[Oberon] Re (2): OLR and data storage.

[Jan de Kruyf]

Right then,

I notice this in Peter(E)'s kernel.log:

In the case where there is a wrong file listed in the Directory listing
 either the next one
(most of the time) or the one after the next one (not so often) is a *.Tool
file.
Unfortunately I am to lazy tonight to work out the statistical significance
of that.
It might or might not be significant.

But a good test might be to run Oberon on one processor only. At the moment
it will run on
all processors as the scheduler sees fit.

So Peter could you do this in a terminal:

jan at janDell:~/Oberon-olr/olr$ ./olx
(to start oberon)

in another terminal:
jan at janDell:~$ ps -A|grep olx
(to find the pid of the process)
this gives something like:
 9664 pts/0    00:00:00 olx
where 9664 is the pid for this process on my machine, it will be different
every time.
Then:
jan at janDell:~$ taskset -p 9664
should give something like
pid 9664's current affinity mask: f
it means that the process has equal affinity to all cpu's.

Close oberon with "System.Quit"
(This is important since closing the viewer does not kill the process and
you are left with a zombie.)

So this was our warm-up session.
Now for real:
Restart oberon with affinity to the 2nd cpu:

jan at janDell:~/Oberon-olr/olr$ taskset 2 ./olx

Now you can repeat from "ps -A|grep olx" in another terminal to confirm
that you get something like:

pid 9640's current affinity mask: 2
(this is a binary mask, so the 2nd bit is set; meaning the process likes
the 2nd processor best)

So now we are all set to repeat the "System.Directory *.Tool"  test and see
if there is any difference at all.

Could you please do this and let us know the results?
I will try to make some time in the weekend for messing around with
ImportGraph.Mod,
where I has the lost characters in the viewer. I did not do that yet.

Cheers,

j.




On Fri, Oct 16, 2015 at 8:28 PM, Jan de Kruyf <jan.de.kruyf at gmail.com>
wrote:

> Peter(E) and Peter(M).
>
>
> Here are the results from my machine for the System.Directory *.Tool
>
> Backup.Tool
> BootLinker.Tool
> Compiler.Tool
> Configuration.Tool
> ET.Tool
> Edit.Tool
> FAT.Tool
> FileSystem.Tool
> Files.Tool
> GZip.Tool
> Greek.Tool
> IFS.Tool
> Math.Tool
> Miscellaneous.Tool
> NetSystem.Tool
> OLR.Gadgets.Tool
> OLR.Make.Tool
> PPP.Tool
> Partitions.Tool
> Rel.DOS.Tool
> Rel.Network.Tool
> Script.Tool
> System.Tool
> Tools.Tool
> Usb.Tool
> Zip.Tool
>
> And here is the same result but in Linux:
>
> jan at janDell:~/Oberon-olr/olr$ ls -1 *.Tool
> Backup.Tool
> BootLinker.Tool
> Compiler.Tool
> Configuration.Tool
> Edit.Tool
> ET.Tool
> FAT.Tool
> Files.Tool
> FileSystem.Tool
> Greek.Tool
> GZip.Tool
> IFS.Tool
> Math.Tool
> Miscellaneous.Tool
> NetSystem.Tool
> OLR.Gadgets.Tool
> OLR.Make.Tool
> Partitions.Tool
> PPP.Tool
> Rel.DOS.Tool
> Rel.Network.Tool
> Script.Tool
> System.Tool
> Tools.Tool
> Usb.Tool
> Zip.Tool
> jan at janDell:~/Oberon-olr/olr$
>
> The only difference is the sorting method, but I can live with that. Its
> probably an artifarct from NW's String comparison idea.
>
> So I think we might conclude that sofar the problem exibits differently on
> a 64 bit machine.
> Since we now did a System.Directory with exactly the same FileDir.Mod
> software.
> But it would be nice if Peter Matthias repeated this test also.
>
> And perhaps you might tell us your machine details Peter M?
>
> I will try to look through the Kernel.Log a bit later.
>
> Cheers,
>
> j.
>
>
>
> On Thu, Oct 15, 2015 at 10:36 PM, Peter Easthope <peter at easthope.ca>
> wrote:
>
>> On Thu, October 15, 2015 3:28 pm, Jan de Kruyf wrote:
>> > And if you do *.Mod?
>>
>> ??
>>
>> > I will try tomorrow or on the weekend. Bedtime now.
>>
>> Such a problem can be a very long ordeal.
>>
>> > You saw my struggle with the Ulm package I take it.
>>
>> Yes but not in detail.  I've never explored beyond Native
>> Oberon and UnixAos.
>>
>> If you have another test for OLR, System.Directory, let me know,
>>                           ... Peter E.
>>
>> --
>> Telephone 1 360 639 0202 or +13606390202 or 12506293757.
>> Bcc: peter at easthope.ca  http://easthope.ca/Shop.html
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.inf.ethz.ch/pipermail/oberon/attachments/20151016/400f6fdd/attachment.html>