RAM Test

[anonprivate]

Not quite good enough ... be sure to pull the plug out of the wall first!

Would leaving the plug in, but switched off, be a good idea because it give a route to earth

[anonprivate]

Hmmm, that sounds about right!! A 220 volt leak into the case in anonprivate's UK, would be a stinker ...

More like 240v

[jglen490]

Ahh, even better/worse!

Seriously, unplug it. As long as you DO NOT ground to the motherboard or the RAM unit, you'll be fine. Grounding yourself to the case/chassis is perfectly acceptable.

[GreyGeek]

Would leaving the plug in, but switched off, be a good idea because it give a route to earth

Never do that!
If the PC is plugged in but even switched off there are still one or more hot points in the power supply. If a component is faulty or leaky YOU might become a path to ground.

Unplugged, all that is needed is that you wait for a minute or so to give power capacitors a chance to drain before you attach a grounding wire to the PS chassis and your wrist. That allows you and the mobo to be at the same electromotive potential before you start pulling old RAM and plugging in the new RAM.

If your work area is carpeted and you walk across the carpet to the PC and touch the unplugged and drained PC you could still build up enough static electric charge to create a 5,000 to 15,000 volt discharge spark. It won’t have enough current to kill you but it could kill a delicate component.



Sent from my iPhone using Tapatalk

[anonprivate]

I have removed the existing RAM and installed 4 x 1GB of memory. It is reading 3.9 GB

It looks a lot better:

andrew@andrew-Dell-DM061:~$ free -h
          total       used       free     shared    buffers     cached
Mem:          3.9G       2.2G       1.7G        83M        70M       722M
-/+ buffers/cache:       1.4G       2.5G
Swap:         1.0G         0B       1.0G

I have seven Firefox windows minimised and the Terminal; and four virtual desktops (no open windows)

I am still experimenting

[anonprivate]

Why is the system using the swap partition when RAM is still available?

ndrew@andrew-Dell-DM061:~$ free -h
            total       used       free     shared    buffers     cached
Mem:          3.9G       3.3G       570M        94M        91M       985M
-/+ buffers/cache:       2.3G       1.6G
Swap:         1.0G        13M       1.0G

[GreyGeek]

Using 13/1000th of your swap is using much. Notice that your Free equals 1.0G as well. Like memory, it often isn't cleared, just overwritten the next time it is needed.

[anonprivate]

Using 13/1000th of your swap is using much. Notice that your Free equals 1.0G as well. Like memory, it often isn't cleared, just overwritten the next time it is needed.

Does this mean that I must have exceeded 4 GB of RAM at some stage, and this partition has not cleared completely. Hence I see usage.

[GreyGeek]

Not necessarily.

[jlittle]

Does this mean that I must have exceeded 4 GB of RAM at some stage, and this partition has not cleared completely. Hence I see usage.

I'm not sure about this, but I think some software can map a lot of memory initially, without using it.* So, the swap file usage does not always indicate anything has been swapped out. Once some OS stuff has been mapped to swap it tends to sit there.

Firefox used to (a few years ago) do this for me for each tab on Google groups, about 300 MiB, when I only had 1 and 2 GiB; it took me months to work out what was happening, Kubuntu thrashed for several minutes, and by the time I could run monitors like top the usage had dropped. Firefox fixed part of the problem, then reduced memory usage again, and I got more RAM, so my problem went away. I'm not a Google Chrome or Chromium user, but it has a reputation for doing things like this for script-laden news web sites; for some users Windows doesn't cope well and tabs crash a lot.

Linux I/O buffering can have a similar effect when copying large files, on some old-ish hardware, especially using the CFQ scheduler. I never understood what was happening with that, or why it wasn't considered a serious bug.

*(I've written code like that, for some tasks it's quite liberating, no I/O to worry about.)

[anonprivate]

I'm not sure about this, but I think some software can map a lot of memory initially, without using it.* So, the swap file usage does not always indicate anything has been swapped out. Once some OS stuff has been mapped to swap it tends to sit there.

Firefox used to (a few years ago) do this for me for each tab on Google groups, about 300 MiB, when I only had 1 and 2 GiB; it took me months to work out what was happening, Kubuntu thrashed for several minutes, and by the time I could run monitors like top the usage had dropped. Firefox fixed part of the problem, then reduced memory usage again, and I got more RAM, so my problem went away. I'm not a Google Chrome or Chromium user, but it has a reputation for doing things like this for script-laden news web sites; for some users Windows doesn't cope well and tabs crash a lot.

Linux I/O buffering can have a similar effect when copying large files, on some old-ish hardware, especially using the CFQ scheduler. I never understood what was happening with that, or why it wasn't considered a serious bug.

*(I've written code like that, for some tasks it's quite liberating, no I/O to worry about.)

This is obviously a complex subject. Things are running so much better after I installed the new RAM cards. This thread has engendered much interest. I have a couple of other questions which are unrelated to RAM and will form new threads.

I would just like to thank all users who have contributed to this thread.

Best wishes.