i-7 Hex Unstable

[GhosT15]

Hey Guys,

Got my i7-Hex but I cannot get it stable which is really killing me. I have an EVGA Board, evga x58 3x SLI Board. Decided to abandon Asus as my other board has the unlocked multiplier problem (wont stick). Seems like this was a mistake as I cannot get the evga board stable.

Here's the problem; The Stock or OC'ed the computer will randomly reboot. Sometimes after 10-15 minutes sometimes after hours.

I've done MemTest64+ for 2 passes, little over 1.5 hours and both passes were clean (no errors).

Full Hardware:
1x SSD
1x Raptor HDD
2x Green HDDs
1x GEForce 280 OC
1x Sound Card (x-fi gamer)
1x i7-980x @ 143x29 @ 1.3vcore (4136.73MHz)
1x OCZ Repear 2000MHz 3x2GB Memory (@ 1996MHz)
1x EVGA x58 3x SLI (132-BL-E758-A1)
1x BFG ES-800 , 800 Watt PSU

My CPU Temps are all fine, peaking at 43c Stock, 55c OC'ed (I haven’t really finished pushing the OC, but I wanted to see if using manual voltage and clocks would help the issue)

Called EVGA they think it's a Power Issue, this PSU was working fine with my old x58 board, and my old i7 (with the same video card and rest of the hardware).
Using Kill-a-watt my old setup on this PSU pulled 485w from the wall
Using Kill-a-watt my new setup on this PSU pulled 480w from the wall

No real change in pull so no excuse that a 800 watt PSU cannot handle 480w under full load, or roughly ~400w of real draw (assuming 80%ish efficiency which it’s rated for)

Anyone have any ideas while I try to get this system folding. While it's going on OC i'm getting almost 22k PPD on regular SMP w/ ~8k PPD from the GPU at the same time!

Just need to get this thing stable now. P.S. - This happened at stock speed and voltage, the OC isn't a factor. Fresh Install of Windows, No Errors / Blue screens

 

[DooKey]

Reboots are almost always PSU related in my experience. If you are getting reboots at stock and OC then I am leaning PSU myself.

 

[Tobit]

I agree with DooKey

 

[Jon855]

What PSU do you have? And as others said - it may be the PSU itself. RMA it or get a better one.

Cannot say without knowing what exactly you have there.

 

[LOCO LAPTOP]

What PSU do you have? And as others said - it may be the PSU itself. RMA it or get a better one.

Cannot say without knowing what exactly you have there.

he said he has a BFG ES-800, cant RMA he is shit out of luck.

if that the problem atleast.

 

[musky]

Do you have the same issue without folding on the video card? You will be surprised at how little the video card will gain you with the hex core. The machine will probably only pull 310W without the video card folding, which gives you a lot more headroom for your PSU.

 

[GhosT15]

Do you have the same issue without folding on the video card? You will be surprised at how little the video card will gain you with the hex core. The machine will probably only pull 310W without the video card folding, which gives you a lot more headroom for your PSU.

Yep, same problem without folding.


I know the PSU seems likely, but this same PSU was running a x58 Asus MB the EE i-7 980 i think it was, and the same video card flawlessly. Why would switching out MB, CPU, and RAM suddenly cause the PSU to fail?

Just did a full reset on the bios; pulled battery, psu and memory, waited 15 mins. Only installed 1 stick of ram, and so far *knock on wood* been up for almost an hour. Wonder if the flash and normal BIOS reset didn't quite clear right.

 

[roguespearus]

GhosT15,

Evga recently pushed out a new bios; usually 74 for most boards, that is supposed to make overclocking HEX cpus more stable. I have had strange operation for my 930 which I can not overclock at all with Eleet but my bios overclocks are rock solid on my E762.

I also have the X-fi gamer card and creative has released a driver in the last month which gave me more consistent frame rates in games.

What cooling are you using? Some aftermarket coolers cause an electrical short due to contacting the electrical connections on the back of the motherboard. I had this problem a couple years ago and it caused the same random reboots you are having. I do not remember the exact cooler but I think it was a Zalman 9500.

Is the motherbaord in a case? Some power supplys need to have the grounding that the motherboard get when mounted in the case.

 

[GhosT15]

GhosT15,

Evga recently pushed out a new bios; usually 74 for most boards, that is supposed to make overclocking HEX cpus more stable. I have had strange operation for my 930 which I can not overclock at all with Eleet but my bios overclocks are rock solid on my E762.

I also have the X-fi gamer card and creative has released a driver in the last month which gave me more consistent frame rates in games.

What cooling are you using? Some aftermarket coolers cause an electrical short due to contacting the electrical connections on the back of the motherboard. I had this problem a couple years ago and it caused the same random reboots you are having. I do not remember the exact cooler but I think it was a Zalman 9500.

Is the motherbaord in a case? Some power supplys need to have the grounding that the motherboard get when mounted in the case.

Roguespearus, great points.

Since this was a new install I did do a bios flash to the latest version (74). The OC Is stable from what I can tell, other than this reboot issue.

on the X-Fi, since this is a fresh install of windows I did pick up these newer drivers for the install.

The cooling is a custom water loop w/ 4x120mm radiator. It uses the Koolance CPU block, it is metal but has a silicon(ish?) layer between the backend mount and the motherboard. Although possible, it makes a short unlikely.


------------------

For an update, the system stayed up overnight downloading SC2 and doing SMP + GPU Folding. I've also been able to write this whole update on it. Only thing is I only have 1 of 3 sticks of memory. Memory passed MemTest but who knows. If it makes it to after work I'll put in the other two sticks and see what happens

 

[lehmann]

Ghost from what i have seen. you might wanna try 1.35 vcore. i have had this happen while trying out new overclocks.

 

[Wolfie]

I know the PSU seems likely, but this same PSU was running a x58 Asus MB the EE i-7 980 i think it was, and the same video card flawlessly. Why would switching out MB, CPU, and RAM suddenly cause the PSU to fail?

Some motherboard are more sensitive to voltage fluctuations, (particular parts such as CPU core voltage, )
It may or may not be noticeable but the extra load on the 12V rail might be causing cross talk noise or a voltage ripple on the12v line, thus causing instability and power losses.

Have you turned off the "automatically Reboot on Bluescreen" in windows, so you can see what the reboot error is?

What about running memtest overnight? (I had a P!!! with an intermittent problem, but it would run for about 6 hours without error then start dropping random memory locations, I backed the memory timing to a less aggressive setting and it was fine)

 

[10e]

If things are working well, try adding a stick at a time.

Until that point you want to run the memory at 1333mhz DDR3 if you are not touching VTT to ensure your memory is all good, and your first three DIMM slots are all good. If you are trying to run memory at 2000mhz with default VTT of 1.15, one stick may work, but as you add more sticks, your IMC on the chip will start to freak out. I haven't used one Core I7 yet that could run at higher than 1400mhz DDR3 with dual or triple channel without some form of VTT rise.

Also IntelBurnTest and LinX are way better than Memtest to help test CPU and memory issues. It will also help you see what max temps will be on your water loop.

 

[amdgamer]

Thanks for the heads up. My Auzentech uses those X-FI drivers so i'll have to check and see if Auzentech has released new drivers yet for my X-Fi Forte.

GhosT15,

Evga recently pushed out a new bios; usually 74 for most boards, that is supposed to make overclocking HEX cpus more stable. I have had strange operation for my 930 which I can not overclock at all with Eleet but my bios overclocks are rock solid on my E762.

I also have the X-fi gamer card and creative has released a driver in the last month which gave me more consistent frame rates in games.

What cooling are you using? Some aftermarket coolers cause an electrical short due to contacting the electrical connections on the back of the motherboard. I had this problem a couple years ago and it caused the same random reboots you are having. I do not remember the exact cooler but I think it was a Zalman 9500.

Is the motherbaord in a case? Some power supplys need to have the grounding that the motherboard get when mounted in the case.

 

[GhosT15]

First, Thanks everyone for the ideas, suggestions, etc. I greatly appreciate it even if I've already tried it and didn't mention it.

I'll try a bunch of these newest ones when I get home from work tonight.

Ghost from what i have seen. you might wanna try 1.35 vcore. i have had this happen while trying out new overclocks.

This was happening at Stock Speeds, Auto and 1.25 vcore. I think Its OC'ed to 4.13Ghz @ 1.325 vcore atm. Prime95 runs fine w/o errors

Some motherboard are more sensitive to voltage fluctuations, (particular parts such as CPU core voltage, )
It may or may not be noticeable but the extra load on the 12V rail might be causing cross talk noise or a voltage ripple on the12v line, thus causing instability and power losses.

Have you turned off the "automatically Reboot on Bluescreen" in windows, so you can see what the reboot error is?

What about running memtest overnight? (I had a P!!! with an intermittent problem, but it would run for about 6 hours without error then start dropping random memory locations, I backed the memory timing to a less aggressive setting and it was fine)

is there a greater load on the 12v rail? Ironically kill-a-watt shows a slightly lower power draw! I will try and watch the 12v rail and see if I notice anything wacky before a failure.

There is no blue screen, and no entry in the Event Viewer (other than unsafe shutdown power loss from the kernel) I did however turn off the Auto Reboot option just to be sure!

I have not tried running it overnight with MemTest, Worth a shot, although how will I know if it failed if it happens to reboot overnight? Webcam time? lol.

If things are working well, try adding a stick at a time.

Until that point you want to run the memory at 1333mhz DDR3 if you are not touching VTT to ensure your memory is all good, and your first three DIMM slots are all good. If you are trying to run memory at 2000mhz with default VTT of 1.15, one stick may work, but as you add more sticks, your IMC on the chip will start to freak out. I haven't used one Core I7 yet that could run at higher than 1400mhz DDR3 with dual or triple channel without some form of VTT rise.

Also IntelBurnTest and LinX are way better than Memtest to help test CPU and memory issues. It will also help you see what max temps will be on your water loop.

Interesting points about the VTT, I have the QPI/VTT at 1.30 currently (Even with the one stick) I found during my other overclocking 1.30-1.35v worked best with the high OC, although the memory was only OC'ed to 1600. You might be onto something, and its worth looking into further if the system is still up when I get home on the one stick of memory.

I will definitely take a look at these Apps, if anything it might prove to be another avenue to help verify the new hardware, thanks!

 

[GhosT15]

Thanks to everyone for all the advice, I've been running all 3 sticks of memory for over an hour, with F@H going on both CPU and GPU. I'm pretty happy, and hope this stays up. I'm stable per the Intel Burn Test as well with the OC.

Final Settings:
vcore: 1.3625
vDimm: 1.65
QPI PLL: 1.35
(Defaults) IOH, CPU VTT, ICH

Bclck: 143
Mult: 29
Frequency: 4.136GHz
Memory: 2006MHz @ 9-9-9-30-1T (6GB)
CPU: i7-980X

Getting some pretty crappy TPF on a 6702 ATM, 5:30 , PPD 17,328


Real Temp Shot:

That is after running Intel Burn in Test, and it has now been folding for about 45 Minutes.

 

[rflcptr]

Off topic, but what are you cooling the CPU with?

 

[GhosT15]

Off topic, but what are you cooling the CPU with?

Custom built watercooling loop. 4x120MM Koolance Radiator with the older Koolance i7 CPU Cooler (first generation). Have the fans and pump at 7 out of 10 for the setting, crank it to 10/10 in the summer to keep everything nice and cool. I will take some pictures this weekend and post them in the hardware picture thread.

 

[Red Falcon]

Called EVGA they think it's a Power Issue, this PSU was working fine with my old x58 board, and my old i7 (with the same video card and rest of the hardware).
Using Kill-a-watt my old setup on this PSU pulled 485w from the wall
Using Kill-a-watt my new setup on this PSU pulled 480w from the wall

No real change in pull so no excuse that a 800 watt PSU cannot handle 480w under full load, or roughly ~400w of real draw (assuming 80%ish efficiency which it’s rated for)

Well, even though your only pulling 480 watts from your PSU, it could be needing more watts/amps from one of the voltage rails, be it 3.3v 5v or 12v.

Example: if your PSU can only handle a 20amp pull from the 5v rail, and your setup is trying to pull 28amps from it, that can cause a stability issue.

I'm not saying this is what the problem is, but for a PSU that beefy to be having problems, it's either not giving enough amps on a specific rail, or it is overheating and/or dying.

Best of luck to you.

 

[GhosT15]

Update for everyone...

Once I go 3 Sticks of memory the system reboots. This last time It lasted almost 24 Hours before rebooting. I Swapped Stick 2 with Stick 3, and I'm running 2 sticks again. If it happens again then I'm guessing I have a bad stick, otherwise Motherboard? PSU?

I have another exact same PSU in a different system, BFG ES-800. Is it worth all the pain of swapping the PSU's out to see if the one in this system is dying?

If it is not enough Watt/Amp , any suggestions on what Power Supply to pick up? (Especially Brand & Watt)


Thanks again for all the advice/help so far! I still have it crunching some SMP WU's as best as it can, as long as it doesn't reboot while I'm gone it keeps going fine.

 

[lehmann]

Corsair makes some good psu's the Tx's are good the HX AX's are great