Upgrading memory modules

Hi, I am wondering if it may be worth upgrading the memory modules in my computer. Currently installed I have 2, 667mhz DIMMS and 2, 800mhz DIMMS installed, and the machine is brought down to 667mhz (respectively). I passed inxi -m and the result is as follows:

*-memory
          description: System Memory
          physical id: 1f
          slot: System board or motherboard
          size: 8GiB
        *-bank:0
             description: DIMM DDR2 Synchronous 667 MHz (1.5 ns)
             product: 4A4D383030514C552D324720202020202020
             vendor: Transcend Information
             physical id: 0
             serial: 0008770C
             slot: J6G1
             size: 2GiB
             width: 40968 bits
             clock: 667MHz (1.5ns)
        *-bank:1
             description: DIMM DDR2 Synchronous 667 MHz (1.5 ns)
             product: 4A4D383030514C552D324720202020202020
             vendor: Transcend Information
             physical id: 1
             serial: 0008770C
             slot: J6G2
             size: 2GiB
             width: 41480 bits
             clock: 667MHz (1.5ns)
        *-bank:2
             description: DIMM DDR2 Synchronous 667 MHz (1.5 ns)
             product: 393955353331362D3030312E4130324C4600
             vendor: Kingston
             physical id: 2
             serial: 00000009
             slot: J6H1
             size: 2GiB
             width: 41992 bits
             clock: 667MHz (1.5ns)
        *-bank:3
             description: DIMM DDR2 Synchronous 667 MHz (1.5 ns)
             product: 393955353331362D3030312E4130324C4600
             vendor: Kingston
             physical id: 3
             serial: 00000009
             slot: J6H2
             size: 2GiB
             width: 42504 bits
             clock: 667MHz (1.5ns)

While I reviewed some of the ‘Hardware Maintenance Manual’ the DIMM (one of the memory sections) declair: Memory module 2GB PC2-5300 (667) mhz DDR2 SDRAM models (B2U B2F B3U B3F) looks like the type of memory the machine calls for. (page 293)

  1. My question for you is what will happen if one installs all PC2-5300 (800) mhz modules (in all 4 DIMMS)? Will the system revert to 667 mhz or perhaps have a mainboard failure? Is it even possible?
  2. Should I just leave well enough alone for now so to say?
  3. If the Hardware Service manual specifies this type of memory is there little or no room for difference or change of DIMMS here?

“Hardware Maintenance Manual” M57p type 9088. → https://download.lenovo.com/ibmdl/pub/pc/pccbbs/thinkcentre_pdf/43c8797.pdf

Thanks for your insight on this.

PC2-5300 is 667 mhz (DDR2-667). PC2-6400 is 800 mhz (DDR2-800). PC3-8500 is 1066 mhz (DDR3-1066) and requires ~40% of the power DDR2 requires.

Some chipsets of the transition from 667 to 800 era only support 800 for installation of two sticks, and run at 667 with 4 sticks installed. If you can’t find a manual to say so, the only way to know is to have the required hardware pieces and use a memory tester. Speed differences resulting from improperly matching will be readily apparent.

What’s also important is to realize most if not all support dual channel mode, which depends on putting matching pairs in correct slots. Many motherboards colored the two pairs differently to facilitate matching whenever all 4 sticks installed were constituted of two matched pairs that don’t match each other. That description matches your parts, so you should be sure which slots comprise each pair, and be sure to properly match Transcend with Transcend and Kingston with Kingston. Not doing so can cost almost half the potential RAM speed.

It is worth upgrading the whole PC. DDR2 in 2024?
In your case: it is possible to run all memory modules at 800 MHz by replacing 667 MHz modules with 800 MHz. And it is possible to overclock 667 MHz modules to 800 MHz, possibly with raising voltage.

You can buy AM4 system at bargain price and get 2400 MHz at least with DDR4. New AM4 A6 APU costs about $10.

Thank you for your insight on this. I opened the machines case and have taken a more in depth look at the mainboards DIMM slots. There are 2 blue colored slots next to one another and 2 green colored slots next to each other. The Transcend DIMMS are both in the blue slots and the Kingston Dimms are both in the green slots. So that part of it is sorted out to a degree.

What is the type of memory tester you most recommend from your experiences?

Could doing this cause the overclocked DIMMS to create more heat and fail sooner?
What is the type of memory tester you most recommend from your experiences?

-Thanks

Likely you have one ready for your use in your Grub menu’s advanced options. If not, you can put one of those suggested by Svyatko there using package management.

That’s funny because I just purchased a z820 with 128 gb of memory and ran memtest 10.6 in form of a bootable usb medium. 9 passes took about 11 days it didn’t find a problem ((there is a random powercycle problem (z820 not M57p) though)). I would like to open up another post about the situation with it at some point here.

-Thanks

@panorain if running openSUSE then install rasdaemon to keep an eye on the system…

ras-mc-ctl --summary
No Memory errors.

No PCIe AER errors.

No ARM processor errors.

No Extlog errors.

No devlink errors.
No disk errors.
No Memory failure errors.

No MCE errors.

I was not aware of rasdaemon whatsoever. I have installed rasdaemon and powercycled.

Thinkcentre-M57p:~> sudo ras-mc-ctl --summary
[sudo] password for root: 
DBD::SQLite::db prepare failed: no such table: mc_event at /usr/sbin/ras-mc-ctl line 1183.
Can't call method "execute" on an undefined value at /usr/sbin/ras-mc-ctl line 1184.
Thinkcentre-M57p:~>

I would like to ask about these output messages displayed above. Is this because there is log file as of yet?

-Thanks

@panorain has the daemon been enabled and started? Not sure it will work on that system, but will on the Z series workstation since it has ECC RAM…

I’m not certain yet as to how to enable or start the daemon. I looked a little more deeply. Looks like rasdaemon mapage https://manpages.opensuse.org/Tumbleweed/rasdaemon/rasdaemon.1.en.html

Thinkcentre-M57p:~ # ras-mc-ctl --summary
DBD::SQLite::db prepare failed: no such table: mc_event at /sbin/ras-mc-ctl line 1183.
Can't call method "execute" on an undefined value at /sbin/ras-mc-ctl line 1184.
Thinkcentre-M57p:~ # ras --help

The program 'ras' can be found in the following package:
  * opae [ path: /usr/bin/ras, repository: openSUSE:repo-oss ]

Try installing with:
    sudo zypper install opae

Thinkcentre-M57p:~ #
Thinkcentre-M57p:~> zypper search -i opae
Loading repository data...
Reading installed packages...
No matching items found.
Thinkcentre-M57p:~>
Thinkcentre-M57p:~> zypper search -i rasdaemon
Loading repository data...
Reading installed packages...

S  | Name      | Summary                               | Type
---+-----------+---------------------------------------+--------
i+ | rasdaemon | Utility to receive RAS error tracings | package
Thinkcentre-M57p:~>

Wondering if I should try installing opae? Why did yast software manager not pull opae in as dependency I wonder to myself? hmm

What do you think? I will rest now for a while.
Going to powercycle and run memtest on the machine while sleeping.

-Thanks

@panorain it’s ras-mc-ctl --help did you start the rasdaemon service?

systemctl start rasdaemon

Hi, I unfortunately forgot to start the service.

Thinkcentre-M57p:~> systemctl status rasdaemon
● rasdaemon.service - RAS daemon to log the RAS events
     Loaded: loaded (/usr/lib/systemd/system/rasdaemon.service; disabled; preset: disabled)
     Active: active (running) since Sat 2024-01-13 11:11:12 CST; 6min ago
    Process: 7110 ExecStartPost=/usr/sbin/rasdaemon --enable (code=exited, status=0/SUCCESS)
   Main PID: 7108 (rasdaemon)
      Tasks: 1 (limit: 4915)
        CPU: 61ms
     CGroup: /system.slice/rasdaemon.service
             └─7108 /usr/sbin/rasdaemon -f -r
Thinkcentre-M57p:~>

rasdaemon appears to be correctly monitoring now:

I let memtest86 run 5 passes, it displayed no errors on the Lenovo m57P.
Am I correct to assume that rasdaemon is now monitoring the system (I believe this is the case)?
Thanks for making me aware of rasdaemon.

Thinkcentre-M57p:~> sudo ras-mc-ctl --summary
[sudo] password for root: 
No Memory errors.

No PCIe AER errors.

No ARM processor errors.

No Extlog errors.

No devlink errors.
No disk errors.
No Memory failure errors.

No MCE errors.
Thinkcentre-M57p:~>

I possibly never could have installed rasdaemon on the HP z820 here due to it’s random powercycling after 10-15 minutes of uptime. Then I made things worse by cooking it’s memory in the oven (415 F for 15 minutes) and reinstalling the memory. I think doing that, (I) took out the mainboard by doing so.

Thanks

@panorain why would you do that with the memory?

ras-mc-ctl --status
ras-mc-ctl: drivers are loaded.

ras-mc-ctl --mainboard
ras-mc-ctl: mainboard: Hewlett-Packard model 212B

ras-mc-ctl --layout
       +-----------------------------------------------------------------------------------------------------------------------------------------------+
       |                                  mc0                                  |                                  mc1                                  |
       | channel0  | channel1  | channel2  | channel3  | channel4  | channel5  | channel0  | channel1  | channel2  | channel3  | channel4  | channel5  |
-------+-----------------------------------------------------------------------------------------------------------------------------------------------+
slot2: |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |
slot1: |  16384 MB  |  16384 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |  16384 MB  |  16384 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |
slot0: |  16384 MB  |  16384 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |  16384 MB  |  16384 MB  |     0 MB  |     0 MB  |     0 MB  |     0 MB  |
-------+---------------------------------------------------------------------------------------------------------------------------------------------------+

It is some nonsense spread by some ppl …not recommended if you want to keep your hardware any longer…
https://www.google.com/search?q=baking+ram

I did that with a ACER laptop mainboard prior and had success for a year (It seems now to have reverted to it’s prior state again). I had to heat the laptop base up with a hairdryer in order for the system to POST then continue to the OS.

I made a bad decision with the z820 that’s exactly what I did.

Thinkcentre-M57p:~> sudo ras-mc-ctl --status
ras-mc-ctl: drivers not loaded.
Thinkcentre-M57p:~>

Oh, wow no drivers are loaded as of yet.

-Thanks

No, the service is disabled.

Thinkcentre-M57p:~> systemctl status rasdaemon
● rasdaemon.service - RAS daemon to log the RAS events
     Loaded: loaded (/usr/lib/systemd/system/rasdaemon.service; disabled; preset: disabled)
     Active: active (running) since Sat 2024-01-13 11:11:12 CST; 1h 0min ago
    Process: 7110 ExecStartPost=/usr/sbin/rasdaemon --enable (code=exited, status=0/SUCCESS)
   Main PID: 7108 (rasdaemon)
      Tasks: 1 (limit: 4915)
        CPU: 61ms
     CGroup: /system.slice/rasdaemon.service
             └─7108 /usr/sbin/rasdaemon -f -r
Thinkcentre-M57p:~>

-Thanks

@panorain It’s designed for edac drivers, so may not be working, will run but not log anything. It will work on the HP…

https://documentation.suse.com/smart/systems-management/html/reference-systemctl-enable-disable-services/index.html

Thinkcentre-M57p:~> systemctl enable rasdaemon
Created symlink /etc/systemd/system/multi-user.target.wants/rasdaemon.service → /usr/lib/systemd/system/rasdaemon.service.
paul@paul-Thinkcentre-M57p:~> systemctl status rasdaemon
● rasdaemon.service - RAS daemon to log the RAS events
     Loaded: loaded (/usr/lib/systemd/system/rasdaemon.service; enabled; preset: disabled)
     Active: active (running) since Sat 2024-01-13 11:11:12 CST; 1h 11min ago
   Main PID: 7108 (rasdaemon)
      Tasks: 1 (limit: 4915)
        CPU: 61ms
     CGroup: /system.slice/rasdaemon.service
             └─7108 /usr/sbin/rasdaemon -f -r

I’m going to powercycle the machine now.

This is the current rasdaemon service state after powercycle:

Thinkcentre-M57p:~> systemctl status rasdaemon
● rasdaemon.service - RAS daemon to log the RAS events
     Loaded: loaded (/usr/lib/systemd/system/rasdaemon.service; enabled; preset: disabled)
     Active: active (running) since Sat 2024-01-13 12:32:07 CST; 1min 11s ago
    Process: 1179 ExecStartPost=/usr/sbin/rasdaemon --enable (code=exited, status=0/SUCCESS)
   Main PID: 1178 (rasdaemon)
      Tasks: 1 (limit: 4915)
        CPU: 23ms
     CGroup: /system.slice/rasdaemon.service
             └─1178 /usr/sbin/rasdaemon -f -r

Warning: some journal files were not opened due to insufficient permissions.
Thinkcentre-M57p:~> 

Warning: some journal files were not opened due to insufficient permissions.

Is this what you are referring to when you say:

Can I force this service somehow that you are aware of? Could there be an error in a configuration file here that you aware of?

Is this z820 barebones a good price? I verified I have v2 cpu here (003 or 004 top line end of identification tag on mainboard). HP Z820 Workstation Barebones 2x Heatsink 4x Trays 1125W PSU No CPU No RAM No HD | eBay

-Thanks