Overclock.net › Forums › Components › Memory › Memory won't pass IBT in dual channel mode
New Posts  All Forums:Forum Nav:

Memory won't pass IBT in dual channel mode

post #1 of 5
Thread Starter 
Alright, I'm a bit lost right now on how to solve this problem, so I hope I can get some help.

I'm having some troubles with my sig rig. I bought 8Gb of G.Skill Snipers 1600 (2x4GB), but I can't get them to work properly.

I've had to RMA them twice, since they would throw errors at Memtest withing the first pass. Additionally, they wouldn't accept any kind of tinkering with the speed/timings, leading the motherboard to give an "overclock error" on boot.


Now, this is the third set, and I'm actually able to change the settings without the motherboard throwing a fit. But they still don't work properly.

If I run Intel Burn Test (working on my OC again), it will give a warning error after some time running. I followed the guide here on how to run it properly, closed all unneeded programs, stopped services, etc. Then I popped my old memory (Mushkin, 2x2GB), and it worked (but it doesn't work at 1600, so I need to use it at 1333, even if it's rated for more). It ran 200 passes of IBT without a hitch.

Now, here are the things I've tried:

-Locked the cores again, to run is as dual core.
-Went back to stock settings.
-Gave the CPU/NB more volts. Right now, it's running stock speeds, 1.3v (which should be safe).
-Ran the memory at 1333-8-8-8-24, which according to G.Skill support should be a 100% compatible setting.
-Ran Memtest86+, but it didn't detect any errors in 4 passes.
-Ran HCI Memtest, detected errors.
-Tried upping the NB voltage, from 1.120 stock to 1.140. I read that it can help with memory unstability.
-Ran the memory at 1.65v.

Nothing of this worked.

Now, I tried running 1 module at a time. I ran them at 1333-8-8-8-24. And both modules passed 160 passes of IBT by themselves. This seems to indicate that the memory modules are working correctly, right?

Now I'm at a loss. Have I run into a motherboard incompatibility or something? Do I need to do anything else to make both modules work?

I've just updated the BIOS, although it doesn't mention anything about memory. I'll run it overnight with both modules, and see how it goes.

But is there anything else I can try? Any advice will be welcome.
Dustbox
(15 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 1045t Gigabyte GA-790XTA-UD4 Gigabyte HD 6850 @ 970/1100 G.Skill Sniper @ Working on it 
Hard DriveCoolingOSMonitor
2xSeagate 7200.12 500GB (RAID0) + 7200.12 1TB +... Prolimatech Megahalems + 2xScythe Gentle Typhoo... Windows 7 Professional x64 Dell U2311H 
PowerCaseAudio
Corsair HX850 Cooler Master HAF 922 Asus Xonar DX 
  hide details  
Reply
Dustbox
(15 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 1045t Gigabyte GA-790XTA-UD4 Gigabyte HD 6850 @ 970/1100 G.Skill Sniper @ Working on it 
Hard DriveCoolingOSMonitor
2xSeagate 7200.12 500GB (RAID0) + 7200.12 1TB +... Prolimatech Megahalems + 2xScythe Gentle Typhoo... Windows 7 Professional x64 Dell U2311H 
PowerCaseAudio
Corsair HX850 Cooler Master HAF 922 Asus Xonar DX 
  hide details  
Reply
post #2 of 5
Quote:
I've had to RMA them twice, since they would throw errors at Memtest withing the first pass. Additionally, they wouldn't accept any kind of tinkering with the speed/timings, leading the motherboard to give an "overclock error" on boot.
Now, this is the third set, and I'm actually able to change the settings without the motherboard throwing a fit. But they still don't work properly.

Sort of sounds like the board is having a hard time, especially after RMAing the memory that many times. Possible weak circuit traces or line ringing, do you have the option of adjusting the DIMM termination and clock delays with that board? Have you tried the other pair of DIMM slots?
Quote:
Then I popped my old memory (Mushkin, 2x2GB), and it worked (but it doesn't work at 1600, so I need to use it at 1333, even if it's rated for more).

Do these work properly in dual channel with the settings you want for the Gskill modules? Or have tested with the Gskill modules? (1333-8-8-8-24)
Quote:
Tried upping the NB voltage, from 1.120 stock to 1.140. I read that it can help with memory unstability.

You're safe to take this up to 1.35+ easily, just watch the heat on the chip. IMC contributes quite a bit to the heat output. I'd give 1.20-1.25 a go.
Quote:
Now, I tried running 1 module at a time. I ran them at 1333-8-8-8-24. And both modules passed 160 passes of IBT by themselves. This seems to indicate that the memory modules are working correctly, right?

This points to above, something in the board or IMC. Especially if your Mushkin sticks (unsure of which sticks, of course) also fail in dual channel with those settings- assuming they are rated (or overclocked) to do it. Yeah, I'd say the sticks work, but IBT isn't the do all end all, definitely try HCI Memtest as well- but after that many passes, sticks are probably not at fault.

As far as unrelated things to try, try bumping up the vCore an increment or two, it -can- help, especially with untested (or -testing-) modules, I had to do this when I got my Samsung sticks due to it failing about 8 hours in, needed 1 notch more vCore. Also, if you have DCT set to Ganged Mode, disable that. Unganged generally works better outside of benchmarking, and can help stability. What are your sticks RATED command rate? This can be an extremely sensative timing, as is tRFC. Verify the correct timings/delays, and all that stuff (probably already have).

Best wishes with getting the bugs sorted out! thumb.gif

EDIT: Forgot to mention, take it easy with the IBT. Running that many passes is absolutely terrible for your processor and regulation subsystems. IBT is more of a short and dirty CPU test, I wouldn't recommend it for long-run memory tests.
Edited by mezmenir - 9/15/12 at 7:23am
    
CPUMotherboardGraphicsGraphics
Intel Core i7-4790K ASRock Fatal1ty Z97X NVIDIA GeForce GTX 780 NVIDIA GeForce GTX 780 
RAMHard DriveHard DriveHard Drive
32GB ADATA XPG DDR3-2400 1x 256GB Plextor M6e M.2-PCIe 2x 256GB Samsung SSD 830 2x 2TB Western Digital Black 
OSMonitorKeyboardPower
Microsoft Windows 6.4.9841 CTP Qnix QX2710LED Logitech G710 EVGA Supernova 1000 G2 
CaseMouseAudio
Corsair Carbide 500R Logitech G3 Creative Zx, Sennheiser HD598 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
Intel Core i7-4790K ASRock Fatal1ty Z97X NVIDIA GeForce GTX 780 NVIDIA GeForce GTX 780 
RAMHard DriveHard DriveHard Drive
32GB ADATA XPG DDR3-2400 1x 256GB Plextor M6e M.2-PCIe 2x 256GB Samsung SSD 830 2x 2TB Western Digital Black 
OSMonitorKeyboardPower
Microsoft Windows 6.4.9841 CTP Qnix QX2710LED Logitech G710 EVGA Supernova 1000 G2 
CaseMouseAudio
Corsair Carbide 500R Logitech G3 Creative Zx, Sennheiser HD598 
  hide details  
Reply
post #3 of 5
Thread Starter 
Quote:
Originally Posted by mezmenir View Post

Sort of sounds like the board is having a hard time, especially after RMAing the memory that many times. Possible weak circuit traces or line ringing, do you have the option of adjusting the DIMM termination and clock delays with that board? Have you tried the other pair of DIMM slots?

For the termination voltage, my board doesn't allow you to tweak it if you set the DRAM voltage manually. It seems to be set at 0.75v by default, which should be correct for 1.5v DRAM voltage. As for clock delays, I guess I can.

The other pair of DIMM slots is unavailable right now due to the CPU cooler, the fan presses the modules. In the worst case I could run with just 1 fan, but I prefer to try other alternatives before.
Quote:
Do these work properly in dual channel with the settings you want for the Gskill modules? Or have tested with the Gskill modules? (1333-8-8-8-24)

The Mushkin modules were working in dual channel configuration, in the same slots that the Gskill modules. Their timings were different though, 7-10-10-24. They also ran at 1.65v instead of 1.5v.
Quote:
You're safe to take this up to 1.35+ easily, just watch the heat on the chip. IMC contributes quite a bit to the heat output. I'd give 1.20-1.25 a go.

Mind you, this is the chipset NB. Not CPU-NB. I read here that increasing NB voltage (again, not CPU-NB) could help with memory instability. CPU-NB is running at 1.3v, with stock speed of 2000Mhz.
Quote:
As far as unrelated things to try, try bumping up the vCore an increment or two, it -can- help, especially with untested (or -testing-) modules, I had to do this when I got my Samsung sticks due to it failing about 8 hours in, needed 1 notch more vCore.

vCore shouldn't be a problem. Right now, I'm running it at 1.4v, which is a couple notches more than it probably needs for 3.6Ghz (last time I tried, it could do it at 1.325v). Also, when I turned everything to stock, I tried running higher vCore and it still failed.
Quote:
Also, if you have DCT set to Ganged Mode, disable that. Unganged generally works better outside of benchmarking, and can help stability. What are your sticks RATED command rate? This can be an extremely sensative timing, as is tRFC. Verify the correct timings/delays, and all that stuff (probably already have).

DCT is set to unganged, I haven't changed it. Command Rate is 2T, and that's what I'm using. tRFC is set at 160ns I think (or whatever unit Gigabyte uses), which might be higher than needed, or perhaps not since it isn't working. As for the rest of the timings, I'm setting them as indicated by SPD, so I hope they are right.

Thanks a lot for the answer, and I'll be sure to stop using IBT for such long runs. HCI Memtest should work anyway to check the memory, I'm not really sure why I kept using IBT.
Dustbox
(15 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 1045t Gigabyte GA-790XTA-UD4 Gigabyte HD 6850 @ 970/1100 G.Skill Sniper @ Working on it 
Hard DriveCoolingOSMonitor
2xSeagate 7200.12 500GB (RAID0) + 7200.12 1TB +... Prolimatech Megahalems + 2xScythe Gentle Typhoo... Windows 7 Professional x64 Dell U2311H 
PowerCaseAudio
Corsair HX850 Cooler Master HAF 922 Asus Xonar DX 
  hide details  
Reply
Dustbox
(15 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 1045t Gigabyte GA-790XTA-UD4 Gigabyte HD 6850 @ 970/1100 G.Skill Sniper @ Working on it 
Hard DriveCoolingOSMonitor
2xSeagate 7200.12 500GB (RAID0) + 7200.12 1TB +... Prolimatech Megahalems + 2xScythe Gentle Typhoo... Windows 7 Professional x64 Dell U2311H 
PowerCaseAudio
Corsair HX850 Cooler Master HAF 922 Asus Xonar DX 
  hide details  
Reply
post #4 of 5
Or you can use Prime95 Blend Test with Rounding Error Check enabled.

"CPU-NB is running at 1.3v, with stock speed of 2000Mhz."
The speed is stock but CPU-NB voltage is not. Stock cpu-nb voltage is 1.15 or 1.2v

Regular NB voltage is for Reference Clock (def 200) overclocking stability, not for rams.
Edited by Catscratch - 9/15/12 at 9:32am
Intel Evilnow
(18 items)
 
   
CPUMotherboardGraphicsRAM
i5 2500k 4ghz @ Offset -0.015 Asus P8P67 Evo (bios 3207) Sapphire 280x Tri-x 3GB OC (Stock 1020/1500 Non... G.Skill RipjawsX 2x4gb 1866mhz 9-10-9-28-2n @ 1.5v 
Hard DriveHard DriveHard DriveHard Drive
SHSS37A120G WD5000AAKX-001CA0 WD20EARX WD20EZRZ 
Hard DriveOptical DriveCoolingOS
WD5001AALS-00L3B2 (Now External) ASUS DRW-1814BLT Noctua NH-u12p SE2 Windows 10 Pro 
MonitorKeyboardPowerCase
Asus VH242H Wobbly Stand :) Microsoft Ergo 4000 Enermax Infiniti 650 (28a,28a,30a) Cooler Master haf 912 Advanced 
MouseOther
A4tech x7 F3 Sunbeam RHK-EX-BA Rheobus-Extreme Fan Controlle... 
CPUMotherboardGraphicsRAM
Phenom II x6 1090t BE 3.6/4.0 Turbo@def.volt MSI K9A2 Platinum v1 Sapphire HD6850 1GB 850/1100@def.volt Kingston 2x2gb Hyperx 1066 5-5-5-15 
Hard DriveHard DriveOptical DriveOS
Western Digital WD5001AALS Seagate Barracuda ST3250410AS Asus DRW-1814BLT Windows 7 Ultimate x64 SP1 
MonitorKeyboardPowerCase
Asus VH242H 23.6" Wobbly Stand :D Microsoft Ergo 4000 Enermax Infiniti 650w (28a,28a,30a) Thermaltake Kandalf SuperTower 
Mouse
A4 tech Swop-3 
  hide details  
Reply
Intel Evilnow
(18 items)
 
   
CPUMotherboardGraphicsRAM
i5 2500k 4ghz @ Offset -0.015 Asus P8P67 Evo (bios 3207) Sapphire 280x Tri-x 3GB OC (Stock 1020/1500 Non... G.Skill RipjawsX 2x4gb 1866mhz 9-10-9-28-2n @ 1.5v 
Hard DriveHard DriveHard DriveHard Drive
SHSS37A120G WD5000AAKX-001CA0 WD20EARX WD20EZRZ 
Hard DriveOptical DriveCoolingOS
WD5001AALS-00L3B2 (Now External) ASUS DRW-1814BLT Noctua NH-u12p SE2 Windows 10 Pro 
MonitorKeyboardPowerCase
Asus VH242H Wobbly Stand :) Microsoft Ergo 4000 Enermax Infiniti 650 (28a,28a,30a) Cooler Master haf 912 Advanced 
MouseOther
A4tech x7 F3 Sunbeam RHK-EX-BA Rheobus-Extreme Fan Controlle... 
CPUMotherboardGraphicsRAM
Phenom II x6 1090t BE 3.6/4.0 Turbo@def.volt MSI K9A2 Platinum v1 Sapphire HD6850 1GB 850/1100@def.volt Kingston 2x2gb Hyperx 1066 5-5-5-15 
Hard DriveHard DriveOptical DriveOS
Western Digital WD5001AALS Seagate Barracuda ST3250410AS Asus DRW-1814BLT Windows 7 Ultimate x64 SP1 
MonitorKeyboardPowerCase
Asus VH242H 23.6" Wobbly Stand :D Microsoft Ergo 4000 Enermax Infiniti 650w (28a,28a,30a) Thermaltake Kandalf SuperTower 
Mouse
A4 tech Swop-3 
  hide details  
Reply
post #5 of 5
Quote:
For the termination voltage, my board doesn't allow you to tweak it if you set the DRAM voltage manually. It seems to be set at 0.75v by default, which should be correct for 1.5v DRAM voltage. As for clock delays, I guess I can.

Not the VTT Ref, sorry. I meant the On-Die Termination. It's a value measured in ohms and is usually listed along with the fine delay/skew settings. Higher may help as Gigabyte boards (and most for that matter), like to default to 60ohms. It's purely a signal integrity setting, so performance variance isn't really a matter when changing it.
Quote:
The Mushkin modules were working in dual channel configuration, in the same slots that the Gskill modules. Their timings were different though, 7-10-10-24. They also ran at 1.65v instead of 1.5v.

That is really bizarre. I mean, it may have something to do with the timings and your IMC, but that is still really odd. Out of curiousity, have you tried running the GSkill sticks with the exact same settings as the Mushkins that don't fail? St 1.65v, the Gskill sticks SHOULD be able to run those timings at 1333.
Quote:
Mind you, this is the chipset NB. Not CPU-NB. I read here that increasing NB voltage (again, not CPU-NB) could help with memory instability. CPU-NB is running at 1.3v, with stock speed of 2000Mhz.

Sorry about that, always jump to memory controller when I see NB in an AMD thread, hah. Not really sure what benefits it has, but I've always just run mine at 1.15v with the 900 series chipsets.
Quote:
vCore shouldn't be a problem. Right now, I'm running it at 1.4v, which is a couple notches more than it probably needs for 3.6Ghz (last time I tried, it could do it at 1.325v). Also, when I turned everything to stock, I tried running higher vCore and it still failed.

Guess you've got this covered then. Though, I am really leaning away from it being a board or IMC problem from your last post. It just seems rather odd, like the board just really doesn't play nicely with those DIMMs. Do you have another computer you can try them in?
Quote:
tRFC is set at 160ns

Should be plenty. At 200 HTT REF, most DIMMs work fine at 110ns. Some have higher ratings of course, but much of the time, you wont need to increase tRFC unless you are aiming for 300+ HTT REF.
Quote:
The other pair of DIMM slots is unavailable right now due to the CPU cooler, the fan presses the modules. In the worst case I could run with just 1 fan, but I prefer to try other alternatives before.

Would try another computer if you have one available, or a friend who'd let you test em. Otherwise, you could just un-heatspreader the sticks, they are only there for show. Personally, I'd rather keep the fan and pull of the spreaders, but that's just me.
    
CPUMotherboardGraphicsGraphics
Intel Core i7-4790K ASRock Fatal1ty Z97X NVIDIA GeForce GTX 780 NVIDIA GeForce GTX 780 
RAMHard DriveHard DriveHard Drive
32GB ADATA XPG DDR3-2400 1x 256GB Plextor M6e M.2-PCIe 2x 256GB Samsung SSD 830 2x 2TB Western Digital Black 
OSMonitorKeyboardPower
Microsoft Windows 6.4.9841 CTP Qnix QX2710LED Logitech G710 EVGA Supernova 1000 G2 
CaseMouseAudio
Corsair Carbide 500R Logitech G3 Creative Zx, Sennheiser HD598 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
Intel Core i7-4790K ASRock Fatal1ty Z97X NVIDIA GeForce GTX 780 NVIDIA GeForce GTX 780 
RAMHard DriveHard DriveHard Drive
32GB ADATA XPG DDR3-2400 1x 256GB Plextor M6e M.2-PCIe 2x 256GB Samsung SSD 830 2x 2TB Western Digital Black 
OSMonitorKeyboardPower
Microsoft Windows 6.4.9841 CTP Qnix QX2710LED Logitech G710 EVGA Supernova 1000 G2 
CaseMouseAudio
Corsair Carbide 500R Logitech G3 Creative Zx, Sennheiser HD598 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Memory
Overclock.net › Forums › Components › Memory › Memory won't pass IBT in dual channel mode