On 12/29/2015 03:03 PM, Florian Gleixner wrote:

Run dmidecode and check which dimms are working. Try to find the one dimm that does not work and swap it with another one to see if the dimm or the slot is broken.

+1 Greg It will list the individual DIMM slots and show the population of each. I've never played with 64G, so don't know if there are any PAE type limits that apply, but I've never heard of any that would apply. God. Just thinking about the $100's spent for 64K of RAM not too long ago to 64G for roughly the same price.... wow. -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 12/29/2015 09:19 PM, James Knott wrote:

On 12/29/2015 09:50 PM, David C. Rankin wrote:

...

I've never played with 64G, so don't know if there are any PAE type limits that apply, but I've never heard of any that would apply.

PAE was to work around the 4 GB limit of 32 bit addresses. You have to go a bit more than 64 GB to run out of 64 bit address space.

Yes, I know, many terabytes before 64 bits play out. It was more of a euphemism about a generic limit :) There are some potential candidates with hardware/chip manufacturers including chips that do not handle a true 64-bits, that could effect the size reporting pipeline, but I have not read of any specific "be on the lookout for..." articles. (not like those fake 64G pen-drives.... -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

Greg Freemyer composed on 2015-12-30 12:04 (UTC-0500):

jdd wrote:

...

Greg Freemyer composed:

...

...

Memtest starts fine, but it is only reporting 48GB.

...

memtest list the memory slots, check this first

Per Memtest86 v4.20 in the DMI table area:

32G with success? Maybe it isn't even supposed to support 64. Regardless,

Could be test binary is your problem. Do we know anyone who has used it for try memtest86+ 5.01 or memtest 6.2 (neither of which I've been able to start from HD boot).

I have 8 slots. 3 have a 16GB DIMM installed.

B1 / C1 / D1 have DIMMs. A1 seems like it should. I'll take a look.

That is very helpful, thanks. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation)

Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Wed, Dec 30, 2015 at 1:14 PM, Brandon Vincent wrote:

On Wed, Dec 30, 2015 at 10:04 AM, Greg Freemyer wrote:

...

B1 / C1 / D1 have DIMMs. A1 seems like it should. I'll take a look.

The easiest solution is to remove all DIMMs except for A1. Then move A1 to D1 and see if the stick is recognized. If the stick refuses to show you probably have an issue with the memory. If the stick shows up, then try populating the rest back into the system and see if a different stick shows up in A1. If A1 is still missing a stick, the slot is likely bad.

64 GB is nothing for older versions of Memtest86+. The latest version now supports up to 2 TB of RAM.

Brandon Vincent

I have tried the "bad" DIMM in 3 slots standalone. In all 3 it works. In all 3 if I add a second DIMM, then the "bad" DIMM disappears and only the second one is visible. Is that a normal failure mode? If so, is it a bad DIMM, or something wrong with the MB? Thanks Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Wed, Dec 30, 2015 at 3:56 PM, John Andersen wrote:

On 12/30/2015 12:51 PM, Greg Freemyer wrote:

...

I have tried the "bad" DIMM in 3 slots standalone. In all 3 it works.

In all 3 if I add a second DIMM, then the "bad" DIMM disappears and only the second one is visible.

Is that a normal failure mode?

If so, is it a bad DIMM, or something wrong with the MB?

Are these all matched DIMMs? Same brand, same speed, same features?

I have seen this before, back in the day when trying to use ECC and non ECC in the same banks (paired dim slots).

In theory they are all the same. Corsair Vengeance DDR4. The PC was a custom build from Fry's Hopefully that didn't do any mix and match tricks. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Wed, Dec 30, 2015 at 4:26 PM, Brandon Vincent wrote:

On Wed, Dec 30, 2015 at 2:15 PM, Greg Freemyer wrote:

...

In theory they are all the same. Corsair Vengeance DDR4.

The PC was a custom build from Fry's Hopefully that didn't do any mix and match tricks.

Greg, what is the exact model of the motherboard and the BIOS revision you are on?

It looks like there are a few BIOS updates that add support for 16 GB DDR4 modules.

Brandon Vincent

The MB is a X99-Deluxe/U3.1 Part No: 90MB0JF1 I've just upgraded the BIOS to 2001, but it hasn't helped. I was at 1801 before. I also enabled XMP mode in the BIOS. That is supposed to optimize the speed of the RAM. No change in behavior. The RAM is Corsair Vengence LPX 16GB. I just put the "bad" DIMM in D2 (The others are A1, B1, C1). It still is not seen by the MB. I've tried the "bad" DIMM in all the x1 slots and now D2. It only works if it is the only DIMM in the MB. I did "chat" with ASUS support, but they didn't help much and when I didn't respond for a minute they terminated the chat. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Wed, Dec 30, 2015 at 4:48 PM, Greg Freemyer wrote:

I did "chat" with ASUS support, but they didn't help much and when I didn't respond for a minute they terminated the chat.

After looking up your particular motherboard, I'd be tempted to return it. That particular ASUS model seems to have a lot of issues detecting large amounts of RAM. Brandon Vincent -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 12/30/2015 07:50 PM, Greg Freemyer wrote:

I called Corsair and they informed me that since I bought 2 kits of 2x16 GB they were not guaranteed to work if installed at the same time. Wow!!! That's fussy.

I have added memory to both my desktop and ThinkPad computers. I didn't even use the same make of memory and all works fine. For example, my ThinkPad came with 4 GB of Samsung memory and added 4 GB of Kingston. It just worked without problem. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 12/31/2015 12:12 AM, Lew Wolfgang wrote:

...

I have added memory to both my desktop and ThinkPad computers. I didn't even use the same make of memory and all works fine. For example, my ThinkPad came with 4 GB of Samsung memory and added 4 GB of Kingston. It just worked without problem.

I have a hunch it's more related to signal timing and clock speed rather than capacity. Lower speed DRAM probably has fewer issues than high speed ones. I had a problem like Greg's on a SuperMicro server mobo some years ago. I never did get it working and just lived with 48-GB. I've got one now with 128-GB and it's rock-solid. Go figure...

That boils down to board quality. The desktop system got a new motherboard a couple of years back and has an Intel I7 CPU. I bought some memory initially and more some time later. Incidentally, many years ago, I used to work with core memory, which was made with tiny ferrite rings arranged in an array strung over extremely fine wires. Now that was fussy about timing in that if you replaced one inhibit line transistor or one of the driver/sense ICs connected directly to the array, you had to replace all of them with devices from the same batch. Those memory boards were either 8 or 16 KB on a 15" square board. On my shelf behind me, I have a core plane from a Collins B8500 computer. it contains 4096 cores and is about 5" square. Here's a picture of it: https://drive.google.com/file/d/0B5LapMwk8iPrNkFZUXA4cmd5dTg/view?usp=sharin... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Thu, Dec 31, 2015 at 11:13 AM, Lew Wolfgang wrote:

On 12/31/2015 04:17 AM, James Knott wrote:

...

Incidentally, many years ago, I used to work with core memory, which was made with tiny ferrite rings arranged in an array strung over extremely fine wires. Now that was fussy about timing in that if you replaced one inhibit line transistor or one of the driver/sense ICs connected directly to the array, you had to replace all of them with devices from the same batch. Those memory boards were either 8 or 16 KB on a 15" square board. On my shelf behind me, I have a core plane from a Collins B8500 computer. it contains 4096 cores and is about 5" square.

Wow! That jarred some memory bits loose! I used to care for early minicomputers that used core too. The HP 2100A was one of them, and I remember saving the day once by replacing one of those line transistors with a soldering iron. But in this case I didn't have to replace all of them. Those were the daze! I remember having to physically clean the disk drive heads on a monthly basis too. I wish I had managed to save one of those core planes...

I guess you were an expert debugger at the time! Even I'm too young to ever had to debug by removing bug carcasses from the electronics. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 12/31/2015 11:13 AM, Lew Wolfgang wrote:

I remember saving the day once by replacing one of those line transistors with a soldering iron

I didn't know you could replace a transistor with a soldering iron. I usually replaced them with another transistor. ;-) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 12/31/2015 01:28 PM, Lew Wolfgang wrote:

On 12/31/2015 09:53 AM, James Knott wrote:

...

On 12/31/2015 11:13 AM, Lew Wolfgang wrote:

...

I remember saving the day once by replacing one of those line transistors with a soldering iron I didn't know you could replace a transistor with a soldering iron. I usually replaced them with another transistor. ;-)

Touche! s/with/using

BTW, are the cores visible on your Collins module? There were always covers over the planes on the one's I've seen. They were so expensive I could never find the courage to take one apart.

Regards, Lew

That photo showed a single plane. There was a stack of 32 of them wired together, with the stack enclosed in the memory module case. The other memory I mentioned, on the 15" square board, had a plastic cover over the cores. You'd never want to leave the core array exposed, as it was so fragile. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

Greg Freemyer composed on 2015-12-30 15:51 (UTC-0500):

I have tried the "bad" DIMM in 3 slots standalone. In all 3 it works.

In all 3 if I add a second DIMM, then the "bad" DIMM disappears and only the second one is visible.

Is that a normal failure mode?

If so, is it a bad DIMM, or something wrong with the MB?

Maybe it's time to inquire of Asus tech support? -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Wed, Dec 30, 2015 at 5:21 PM, jdd wrote:

Le 30/12/2015 21:51, Greg Freemyer a écrit :

...

In all 3 if I add a second DIMM, then the "bad" DIMM disappears and only the second one is visible.

Is that a normal failure mode?

If so, is it a bad DIMM, or something wrong with the MB?

do memtest report exactly the same ram model number?

fyi: The model number was the same. So was the "version" number on the paper label on the DIMMs. Per the Tech Support guy, the version number reflects the source of the ICs (Integrated Circuits) on the DIMM. Every time the supplier of the ICs or the specs of the ICs change they change the version number. In my case all 4 sticks had the same version number. Even then the Tech Support guy said they was no guarantee they would work together. Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Thu 31 Dec 2015 10:33:12 AM CST, Greg Freemyer wrote:

On Wed, Dec 30, 2015 at 5:21 PM, jdd wrote:

...

Le 30/12/2015 21:51, Greg Freemyer a écrit :

...

In all 3 if I add a second DIMM, then the "bad" DIMM disappears and only the second one is visible.

Is that a normal failure mode?

If so, is it a bad DIMM, or something wrong with the MB?

do memtest report exactly the same ram model number?

fyi: The model number was the same. So was the "version" number on the paper label on the DIMMs.

Per the Tech Support guy, the version number reflects the source of the ICs (Integrated Circuits) on the DIMM.

Every time the supplier of the ICs or the specs of the ICs change they change the version number.

In my case all 4 sticks had the same version number. Even then the Tech Support guy said they was no guarantee they would work together.

Greg Hi Have you looked in the BIOS Ai Tweaker menu, check the RAM specs and set the voltage etc as per the specs rather than auto?

-- Cheers Malcolm °¿° LFCS, SUSE Knowledge Partner (Linux Counter #276890) SUSE Linux Enterprise Desktop 12 | GNOME 3.10.1 | 3.12.51-52.31-default up 2 days 23:50, 4 users, load average: 0.33, 0.36, 0.31 CPU AMD A4-5150M @ 2.70GHz | GPU Radeon HD 8350G -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Thu, Dec 31, 2015 at 9:17 AM, Greg Freemyer wrote:

Time to take Corsair up on their offer to swap out the DIMMs.

Hopefully that should take care of your issue. The recommended memory compatibility chart from ASUS only shows that they've tested 16 GB registered DIMMs. Next time you start a build I highly recommend you pick memory modules off the list. I'm still wagering that this motherboard is very picky. You have to realize that unbuffered 16 GB DDR4 modules are only a few months old. Good luck with this system. It seems like it will be quite impressive when finished. Brandon Vincent -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

Le 31/12/2015 18:15, Greg Freemyer a écrit :

At the time I just called Fry's and had them pick the components. Last time I will do that I guess.

make them assemble the computer and test it... jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 16/01/16 08:27 PM, Greg Freemyer wrote:

...

To be honest, I would not object to having a second similar spec machine, Maybe I should buy new DIMs which have been tested with the ASUS X99-Deluxe, then buy a new MB / CPU pair that have been tested with the unbuffered DIMs I have?

At this point, this seems like a good idea -- and in future, always check the board mfr's site for memory it has been tested with. Not sure if all mfr's do this, though, but if they do, only using what they've tested would probably be the best idea. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Sat, Jan 16, 2016 at 9:27 PM, Greg Freemyer wrote:

The MB now sees 6 of them for 96GB total.

The missing DIMs are in A1 and A2.

I'm not sure how much speed I'm losing by that config, but I'm more than happy with 96GB.

Strange how Intel claims that you can only use 64GB RAM with that cpu, but you have 96GB working...... http://ark.intel.com/products/82932/Intel-Core-i7-5820K-Processor-15M-Cache-... I would see what combination of old & new sticks work to give you 64GB and RMA the rest. If you can't get any combination to work, then I'd say the board is defective. Seems like the board since A1/A2 are bad. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

John Andersen wrote:

On 01/18/2016 03:09 AM, Larry Stotler wrote:

...

On Sat, Jan 16, 2016 at 9:27 PM, Greg Freemyer wrote:

...

The MB now sees 6 of them for 96GB total.

The missing DIMs are in A1 and A2.

I'm not sure how much speed I'm losing by that config, but I'm more than happy with 96GB.

Strange how Intel claims that you can only use 64GB RAM with that cpu, but you have 96GB working......

http://ark.intel.com/products/82932/Intel-Core-i7-5820K-Processor-15M-Cache-...

I don't see any such specification on that page.

John, get your glasses on - Memory Specifications Max Memory Size (dependent on memory type) 64 GB /Per -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

Le 18/01/2016 21:01, Greg Freemyer a écrit :

If there were wrap around at 64GB I assume memtest would detect that?

most probably not, because why should a mobo do that? if you want, get a look at f3write http://oss.digirati.com.br/f3/ that do the test for sd cards, it should be easy to change it do do this for ram. the trick is to use a counter to write to the memory files with the count in them. after that you read, if after some time the read do not fit with the number of the file, it's paged jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Thu, Dec 31, 2015 at 5:15 AM, Oliver Kurz wrote:

On Tuesday 29 December 2015 15:49:04 Greg Freemyer wrote:

...

[…] Is there a Live Leap DVD I can try?

There is no official Live DVD for Leap but for Tumbleweed there are: https://en.opensuse.org/openSUSE:Tumbleweed_installation#LiveCDs.2FRescue_CD

Great. I'm surprised to see a 32-bit live DVD. But I will pull only the 64-bit! Greg -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org