RAM Question

I recently had a problem where my computer was shutting down randomly 5-20 mins into games that were very heavy on my CPU/GPU if you saw my other thread. I found that I could remedy this by removing 1 of 4 RAM cards (all 4 the same, Corsair XMS3 8GB (4 x 2GB) 240-Pin DDR3 1333 Desktop Memory Model CMX8GX3M4A1333C9.)

I assumed first of all that one of my RAM cards had an error, so I ran memtest on each stick individually for 7 passes. I found no errors on any of the sticks, as well as when I ran memtest on all 4 sticks at once. I also ran the windows memory diagnostic, which found nothing as expected.

I'm asking now, if anyone knows why this would occur when 8 GB's are in place and not when there are 2, 4 or 6?

PS - The motherboard is an ASUS P7P55D PRO.

7 answers Last reply Best Answer
More about question
  1. I believe this is your RAM

    4 sticks can sometimes require more V to keep things stable. Memtest is good for checking that there are no problems with the RAM. So it's good that they passed that test. But you need a better stress test to make sure you're system is stable. Have you tried running Prime95 blend? Be sure to enable Round Off Checking in the advanced menu.

    Did you make any changes to your RAM settings in BIOS? What settings are you running at? Multiplier, timings, V.
  2. ^+1...

    Couple of additional questions...
    1) What is the make & wattage of your PSU? A weak/defective PSU can cause your issue. Your PSU may just be at the edge of it's limit & by removing the slight voltage of a RAM stick keeps it from crashing.
    2) What are your system temps? CPU, GPU ... Etc. Heat can play a factor in your system shutting down and the GTX 280 is know to run hot.
  3. @First Poster First of all, how do I see my multiplier, timings, and voltage of my RAM? I have not tried running 95 Prime, because I don't understand how it works.

    @Second Poster My system temps are pretty low for the most part besides the videocard, GPU never gets above 76c while playing Mod Warfare 2. I have an aftermarket heatsink on the CPU, so that stays cool.

    I took an online assessment thing on how much power my system needs approx. and it said about 661w. My power supply is 750w, but its an older, cheap power supply. I have a corsair 750w lying around that someone lent me so I guess I could try that.
  4. Also, I'm running Prime95 right now, but how long do I want it to go? I'm running a blend stress test at the moment and my temps are...

    GPU: 45C
    HD1: 37C
    HD2: 35C
    System: 61C
    CPU: 62C
    AUX: 52C
  5. Run Prime95 for at least an hour to achieve max temps. You initial temps aren't in the too hot range and are higher than you would see in Gaming.

    You can see your voltage, settings... etc for your RAM in your BIOS or download CPUID's CPU-Z which will provide you the same information.

    I would switch out to the Corsair PSU to see what happens, as well.
  6. Best answer
    Check out
    Image 14 shows DRAM voltage. Image 15 shows the RAM timings. Somewhere in there you'll find your base speed/bclk and CPU and RAM multipliers. They're more than likely running at 133, 20x and 10x, respectively, since you haven't changed anything.

    CPU-Z is a good way to find some info quickly.

    I feel like I should warn you now - if you do end up needing to increase the V to the RAM, the max the P55 can support is 1.65. Your RAM is rated 1.5V so I don't see any reason why you'd end up needing to get anywhere close to that. Just a word of caution.

    Hopefully it's the PSU. I second tecmo's recommendation. Put in the Corsair and see what happens.
  7. I finally fixed my problem. The motherboard by default was providing 1.59v to the DRAM, but when it was adjusted down to 1.5v it no longer crashed. Thanks for all of your time.
Ask a new question

Read More

Homebuilt RAM Systems Product