, I don't see that it is going to be a problem. The memory controller is on the CPU anyway. And memory incompatibility errors are very rare in my (limited) experience, as long as it is the right type (DDR3, unbuffered or registered, non-ECC or ECC).


    • #3
      Well it depends. When a board has got issues with ram you will notice real bad crashes, ecc ram would at least notify you that there are problems (and could correct 1 bit errors). I would not say that you need 32 gb ram however, thats usually overkill. 8 gb seems to be enough for most workloads, if you want to create linux live images in ram completely maybe get 16 gb.


      • #4

        This is valid question and a valid answer is YES. Your configuration will support it and its VERY CHEAP solution.

        There is no such thing as "mission critical application", simply because if there would be no problem - there would be no solution.
        Server environments usually feature much higher electromagnetic field interferences, this is why to prevent bit-flips ECC memory is critical there.
        However current desktops, especially since DDR2, especially when highly loaded with banks of memory, can result in bitflipping. But virtually since its design the ECC memory was not used on desktop simply because it costs more.

        I would definately go for this protection level, because it will spare you corrupted binaries, data and reboots. The possibility that this happens is present, the technology is cheap, and there is not much point why you shouldn't do this.

        However, please gather as much information as possible to be sure the memory you buy will be able to maintain its speed if you install many banks. There might be limitations. AMD officially supports ECC, so you might also ask you mainboard vendor on possible limitations.

        32GiB of RAM is high-range by today standards,but some insert upto 64GiB in desktop. This depends upon your real need.. so I assume you really need that amount. Otherwise, it will be just wasted money. Its up to you.

        Btw, I will move to ECC fairly soon as well following exactly same route


        • #5
          Originally posted by PreferLinux View Post
          I've got 12 GB of RAM, and never seen a problem.
          That's quite a paradox there. In order to see the problem if there is one, you need ECC RAM. But if you don't have ECC RAM, there's no way to see it.


          • #6
            If you don't store any data on the PC itself and use it only as browsing / video streaming client, then you don't need ECC. In fact, using ECC will make the system slower due to the additional error detection calculations.

            But once you start storing data on it, ECC becomes important. This has been shown in a Google publication from 2009:
            For example, we observe DRAM error rates that are orders of magnitude higher than previously reported, with 25,000 to 70,000 errors per billion device hours per Mbit and more than 8% of DIMMs affected by errors per year.
            DRAM errors can cause nasty data corruption on disk, which can be catastrophic (if important filesystem structures are affected) or silent. An error rate of 8% per DIMM per year (for server memory, which usually sits on a well-designed mobo behind high-quality PSUs) is not negligible.


            • #7
              I am also looking at these specific parts for a new build. My current box is similar: PhenomII x6 running at 3.9GHz and 16G of DDR3-1333 ECC which I run at 1600. I use the EDAC module and utilities so I will know if my RAM becomes faulty, before I go and destroy some important data. I've been running my box for a few years nonstop - I only restart to change kernels. I'm very happy with the quality of these Kingston DDR3.


              • #8
                ECC all the way

                +1 for ECC RAM, but you know what sucks ? that High-Performance RAM makers (such as corsair, G.Skill etc..) don't make them
                and there is no way I would go for a 1333/1600 mhz DDR3, I want 2800 Mhz ECC DDR3 RAM


                • #9
                  Hello everyone!

                  I am new to this forum. I like computer hardware discussion very much. There are many interesting and knowledge able discussion threads in this forum. The forum community have knowledge and sense about the discussion topics.


                  • #10
                    Originally posted by chithanh View Post
                    An error rate of 8% per DIMM per year (for server memory, which usually sits on a well-designed mobo behind high-quality PSUs) is not negligible.
                    It means you might get one bit corrupted on a four-DIMM system in three years. That's probably an issue for a server that's on 24/7 and processing important data, but not for a desktop PC.

                    There was also an interesting study a few years ago which I think was mentioned on a Solaris mailing list; they found that some DIMMs had no errors over the duration of their study, while those that had even one error tended to have quite a few. The errors weren't randomly distributed across all DIMMs, so they could minimise errors by doing a burn-in test and tossing any DIMMs that showed any errors.