Sli or possible mobo problem, please help!

Intel Core i5 750
Asus p7p55d-e pro(BIOS version 1602, the most recent one)
Two EVGA 570 DS in Sli
16 Gb DDR3 PC 1600
900w psu

so heres my problem, this motherboard should run my GPU's at x8 x8, and it does however it only does every so often, most of the time it runs the cards at x4 x8, ive flashed my bios to the most recent version and this hasnt seemed to fix it, ive reseated the cards messed with sli settings, im lost i cant seem to figure out why it runs at x4 x8 and on random restarts will run at x8 x8. Anyone have ideas?

Heres Gpuz screens:

32 answers Last reply
More about mobo problem help
  1. Are there any things in your regular PCI slots?

    It sounds like another thing wants to use the bandwidth that the card is trying to use.
  2. Nope just the Graphic cards
  3. Do you have anything that uses USB 3.0 plugged in?
  4. had to check on that one i use usb3.0 for my external but its not plugged in atm.
  5. Try giving me a maker and model for everything you have attached, maybe that will help.

    That includes the PSU which wasn't specified earlier.

    Also, a picture of the inside of the case that works would be nice. The other ones you posted don't show anything on my computer. Not sure where they are hosted.
  6. PSU: Apevia warlock 900w
    Ram: 16 GB DDR3 Gskill Sniper pc 1600
    Steelseries 5H V2 usb soundcard

    The images i attached were screenshots of GPU-z ill upload them on a different site,

    Heres pictures of inside my case:

    i recently broke my nice 22' monitor so im on a old 19' vga :(
  7. Anything else you have attached that isn't listed?

    Anything at all?

    Is there anything USB that you can live without other than the Steelseries 5h V2 for the purposes of these tests?

    Get rid of that and anything else that isn't mouse/keyboard and see if it is still x8/x4.
  8. nope thats it
  9. unplugged the sound-card then restarted still at x8/x4
  10. Ok, so lets try this angle.

    Some of the time when you try it the output says x8/x8 and sometimes x8/x4, what is happening different between the times that this is happening?

    Is there something you do before testing that could have some effect?
  11. possibly entering the bios, ive been messing around with my overclocking a bit lately. ill give it a try.
  12. Try to down clock everything back to stock settings and see if that changes anything.

    Do you also OC your video cards? If so, set those back to normal too.
  13. i also just uninstalled vmware i thought the hardware virtualization might have something to do with it. its at x8/x8 right now im going to restart again and not enter the bios this time.
  14. so im at x8/x8 again, im going to leave it up for awhile and restart later to see if it goes back to x8/x4.
  15. OK, I will be around. I am travelling in Europe atm so its almost midnight now. It may be a while before I respond again, but I will be around within 24 hrs.
  16. Cool not to get too off topic but where are you going/have you been? Ive been to quite a few places in Europe myself.
  17. I am from the U.S. but my wife is German. I am in Germany for the entirety of this trip, but I have been to surrounding countries. I had my honeymoon in the Czech Republic, I have been to the Netherlands, etc.
  18. well happy travels, Ive been in Belgium, France, Italy, UK, and Amsterdam.
  19. Thanks.
  20. after a couple of restarts the problems still here :/
  21. I am sorry to hear it.

    Is there any chance you can borrow another rock solid PSU from somebody?

    800w+ made by Antec, Corsair, Seasonic, XFX, or anything with a Seasonic or perhaps Channel Well, Delta, or Enermax OEM?
  22. unfortunately no, you think its a psu issue? because i had everything running great a couple of months ago and i was playing BF3 and my PSU blew, it was the same model, luckly i had a 650w XFX psu so i could at least still have a functioning rig even though it wouldnt run sli, my 900w just got back from RMA and now im having these issues.
  23. OK so let me know if I got this right.

    You had the Apevia 900w.

    You used it with your SLId video cards.

    It blew up from the strain, so you RMAd it.

    You took out the SLI and left in 1 card and an XFX 650w PSU.

    It worked fine during the RMA process.

    You got another Apevia 900w.

    You put the new Apevia 900w in the PC and brought the SLI configuration back online.

    Now you have an identical situation as when the first PSU blew up, except that now you are having some different problem.

    If all of that is true, it does cause me to look at the PSU with a suspicious eye.

    I went ahead and read this nice little Apevia 900w Warlock review here:

    which gave the PSU a 5.5 / 10 rating and said it should have been listed as an 800w instead of 900w.

    It also said that the PSU ran out of spec many times on different lines when heavy amounts of load was put on it.

    I am guessing your SLI setup is capable of similarly heavy amounts of draw with 2x 570s.

    I don't know, I am kinda biased against PSUs that get serious ratings of 5.5 / 10 and the thing about 1x working with a 650w and 2x not working with a 900w right, it kinda feels like a PSU thing to me right now.

    The video card can pull 75w from the board, another 75w from a 6 pin pcie connector, and 100w from an 8 pin connector (25 per 2 wires on both types) so that is up to 250w that the second card can potentially max draw.

    With the XFX 650w another +250 could potentially reach to 900w, and this PSU was said in the review to maybe be an 800w, again I can see some potential problems there.

    The thing that really sticks a thorn in my side right now is the fact that these problems exist without heavy loads, as I understand it and why it might work sometimes and not others.

    If it were indeed a PSU problem, I would expect weirdness at load rather than at idle.

    That being said, PSU problems can surface as problems with nearly any other part.

    The exact same setup caused the other PSU to literally blow up, that tells me that the strain this thing must be feeling under load must be unimaginable.

    I don't know, it kinda feels like the PSU isn't giving the right amount of juice to the cards and therefore they can't run optimally.

    Anyway, that is just a stab in the dark with some reasonable grounds.

    Old PSU blew up with the same conditions.
    The rest of the PC worked just fine with no SLI.
    New PSU having problems with the conditions that blew up the same sort of old PSU.
    Review that calls it an 800w, fairly.

    I think that is reasonable grounds to consider another PSU at least. Something from Enermax or Seasonic in the 950w - 1000w range approx.
  24. that sucks alot it cost me $130 from newegg and had 4 stars, unfortunately i bought it months ago so i cant return it. and i dont have the money to get another psu right now.
  25. I am sorry to hear it.

    Might you know someone with a similarly uber system whose PSU you could at least borrow for a few hours for testing purposes?
  26. only at the school i went to, but its winter break and that would involve me driving up there with computer in tow, i just emailed apevia and asked for a refund (lol) they told me they couldnt do it but are willing to rma again i might give that a try seeing as the first psu worked until it blew.
  27. so this morning i started up and i was running at x1/x8..., so i took out that power supply and decided i was going to test each card individually in each slot, with my XFX 650W and now i dont think its the power supply because both cards ran at x8 in each slot when they should have run at x16 in each slot, i think its my mobo.
  28. They are both within the realm of possibility.

    I don't think anyone will fight with you overly much if you want to try replacing the motherboard first.

    I generally try to get people to exhaust all the possible options that don't require paying money first, but if you aren't worried about that then I guess I won't be.
  29. well the mobo is still under warranty so hopefully it will be free.
  30. Here is hoping.
  31. well i have another development, when you told me to reset my overclocking i did but i did by just entering the bios and clearing what i had set, i might have missed something because later i used the jumper on the board and cleared it and the cards functioned at the correct speeds. so with this found i went back through and re entered some overclock settings, with the default bios settings my ram only runs at 1333mhz not its 1600mhz all i did was increase the BCLK Frequency to allow the ram to run at the correct speed i increased it from 130 to 160 and then set the dram frequency to 1600mhz. by doing this my next restart went fine but after it was back to x4/x8, this still is confusing being that before my overclocking ran fine causing no problems, and if i cant even run my ram at the correct speed i feel like something is wrong. got any ideas?
  32. I suppose it is possible that your OCing did cause the change to your video card speeds by default.

    A lot of things I do know about the background workings and a lot of things I don't. In this particular case all I can really help do is isolate what the cause of problems are and not so much explain why they are problems.

    Anyway, It sounds like the x8/x8 is just not detecting quickly enough or something.

    Anyway, about the RAM...

    There are a lot of differences between generic RAM and non-generic RAM.

    There are only 4 companies that can make those little black chips on RAM sticks. Those are Micron(Crucial), Samsung, Hitachi, and Hyundai(Hynix). Each one of those makes these chips and runs a series of tests on them. They test them in many conditions and often ones that are more hash than the conditions they expect the chips to be used in.

    Many of their chips fail those tests due to some flaw or another and those go into a fail bin. The chips in the fail bins get sold to other companies for pennies on the dollar. The other chips that don't fail get branded by the maker and used for their own purposes. The maker requires these chips from the fail bin be sold under the buyer's own brand.

    Thus you can be sure a Micron stick has past every test with flying colors in a wide variety of tests and you can be sure that a generic stick of RAM failed at least one test if not more.

    When the companies buy the fail chips, they aren't incentivized to try to fail these chips themselves, so their testing is usually minimal. Almost all of them run only a few important test types and omit most of the others. The ones within spec on the small number of tests get slapped in a circuit board and sold.

    The generic makers know that a lot of these items will be RMAd because the chips just aren't good. They know a large percentage of this stuff will come back, but it is still effective to run their business this way, because the returns don't cost them all of their profits.

    In any event, I am pretty sure GSkill isn't the name that Samsung and Hitachi use to market their RAM directly to consumers which likely means that GSkill pulls its chips out of the fail bin.

    Anyway, long story shorter, a lot of chip makers test a RAM stick for like 10 seconds at 1600 for instance and if it works then they will call it 1600 RAM. If they had tried it for 100 or 1000 seconds the stick may have revealed a flaw and crashed and they couldn't label it as 1600. Maybe the stick is only stable at 1333, but the test the generic RAM company does doesn't reveal this.

    My wife refuses to spend more than the minimum on her computer. She is using generic RAM herself. It says 800 on the package, but it only runs at 667 mhz. It is probably for the same reason too.

    The more I learn about computer hardware the more I learn not to trust the labels for things because of reasons like these.
Ask a new question

Read More

Homebuilt Systems