Xeon Owners Club

[lexluthermiester]

Bummer I don't have another CPU to toss into the unit till I get my hands on the x5675...

You can still get a bit of life out of it. Under clock it. Run it at a lower than stock voltage at 2.2ghz. It'll be slower but at least you'll have a functional CPU until that replacement arrives, presuming that this is your daily driver system..

 

[SeanBeanDiesInTheEnd]

You can still get a bit of life out of it. Under clock it. Run it at a lower than stock voltage at 2.2ghz. It'll be slower but at least you'll have a functional CPU until that replacement arrives, presuming that this is your daily driver system..

Daily driver 'entertainment' PC haha, Macbook is the legit daily driver for life tasks. Doing your suggestion atm and running 3.8GHz fine for the last 3 hours, just bummed on the dual channel bit... Was gonna carve out some time to play CP2077 tonight, guess I'll have to tolerate choppy frames for the next while till I can test the x5675 :/

Not gonna lie, had a mini heart attack thinking I fried the P6T DIMM's...

 

[Edwired]

I think lexluthermiester is right about the cpu degrading giving off the symptoms. If the memory controller (imc) is dying it would generally drop a memory channel completely sometimes the socket pins not connecting correctly to the memory dimms can cause this as well and last the worse case could be the board itself.

At least try taking out the cpu and inspect for burn marks/corroded pads and give it a clean with a soft rubber eraser but be careful not to knock off the caps on the back of the cpu.

And looking at the voltages for qpi that the cpu experienced pretty much it digging a grave for itself

SeanBeanDiesInTheEnd what memory are you using as it can be found in aida64 as I need to check it up on specs online
Is it this one you got installed

F3-12800CL9D-8GBRL - Overview - G.SKILL International Enterprise Co., Ltd.

G.SKILL Ripjaws is the award winning classic DDR3 memory kit that just works. With the widest range of DDR3 memory compatibility across Intel and AMD platforms, we covered all the bases for you, so you won't have to.

www.gskill.com

 

[80-watt Hamster]

I think lexluthermiester is right about the cpu degrading giving off the symptoms. If the memory controller (imc) is dying it would generally drop a memory channel completely sometimes the socket pins not connecting correctly to the memory dimms can cause this as well and last the worse case could be the board itself.

At least try taking out the cpu and inspect for burn marks/corroded pads and give it a clean with a soft rubber eraser but be careful not to knock off the caps on the back of the cpu.

And looking at the voltages for qpi that the cpu experienced pretty much it digging a grave for itself

SeanBeanDiesInTheEnd what memory are you using as it can be found in aida64 as I need to check it up on specs online
Is it this one you got installed

F3-12800CL9D-8GBRL - Overview - G.SKILL International Enterprise Co., Ltd.

G.SKILL Ripjaws is the award winning classic DDR3 memory kit that just works. With the widest range of DDR3 memory compatibility across Intel and AMD platforms, we covered all the bases for you, so you won't have to.

www.gskill.com

Hmmm... maybe I should try re-populating the "dead" channel of my Sabertooth, now that there's a new CPU in it. (PO of the board said the channel was bad; never fully investigated.) And then perhaps inspect pins.

 

[Edwired]

That what I would do. Then again swap around the rams and test it, it might get lucky it could post all dimms populated then again best method is to get a working ram and test each dimm and do the same method on each ram to see if it is working as it should be tested with memtest via usb or cd this way you can spot if the ram is working in stock speed, timings, voltage.

If the ram starts throwing errors in stock state it pretty much faulty. In some cases it could be dirty pins on the ram pcb that might need to be cleaned with a soft rubber eraser and a blow out with a compressed air can for good measure.

 

[SeanBeanDiesInTheEnd]

SeanBeanDiesInTheEnd what memory are you using as it can be found in aida64 as I need to check it up on specs online
Is it this one you got installed

F3-12800CL9D-8GBRL - Overview - G.SKILL International Enterprise Co., Ltd.

G.SKILL Ripjaws is the award winning classic DDR3 memory kit that just works. With the widest range of DDR3 memory compatibility across Intel and AMD platforms, we covered all the bases for you, so you won't have to.

www.gskill.com

Using exactly these ones (6 x 2GB):

G.SKILL Ripjaws X Series 4GB (2 x 2GB) DDR3 1600 (PC3 12800) Desktop Memory Model F3-12800CL9D-4GBXL - Newegg.com

Buy G.SKILL Ripjaws X Series 4GB (2 x 2GB) DDR3 1600 (PC3 12800) Desktop Memory Model F3-12800CL9D-4GBXL with fast shipping and top-rated customer service. Once you know, you Newegg!

www.newegg.ca

I'll check the CPU out over the weekend, wifey's catching up on big brother so no go today

That what I would do. Then again swap around the rams and test it, it might get lucky it could post all dimms populated then again best method is to get a working ram and test each dimm and do the same method on each ram to see if it is working as it should be tested with memtest via usb or cd this way you can spot if the ram is working in stock speed, timings, voltage.

If the ram starts throwing errors in stock state it pretty much faulty. In some cases it could be dirty pins on the ram pcb that might need to be cleaned with a soft rubber eraser and a blow out with a compressed air can for good measure.

The asus boards are certainly finicky when it comes to all 6 DIMM's populated. Heck once upon a time setting the DRAM manually @ 9-9-9-24-208 (CAS 9, tRCD 9, tRP 9, tRAS 24, tRFC 208) fixed it 2 years ago and then it just "worked" when I set XMP getting rid of the tRFC value to auto. Another time 5 years ago adding one stick at a time and rebooting each time till all 6 banks were full did it register 12288mb triple channel... ASUS at it's finest

 

[Toothless]

Using exactly these ones (6 x 2GB):

G.SKILL Ripjaws X Series 4GB (2 x 2GB) DDR3 1600 (PC3 12800) Desktop Memory Model F3-12800CL9D-4GBXL - Newegg.com

Buy G.SKILL Ripjaws X Series 4GB (2 x 2GB) DDR3 1600 (PC3 12800) Desktop Memory Model F3-12800CL9D-4GBXL with fast shipping and top-rated customer service. Once you know, you Newegg!

www.newegg.ca

I'll check the CPU out over the weekend, wifey's catching up on big brother so no go today


The asus boards are certainly finicky when it comes to all 6 DIMM's populated. Heck once upon a time setting the DRAM manually @ 9-9-9-24-208 (CAS 9, tRCD 9, tRP 9, tRAS 24, tRFC 208) fixed it 2 years ago and then it just "worked" when I set XMP getting rid of the tRFC value to auto. Another time 5 years ago adding one stick at a time and rebooting each time till all 6 banks were full did it register 12288mb triple channel... ASUS at it's finest

I've noticed x58 gets cranky with ram and moreso depending on the brand. My SR-2 rarely boots with all 36GB there, randomly dropping banks. I've resorted to getting a total of 12 G.Skill Sniper for the boards that are overclocking. Helps with bclk and ram clocks.

 

[lexluthermiester]

My SR-2 rarely boots with all 36GB there, randomly dropping banks.

Still? I thought you had that sorted out...

 

[Edwired]

Using exactly these ones (6 x 2GB):

G.SKILL Ripjaws X Series 4GB (2 x 2GB) DDR3 1600 (PC3 12800) Desktop Memory Model F3-12800CL9D-4GBXL - Newegg.com

Buy G.SKILL Ripjaws X Series 4GB (2 x 2GB) DDR3 1600 (PC3 12800) Desktop Memory Model F3-12800CL9D-4GBXL with fast shipping and top-rated customer service. Once you know, you Newegg!

www.newegg.ca

I'll check the CPU out over the weekend, wifey's catching up on big brother so no go today


The asus boards are certainly finicky when it comes to all 6 DIMM's populated. Heck once upon a time setting the DRAM manually @ 9-9-9-24-208 (CAS 9, tRCD 9, tRP 9, tRAS 24, tRFC 208) fixed it 2 years ago and then it just "worked" when I set XMP getting rid of the tRFC value to auto. Another time 5 years ago adding one stick at a time and rebooting each time till all 6 banks were full did it register 12288mb triple channel... ASUS at it's finest

It dawned on me what dram voltage were you using before the ram started disappearing as it might be related to too low voltage to run fully populated dimms. Try bumping the dram voltage to 1.60v

 

[lexluthermiester]

Try bumping the dram voltage to 1.60v

That might work...

 

[Edwired]

That might work...

I hope so as I had a look at his ram spec it stated 1.50v to run. I'm banking on that it might be the possible fix to his problem with disappearing rams. As we know more dimms is populated the more dram voltage is needed same with the qpi/dram core voltage have to be bumped up in small steps for it to work.

 

[SeanBeanDiesInTheEnd]

No go jump starting with the DRAM v. Tried 1.56, 1.6, 1.62, and danger zone 1.66 all still registering 8gb in dual only. Even dropped it down to 1066mhz on stock speeds with an UCLK of 2132, haven’t felt slowness that slow in a looooong time. If it didn’t show me the RAM I think I’d actually sleep easier tonight

 

[Toothless]

Still? I thought you had that sorted out...

I wish. I just keep rebooting til it's there and go on my merry way.

 

[bobbybluz]

Years ago in my conversations with The RAM Guy at Corsair he told me DDR3 is easily safe to 1.65 volts and that large amounts require more voltage plus put heavier strains on the memory controller. This was back when I first put 64GB in my Sabertooth X79 circa late 2012. He said increasing the voltage to 1.65 volts (8 x 8GB DDR3 1866 at that time) would take some of the load off the memory controller. I did so and have repeated with other X79 builds with perfect success. Some may remember that Mushkin, OCZ and a few other brands specified 1.65 volts as the default for their top performance RAM back then.

 

[Edwired]

No go jump starting with the DRAM v. Tried 1.56, 1.6, 1.62, and danger zone 1.66 all still registering 8gb in dual only. Even dropped it down to 1066mhz on stock speeds with an UCLK of 2132, haven’t felt slowness that slow in a looooong time. If it didn’t show me the RAM I think I’d actually sleep easier tonight

View attachment 198685

What concerning me is that your board only goes to
DRAM REF Cycle Time [110 DRAM Clock] = Row Refresh Cycle Time (tRFC) but the ram spec for tRFC is 208.
I'm kinda confused by that

Take a screenshot of chipset -> North Bridge: Intel Westmere IMC
This will list the actual motherboard timings that it currently using

 

[Mr.Scott]

What bclk are you still running? Lower it if you're still in the 190's. High bclk will cause dropped channels sometimes.

 

[80-watt Hamster]

That what I would do. Then again swap around the rams and test it, it might get lucky it could post all dimms populated then again best method is to get a working ram and test each dimm and do the same method on each ram to see if it is working as it should be tested with memtest via usb or cd this way you can spot if the ram is working in stock speed, timings, voltage.

If the ram starts throwing errors in stock state it pretty much faulty. In some cases it could be dirty pins on the ram pcb that might need to be cleaned with a soft rubber eraser and a blow out with a compressed air can for good measure.

I think this is my favorite thread on TPU right now. Lots of good info, and nobody's getting all butthurt over anything.

Currently up and running with 3 x 2gb 1333, btw. But I REALLY don't like the DIMM sockets on this board...

 

[Toothless]

What bclk are you still running? Lower it if you're still in the 190's. High bclk will cause dropped channels sometimes.

It's whatever the maxed multi is with bclk to get 4ghz, which is... gotta boot that desktop... 160 bclk.

 

[Edwired]

I think this is my favorite thread on TPU right now. Lots of good info, and nobody's getting all butthurt over anything.

Currently up and running with 3 x 2gb 1333, btw. But I REALLY don't like the DIMM sockets on this board...

You talking about the Sabertooth dimms. Is it x58 or the x79 I wonder

 

[SeanBeanDiesInTheEnd]

What concerning me is that your board only goes to
DRAM REF Cycle Time [110 DRAM Clock] = Row Refresh Cycle Time (tRFC) but the ram spec for tRFC is 208.
I'm kinda confused by that

Take a screenshot of chipset -> North Bridge: Intel Westmere IMC
This will list the actual motherboard timings that it currently using

It’s an interesting one, if set to auto it hits the value needed by the XMP profile, but manually can only set to 110.

The more I think about it pulling the cpu will likely do the trick by clearing the memory controller in some way if it’s holding some faulty data from the prime95 crash that caused it to fubar in buffer space. There’s no definitive info on this, but it kinda makes sense if you think about it. And if this the case, man that would be swell and clear up so many myths about this board that never went answered.

What bclk are you still running? Lower it if you're still in the 190's. High bclk will cause dropped channels sometimes.

Not sure if this one’s for me, but if so my 3.8 is 19*200, what value do you suggest?

I think this is my favorite thread on TPU right now. Lots of good info, and nobody's getting all butthurt over anything.

Currently up and running with 3 x 2gb 1333, btw. But I REALLY don't like the DIMM sockets on this board...

Not many x58 threads left alive out there. One on [H], and another on LTT that’s mostly server speak, this one still has OG enthusiasts

 

[80-watt Hamster]

You talking about the Sabertooth dimms. Is it x58 or the x79 I wonder

X58. Alignment is finicky, and they require too much insertion force.

 

[Edwired]

It’s an interesting one, if set to auto it hits the value needed by the XMP profile, but manually can only set to 110.

The more I think about it pulling the cpu will likely do the trick by clearing the memory controller if in some way if it’s holding some faulty data from the prime95 crash that caused it to fubar in buffer space. There’s no definitive info on this, but it kinda makes sense if you think about it. And if this the case, man that would swell and clearly up so many myths about this board that never went answered.


Not sure if this one’s for me, but if so my 3.8 is 19*200, what value do you suggest?

That good to know sure keep us updated on the progress

 

[SeanBeanDiesInTheEnd]

That good to know sure keep us updated on the progress

Hundo%

X58. Alignment is finicky, and they require too much insertion force.

My C1 slot, never have I used so much force LOL.

 

[Edwired]

X58. Alignment is finicky, and they require too much insertion force.

Jeez it must be that tight to fit a ram in the slot. You sure the dimm slots isn't warped or looks pinched?

 

[80-watt Hamster]

My C1 slot, never have I used so much force LOL.

A1's the worst on mine. Sometimes it wiggles so badly I fear for the solder joints.

Jeez it must be that tight to fit a ram in the slot. You sure the dimm slots isn't warped or looks pinched?

I don't think so. They're single latch, so you've got to make DAMN sure modules are nested well on the fixed end.