Odd Issue with X58 System

[Devastator0]

Oh great and wise TPU community, I come to you asking for help. Heh, so, ok, my problem is that I have an older X58 system containing the following:

i7 920 (cooled with a Corsair H100i V2)
ASUS P6T V2 Deluxe (running 1202 BIOS)
12GB G.Skill F3-12800CL9T-6GBNQ
ASUS Strix GTX 980
Samsung 840 Pro SSD

Is that for a long time, the system has been working perfectly, then recently I was given an older Intel Xeon X5690 and ever since I installed it, the rig has been having strange issues. From what I can see, randomly at POST, it’s not detecting all of my RAM, but when it does and I get into Windows, it will just randomly blue screen.

So, I’m thinking the issue may be caused by the new CPU, I put my old 920 back in, and the same problem appears to be occurring. When I have all my RAM sticks in (6x2GB), the machine will sometimes detect 12GB & other times, it’ll only detect around 8GB. I’ve also pulled the BIOS battery, left the machine off AC for 24 hours, put a fresh battery in and powered on but problem still persists.

I’m currently running Memtest on 1 stick at a time to see if I can pin down a dodgy stick but I was wondering if anyone could recommend I do something or if they have seen a similar problem before, to point me in the right direction.

I would welcome and very much appreciate any feedback and suggestions (I will apologise in advance if I am delayed in putting updates or responding to your comments, I’m in Australia so there’ll be a time difference)

Cheers!

 

[95Viper]

Check your cpu socket for any problems (bent pins, dust, etc.).
And, make sure the cooler is installed correctly... not lob-sided or to tight, just right.
Remove ram and clean out sockets with canned air. Clean up connector on the ram modules reseat. Test.

 

[Ruru]

Yeah, pins came into my mind also as the possible problem. It's not uncommon to have issues with RAM slots and PCI-E slots* when there's problems with bent/shitty pins on Intel platform.

*X58 has PCI-E on its chipset so it doesn't have those problems, but later sockets do have.

 

[Bill_Bright]

According to the CPU QVL (qualified vendors list) for your motherboard, that board does not support the X5960. So it should never have been tried in it. I have never seen it personally, but I have heard claims from other trusted techs that using an unsupported processor that required different voltages, even slightly different as seen with these two processors, resulted in permanent damage to the motherboard bus/chipset due to excessive current demands through those devices. And I have seen personally where clients fried unsupported CPUs due to those processors being exposed to unsupported voltages. I am not saying that is what happened here, I am just saying it is important to check motherboard QVLs and only use listed CPUs. Just because the socket is the same, that in no way means the CPU is supported.

Since the RAM was working before, I don't see how that would be it. Same with the PCI-e sockets. But it would not hurt to clean them - if done properly. An old tech's trick is to gently use a "clean" pencil erasure on the contacts of the sticks, then a quick blast of compressed air to remove any erasure crumbs. Be sure to observe careful ESD precautions and NEVER EVER touch the electrical contacts of the RAM sticks with your fingers. Always handle by the PCB edges. You can use a acid brush and short squirt with some quality electrical contact cleaner. A crisp, new and "clean" folded dollar bill run through the RAM and PCIe slots after a quick squirt with contact cleaner works great for cleaning too - in place of the brush.

Since you pulled the CMOS battery and left the computer unplugged for 24 hours (15 - 30 seconds would have been enough), that took care of resetting the BIOS, which would have been my suggestion after checking for pin damage. I am concerned that didn't work as that suggests something more permanent to me.

So that leaves ESD. Did you take the essential ESD precautions before removing and/or replacing the CPUs? That is, was the computer unplugged from the wall and did you touch bare metal of the case interior BEFORE reaching in and BEFORE touching the processors, and frequently in between?

The problem with ESD (electrostatic discharges) is ESD sensitive devices like processors and memory modules and other high density ICs (like the chipset!) is they can be destroyed by a discharge from your finger tip that is so tiny you (as a human being) are unable to see, hear or feel that a discharge even occurred. Yet that discharge (arc, spark, microscopic lightning bolt ) has enough potential (voltage) to torch a Grand Canyon size trench (microscopically speaking) through 1000s and even millions of transistor gates in those devices. And note you can easily build up such destructive static electricity potentials in your body just sitting in your chair, squirming in your clothes.

Sadly, you are now in a Catch-22 situation. You need to try those processors in compatible motherboards to see if they still work. But you risk damaging those motherboards if those processors have been damaged. And you risk damaging another CPU if you try another CPU in this motherboard.

I’m currently running Memtest on 1 stick at a time to see if I can pin down a dodgy stick but I was wondering if anyone could recommend I do something or if they have seen a similar problem before, to point me in the right direction.

I would probably test them in pairs, one pair at a time as that will test them in dual-channel mode.

Make sure it is this MemTest86. There are several memory testers out there with very similar names that came from the same original source code. But MemTest86 from PassMark is the only one under constant development to ensure support for the latest RAM and motherboard technologies. Run it through several passes or even overnight. If you get even one error, the RAM is bad.

That said, NO software based memory tester is 100% conclusive. As noted, if they report any errors, even one, the RAM is bad. But it is not uncommon for them to report no problems, yet the RAM still fails in use, and/or when paired with other RAM. So, swapping in all new RAM is often the best test - but of course, not everyone has a bunch of spare RAM laying around. So you might try running with just a single RAM stick (or pair of sticks) in normal Windows mode to see if it fails. Repeat process with remaining modules, hopefully identifying the bad stick through a process of elimination. Just be sure to unplug the computer from the wall and touch bare metal of the case interior BEFORE reaching for the RAM to discharge any destructive static in your body.

 

[jsfitz54]

OP no mention fresh Windows install.
No mention memory set to 1333. Starting point.
Old overclock no worky worky the same with Xeon chip.
All ram slots populated, no mention of voltage bump.
No mention if X5690 came from server pull or it was overclocked to death in private hands.

Missing LOTS of pertinent info.

 

[Regeneration]

Missing RAM on X58: Reinstall the CPU on the socket and make sure all pins aren't bent.

 

[steen]

I echo the previous suggestions: P6T Deluxe V2 should support WSM, check socket pins/cpu pads, reset BIOS to defaults, set RAM to 1066/1333 for 6x modules. Also try 1x DDR3 module per channel @ lower speed & increase if stable.

 

[Durvelle27]

According to the CPU QVL (qualified vendors list) for your motherboard, that board does not support the X5960. So it should never have been tried in it. I have never seen it personally, but I have heard claims from other trusted techs that using an unsupported processor that required different voltages, even slightly different as seen with these two processors, resulted in permanent damage to the motherboard bus/chipset due to excessive current demands through those devices. And I have seen personally where clients fried unsupported CPUs due to those processors being exposed to unsupported voltages. I am not saying that is what happened here, I am just saying it is important to check motherboard QVLs and only use listed CPUs. Just because the socket is the same, that in no way means the CPU is supported.

Since the RAM was working before, I don't see how that would be it. Same with the PCI-e sockets. But it would not hurt to clean them - if done properly. An old tech's trick is to gently use a "clean" pencil erasure on the contacts of the sticks, then a quick blast of compressed air to remove any erasure crumbs. Be sure to observe careful ESD precautions and NEVER EVER touch the electrical contacts of the RAM sticks with your fingers. Always handle by the PCB edges. You can use a acid brush and short squirt with some quality electrical contact cleaner. A crisp, new and "clean" folded dollar bill run through the RAM and PCIe slots after a quick squirt with contact cleaner works great for cleaning too - in place of the brush.

Since you pulled the CMOS battery and left the computer unplugged for 24 hours (15 - 30 seconds would have been enough), that took care of resetting the BIOS, which would have been my suggestion after checking for pin damage. I am concerned that didn't work as that suggests something more permanent to me.

So that leaves ESD. Did you take the essential ESD precautions before removing and/or replacing the CPUs? That is, was the computer unplugged from the wall and did you touch bare metal of the case interior BEFORE reaching in and BEFORE touching the processors, and frequently in between?

The problem with ESD (electrostatic discharges) is ESD sensitive devices like processors and memory modules and other high density ICs (like the chipset!) is they can be destroyed by a discharge from your finger tip that is so tiny you (as a human being) are unable to see, hear or feel that a discharge even occurred. Yet that discharge (arc, spark, microscopic lightning bolt ) has enough potential (voltage) to torch a Grand Canyon size trench (microscopically speaking) through 1000s and even millions of transistor gates in those devices. And note you can easily build up such destructive static electricity potentials in your body just sitting in your chair, squirming in your clothes.

Sadly, you are now in a Catch-22 situation. You need to try those processors in compatible motherboards to see if they still work. But you risk damaging those motherboards if those processors have been damaged. And you risk damaging another CPU if you try another CPU in this motherboard.
I would probably test them in pairs, one pair at a time as that will test them in dual-channel mode.

Make sure it is this MemTest86. There are several memory testers out there with very similar names that came from the same original source code. But MemTest86 from PassMark is the only one under constant development to ensure support for the latest RAM and motherboard technologies. Run it through several passes or even overnight. If you get even one error, the RAM is bad.

That said, NO software based memory tester is 100% conclusive. As noted, if they report any errors, even one, the RAM is bad. But it is not uncommon for them to report no problems, yet the RAM still fails in use, and/or when paired with other RAM. So, swapping in all new RAM is often the best test - but of course, not everyone has a bunch of spare RAM laying around. So you might try running with just a single RAM stick (or pair of sticks) in normal Windows mode to see if it fails. Repeat process with remaining modules, hopefully identifying the bad stick through a process of elimination. Just be sure to unplug the computer from the wall and touch bare metal of the case interior BEFORE reaching for the RAM to discharge any destructive static in your body.

A lot of boards didn’t list support for Xeon Processors but they still worked

I used a Xeon X5650 @4.6GHz with a ASUS P6T Deluxe with no issues after a BIOs mod to support more voltage

CPU support list isn’t always instensive

 

[agent_x007]

1) 6x2GB RAM will require more QPI/DRAM Voltage (1,3V should do it).
If missing memory keep insisting, try using lower DRAM Frequency and/or manual timings (including tRFC), and voltage.
2) ALWAYS Update BIOS to latest version for best CPU/RAM support.
3) You don't have to reinstall windows because you switched CPUs or installed more RAM.

 

[silkstone]

Have you tried cleaning the sticks and slots with alcohol?
I get pretty much the same problem with my X58 HTPC and the only way I ever get it working is by re-cleaning all the contacts, though it doesn't always work.

 

[dorsetknob]

According to the CPU QVL (qualified vendors list) for your motherboard, that board does not support the X5960. So it should never have been tried in it.

I have just Built a ASUS P6T Deluxe with a X5690 and it works Flawlessly (and water Cooled)
almost all retail Vendor Manafactures do not list Xeon Compatability yet they work................>> and that also goes for the workstation CPU's

You go thru the QVL list for Ram and ECC is not supported ( i got the manual open in front of me page 2-13 section 2.4.2)
Yet if you install a Xeon you can use DDR3 ECC (yeh bill we know why but tell us anyway )

 

[Bill_Bright]

A lot of boards didn’t list support for Xeon Processors but they still worked

Not the point. The point of the QVLs for both RAM and CPUs is to list devices that have been tested and verified as compatible. Just because something works, does not mean it is 100% compatible.

But to your point, it doesn't make a point. That CPU is not listed and it does not work on that board. We just don't know why. The fact dorsetknob has the same board and CPU and it works is, however, a good sign. Same BIOS version?

(yeh bill we know why)

Yeah, but RAM is a bit different. There are just too many RAM makers and models for motherboard makers to test them all. So you don't have to buy listed RAM. But if you want to be assured of compatibility, you should still buy RAM with the same specs as listed RAM.

 

[Devastator0]

Check your cpu socket for any problems (bent pins, dust, etc.).
And, make sure the cooler is installed correctly... not lob-sided or to tight, just right.
Remove ram and clean out sockets with canned air. Clean up connector on the ram modules reseat. Test.

Yeah, when I removed the Xeon CPU and put my original i7 back in, I looked at the pins on the mobo thoroughly and they all looked good. When I pull it out again today, I'll take a couple photos, post them up and see if anyone can point out something I've missed. I'll also def do the RAM contacts and try to do the sockets.

OP no mention fresh Windows install.
No mention memory set to 1333. Starting point.
Old overclock no worky worky the same with Xeon chip.
All ram slots populated, no mention of voltage bump.
No mention if X5690 came from server pull or it was overclocked to death in private hands.

When I booted up the Xeon CPU, the machine was configured with the default XMP profile and trying to run the RAM at 1333, but that didn't work so I switched all BIOS settings back to defaults and the RAM is currently trying to run only at 1066. I haven't manually adjusted any voltages and I didn't have any OC applied to the i7 CPU. As I understand it, the X5690 came from a server pull. Also, no, I was using an existing Windows 10 install, I didn't believe there a need for a fresh install just by changing the CPU.

Missing RAM on X58: Reinstall the CPU on the socket and make sure all pins aren't bent.

Yep, I'll pull the i7 again and have a more detailed look. Perhaps I missed something, but tbh, I'm very careful with this sort of thing anyway and I wouldn't jam shit in.

I have just Built a ASUS P6T Deluxe with a X5690 and it works Flawlessly (and water Cooled)
almost all retail Vendor Manafactures do not list Xeon Compatability yet they work................>> and that also goes for the workstation CPU's

You go thru the QVL list for Ram and ECC is not supported ( i got the manual open in front of me page 2-13 section 2.4.2)
Yet if you install a Xeon you can use DDR3 ECC (yeh bill we know why but tell us anyway )

Yeah, like, before I even attempted this new CPU, I did quite a bit of reading and the general consensus was that yes, the CPU should have just worked without me needing to dick around too much.

But as said, I'll pull the CPU again, check the pins/take some photos and post up so a 2nd or many sets of eyes can see if I'm missing anything. Will also pull the RAM and clean their contacts and try to clean out the slots too.

Cheers for all the suggestions, I'll update once I've had the chance to do this.

EDIT 1: Ok so I’ve taken some pics of the CPU socket/pins and of the bottom of the CPU itself. To me, there doesn’t look like anything is a problem there.

EDIT 2: And I've also cleaned all of the contacts of each RAM stick and blasted the slots with air to clean any dust out. Issue still remains.

 

[Devastator0]

I thought I would bump my post for those who previously commented to see if you have any further suggestions for me? Am I screwed?

 

[Caring1]

Socket and CPU appear to be OK, swap sticks of Ram around and ensure they are in the correct slots for triple channel, also do not over tighten the CPU cooler as this can case Ram issues.
You may have to reflash the boards BIOS just to be sure it was done correctly.

 

[Devastator0]

Socket and CPU appear to be OK, swap sticks of Ram around and ensure they are in the correct slots for triple channel, also do not over tighten the CPU cooler as this can case Ram issues.
You may have to reflash the boards BIOS just to be sure it was done correctly.

Yeah, I've been trying to swap the sticks around and the results are inconsistent. Sometimes the rig will detect the full 12GB, other times, only the 8GB. And yes, I also did reflash the bios to 1202 to see if it would fix this, but problem still is occurring.

I have a friend who has a machine with very similar specs and he's dropping it off with me tomorrow so I can pull the RAM from his machine and try it in mine to see if doing that will fix it. If it does, then I, somehow, need to get new triple channel RAM here in Australia...I don't know how that'll go...

 

[Regeneration]

Last time I had a similar issue, I solved it like this: install AIDA64, when the system fails to recognize all memory, launch it and go to Computer > DMI, and check exactly which module and slot cause problems. In my case, it was faulty memory.

 

[steen]

Second photo, lower right quadrant, pins look crescent shaped depressed. Surface corrosion? May be just camera artifact though. There's a bit of dust, so give it a bit of a blow out or gently dislodge with a brush if required. Maybe isopropyl/pcb cleaner/flux remover for socket caps too. Do as Caring1 suggests. Disable XMP & run RAM @ 1066. Will it detect 3x modules reliably?

 

[kastriot]

Your problems started when you changed cpu and when you putted your 920 you had problems still so it's issue with pins on mobo maybe some of them are more depressed than usual and not making good contact always because i don't see other issue than this, btw i have p6t V2 deluxe and 920+X5650 but i am now using other mobo, so look for depressed pins and you can raise them with needle of syringe but you must be very careful.

 

[basco]

sorry

 

[Candor]

Is it possible that the Xeon was faulty and has damaged your motherboard?

The issues remain with your i7 920 put back in so it seems to me that the root of the problem is that the Xeon was the beginning of the issues - therefore the cause.

If possible I'd try a bios flash as a last resort, but I'd say bad Xeon damaged your MB.

 

[Malik Sajid]

I once had a very similar issue on my lga775 P5Q-E mobo, I tried everything with no help!
In a desperate attempt to give my PC the kiss of life, I cut off the power completely of the mobo, here's what I've done..
First, I plugged out the power cord, disconnected the 24 pin & 8 pin cables.
Second, I removed the cmos battery.
Lastly I started to press the power on button like ten times then I held it for seconds. After that I left my PC for maybe an hour, that's of course not a role model but I'm telling my story.
My motherboard get rid of all of the residing power and that solved my problem I hope it'll solve yours.

Do that after you have exhausted all of your options.

 

[Ruru]

I once had a very similar issue on my lga775 P5Q-E mobo, I tried everything with no help!
In a desperate attempt to give my PC the kiss of life, I cut off the power completely of the mobo, here's what I've done..
First, I plugged out the power cord, disconnected the 24 pin & 8 pin cables.
Second, I removed the cmos battery.
Lastly I started to press the power on button like ten times then I held it for seconds. After that I left my PC for maybe an hour, that's of course not a role model but I'm telling my story.
My motherboard get rid of all of the residing power and that solved my problem I hope it'll solve yours.

Do that after you have exhausted all of your options.

In LGA775 the CPU didn't have a memory controller etc, so tips from those era motherboards aren't 100% working

 

[Malik Sajid]

In LGA775 the CPU didn't have a memory controller etc, so tips from those era motherboards aren't 100% working

I agree but he got nothing to lose. As I've said he should give it a try after exhausting all the options.

 

[Gasaraki]

OK, I actually have and am running a X58 system right now, unlike most of the people here that were probably in elementary school when X58 came out and now all they spew is crap like "crappy Intel pins", "not on the QVL list", etc, etc. That QVL list was released when the X58 and the i7s first came out. The X58 boards were updated with Xeon support in the later BIOSes. QVL lists don't get updated that often if at all. Only since Ryzen came out did they have to update them because ram support was so bad.

ANYWAY, I had a similar problem before where suddenly my RAM would disappear when I go change my cooler. Now I think you did what I did already but here's what I did. I pulled all my memory and plugged them back in to different slots. When it didn't work I would go in to the BIOS and see when ram slots it saw ram in and what it didn't. I would then pull those RAM and switch them with each other then see. Rise repeat. Eventually I got all 6 sticks recognized by the board.

What blue screen are you getting? Are you overclocking?

I think the RAM issues are because the board is tired.