Menu

Main Menu
Talk Get Daily Search

Member's Online

    User Name
    Password

    Comparing the two CPU/Memory plugins

    Reply
    Linear2202 | # 1 | 2006-04-26, 13:11 | Report

    I have them both loaded into my tablet, and I've noticed something somewhat interesting.

    The first one shows only one bar in the memory used section.

    The second one shows that upon a cold boot, that 40% of memory is used up. This is with no applications loaded.

    I would expect to see two bars on the first cpu/memory load plugin if 40% of memory were used up on a cold boot.

    Has anyone else tried these two plugins and if so, where similar results discovered?

    Thanks..

    Edit | Forward | Quote | Quick Reply | Thanks

     
    gnuite | # 2 | 2006-04-26, 18:16 | Report

    That depends on what you consider to be "free memory." It's not as simple as it sounds. For example, here is a selected subset of the output from the /proc/meminfo virtual file on my desktop machine (the fields are similar to what is found on the Nokia 770's version - the numbers are, of course, way off):

    Code:
    MemTotal:      1035424 kB
    MemFree:        446228 kB
    Buffers:         34748 kB
    Cached:         147728 kB
    SwapCached:     116792 kB
    SwapTotal:     2040212 kB
    SwapFree:      1896712 kB
    Mapped:         359260 kB
    Slab:            24916 kB
    CommitLimit:   2557924 kB
    Committed_AS:   784532 kB
    The "first" plugin that you referenced uses Committed_AS as the sole indicator of memory. That typically is a worst-case scenario and is not practically accurate; in fact, under load that number often exceeds the number indicated by MemTotal. That plugin also assumes a fixed "MemTotal" and thus does not take Swap into account.

    I don't know what the "second" plugin you mentioned uses, but my own custom version of the "first" plugin (load-plugin-run) uses a combination of MemTotal, MemFree, Buffers, Cached, SwapCached, SwapTotal, SwapFree, and Slab to determine a "practical" memory-load indicator. Under the testing I gave it, its "full" indication much more accurately reflected a low-memory situation than the first original plugin did (which often indicated "full" much earlier than necessary).

    Edit | Forward | Quote | Quick Reply | Thanks

     
    Linear2202 | # 3 | 2006-04-27, 01:28 | Report

    OK, thanks for the explanation. I was wondering.

    I don't have the swap file extended to the memory disk, so I run the unit as is memory wise.

    Edit | Forward | Quote | Quick Reply | Thanks

     
vBulletin® Version 3.8.8
Normal Logout