18 Mar
2015
18 Mar
'15
3:53 p.m.
Hi Alexandre,
On Wed, Mar 18, 2015 at 2:01 PM, Alexandre Bergel <alexandre.bergel(a)me.com>
wrote:
Hi!
Making things scalable is indeed a pain for the user. Currently, using
RTCache remove all the interactions from the elements. This means we cannot
have this cache all the time. At least for now. Also, it may create
oddities with labels.
thanks for the explanation.
In my case I don't need interactive features (at least for now).
You have a large data set to render, which is not the
case of everyone.
Is the current solution acceptable for now?
I can work with the current solution, now that I know of the cache.
Thanks,
Manfred
Cheers,
Alexandre
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
On Mar 18, 2015, at 6:46 AM, Manfred Kröhnert <mkroehnert42(a)googlemail.com>
wrote:
Thanks Alex,
adding the RTCache line to the script helps preventing the
unresponsiveness in the most recent Moose image.
It's only that this feels more like a workaround since this is not
remotely obvious for a beginner like me :-)
Would it be possible to let the preview tab in the inspector do something
like this by default?
Then it would work out of the box.
Thanks again,
Manfred
On Wed, Mar 18, 2015 at 12:45 AM, Alexandre Bergel <
alexandre.bergel(a)me.com> wrote:
Hi Manfred,
I think have fixed your problem.
Consider this script:
-=-=-=-=-=-=-=-=-=-=-=-=
cpuFile := '/Users/alexandrebergel/Downloads/cpu.trace' asFileReference.
cpuReader := NeoCSVReader on: (cpuFile readStream).
cpuReader separator: $|.
cpuHeader := cpuReader readHeader.
cpuData := cpuReader upToEnd.
g := RTGrapher new.
g extent: 300@200.
ds := RTDataSet new.
ds points: (cpuData collect: [ :each |
| timestamp cpuUsage |
timestamp := (each at: 1) asNumber.
cpuUsage := (each at: 4) asNumber.
Array braceWith: timestamp with: cpuUsage]).
ds x: #first.
ds y: #second.
ds connectColor: Color blue.
g add: ds.
g addDecorator: RTHorizontalTickLineDecorator new.
g addDecorator: RTVerticalTickLineDecorator new.
g build.
RTCache new cacheView: g view.
g view.
-=-=-=-=-=-=-=-=-=-=-=-=
The magic line here is "RTCache new cacheView: g view.”
I will describe this in another thread.
Cheers,
Alexandre
_______________________________________________
Moose-dev mailing list
Moose-dev(a)iam.unibe.ch
https://www.iam.unibe.ch/mailman/listinfo/moose-dev
_______________________________________________
Moose-dev mailing list
Moose-dev(a)iam.unibe.ch
https://www.iam.unibe.ch/mailman/listinfo/moose-dev
On Mar 17, 2015, at 1:47 PM, Manfred Kröhnert
<
mkroehnert42(a)googlemail.com> wrote:
Hi Alexandre,
seems like I forgot to add the link.
So here it is:
https://drive.google.com/file/d/0B-EyIKkOzbhybjdyU2dsa29EODA/view?usp=shari…
Thanks for taking a look,
Manfred
On Tue, Mar 17, 2015 at 3:24 PM, Alexandre Bergel <
alexandre.bergel(a)me.com>
wrote:
Hi Manfred,
Can I have access to the cpu.trace file? I promise to not share it :-)
Alexandre
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
> On Mar 17, 2015, at 7:14 AM, Manfred Kröhnert <
mkroehnert42(a)googlemail.com> wrote:
>
> Thanks for the response so far.
>
> Here is a link to the file I am using with the script from my first
post.
> The only thing which needs to be done is to
create a Filesystem object
for the file with
>
>
> cpuFile := 'path/to/cpu.trace' asFileReference
> cpuReader := NeoCSVReader on: (cpuFile readStream).
> cpuReader separator: $|.
> cpuHeader := cpuReader readHeader.
> cpuData := cpuReader upToEnd.
>
> g := RTGrapher new.
> g extent: 300@200.
>
> ds := RTDataSet new.
> ds points: (cpuData collect: [ :each |
> | timestamp cpuUsage |
> timestamp := (each at: 1) asNumber.
> cpuUsage := (each at: 4) asNumber.
> Array braceWith: timestamp with: cpuUsage]).
> ds x: #first.
> ds y: #second.
> ds connectColor: Color blue.
>
> g add: ds.
> g addDecorator: RTHorizontalTickLineDecorator new.
> g addDecorator: RTVerticalTickLineDecorator new.
> g build.
> g view canvas.
>
>
> The responsiveness does get back to normal when I switch from the
Canvas view
to the Raw or Meta view of the TRCanvas object but also when
the tab of the TRCanvas object gets closed.
>
> To measure the memory usage of the vm I used the following statement:
((Smalltalk vm getParameters at: 3) / 1000000) asInteger
> This is adapted from the following SO answer:
http://stackoverflow.com/questions/8556207/in-pharo-how-can-you-measure-the…
> Before executing the script the statement
shows 76 and afterwards 203.
>
> Wrapping the whole script in 'MessageTally spyOn: []' gives me the
following report:
>
https://drive.google.com/file/d/0B-EyIKkOzbhydU9PNzVjMmZ5d3c/view?usp=shari…
>
>> Thanks for your help so far,
>> Manfred
>
>
>> On Tue, Mar 17, 2015 at 7:32 AM, Tudor Girba
<tudor(a)tudorgirba.com>
wrote:
>> How much RAM is what using?
>
>> Doru
>
>> On Tue, Mar 17, 2015 at 7:09
AM, phil(a)highoctane.be <
phil(a)highoctane.be> wrote:
>> How much RAM is this using?
>
>> Phil
>
>> Le 17 mars 2015 07:03,
"Tudor Girba" <tudor(a)tudorgirba.com> a écrit :
>
>> Also, does the responsiveness
get better after you close the
visualization?
>
>> Cheers,
>> Doru
>
>> On Mon, Mar 16, 2015 at 9:03
PM, Alexandre Bergel <
alexandre.bergel(a)me.com> wrote:
>> Hi Manfred,
>
>> You may have pin-pointing a
limitation in Roassal. However, I am sure
we can handle it easily. Can I have this CSV file? to reproduce the
situation on my machine?
>
>> Resend the complete script as
well please :-)
>
>> Cheers,
>> Alexandre
>> --
>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>> Alexandre Bergel http://www.bergel.eu
>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>
>
>
>>> On Mar 16,
2015, at 3:31 PM, Manfred Kröhnert <
mkroehnert42(a)googlemail.com> wrote:
>>
>> Hi Sven,
>>
>>
>>
>> On Mon, Mar 16, 2015 at 7:14 PM, Sven Van Caekenberghe <sven(a)stfx.eu>
wrote:
>> Hi Manfred,
>>
>> I can only help you with the NeoCSV & IO part. Here is how I would do
it:
>>
>> [
>> 'manfred.csv' asFileReference writeStreamDo: [ :out |
>> (NeoCSVWriter on: (ZnBufferedWriteStream on: out)) in: [ :writer |
>> writer separator: $|; writeHeader: #(timestamp two three cpu);
fieldWriter: #raw.
>> 130000 timesRepeat: [
>> writer nextPut: { DateAndTime now asUnixTime. 999 atRandom. 999
atRandom. 999 atRandom } ] ] ].
>> ] timeToRun. "0:00:00:01.917"
>>
>> [
>> 'manfred.csv' asFileReference readStreamDo: [ :in |
>> (NeoCSVReader on: (ZnBufferedReadStream on: in)) in: [ :reader |
>> reader separator: $|; readHeader.
>> reader addIntegerField; addIgnoredField; addIgnoredField;
addIntegerField.
>> reader upToEnd ] ].
>> ] timeToRun. "0:00:00:01.064"
>>
>> I think that is acceptable for a 3Mb file.
>>
>>
>> thanks for your input.
>> This will come in handy once the files get bigger.
>>
>> But the problem is not in reading the file.
>> Even when I use 'reader upToEnd' it is really fast and all following
actions remain fluent.
>>
>> The problem I am facing is that any action I am performing _after_
drawing
the graph takes several seconds.
>> And I just noticed that switching from
the Canvas to the Raw tab of
the TRCanvas object makes the problem disappear.
>> But switching to the Canvas tab itself
takes several seconds
(probably until the graph is rendered) and every action
afterwards is
getting delayed several seconds.
>>
>> The machine I am running this image on has 8 GB RAM and a 2.8 GHz
Core i7
(mobile) processor if this information helps.
>>
>> Best,
>> Manfred
>>
>>
>>
>> HTH,
>>
>> Sven
>>
>> > On 16 Mar 2015, at 17:48, Manfred Kröhnert <
mkroehnert42(a)googlemail.com> wrote:
>> >
>> > Hi everyone,
>> >
>> > after having access to a machine running OS X 10.10.2 I started to
give Moose 5.1 a try for some initial data analysis.
>> > I used Pharo Launcher (build #24)
and downloaded the Moose image
with buildnumber 304.
>> >
>> > The task is to read a CSV log file containing around 130000
entries,
each containing a timestamp, a cpu usage number, and some other
information.
>> > Everything works as expected, but
the responsiveness of the image
drops dramatically after executing the script in
Playground with
Cmd+Shift+g.
>> > After evaluating the script it takes
several seconds to get a
response to any action such as opening the world menu or
switching tabs in
the playground.
>> >
>> > Is this because I am doing something very inefficient (most
probably)
or does this problem occur more often?
>> > Here is the script I am using:
>> >
>> >
>> > cpuReader := NeoCSVReader on: (cpuFile readStream).
>> > cpuReader separator: $|.
>> > cpuHeader := cpuReader readHeader.
>> > cpuData := cpuReader upToEnd.
>> >
>> > g := RTGrapher new.
>> > g extent: 300@200.
>> >
>> > ds := RTDataSet new.
>> > ds points: (cpuData collect: [ :each |
>> > | timestamp cpuUsage |
>> > timestamp := (each at: 1) asNumber.
>> > cpuUsage := (each at: 4) asNumber.
>> > Array braceWith: timestamp with: cpuUsage]).
>> > ds x: #first.
>> > ds y: #second.
>> > ds connectColor: Color blue.
>> >
>> > g add: ds.
>> > g addDecorator: RTHorizontalTickLineDecorator new.
>> > g addDecorator: RTVerticalTickLineDecorator new.
>> > g build.
>> > g view canvas.
>> >
>> >
>> > I also noticed that the 'Meta' representation of an object looks
very crammed since it does not use the full size available in the tab.
>>> > Should issues like this be reported somewhere else?
>>> >
>>> > Thanks for taking a look at this,
>>> > Manfred
>>> > _______________________________________________
>>> > Moose-dev mailing list
>>> > Moose-dev(a)iam.unibe.ch
>>> > https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> Moose-dev(a)iam.unibe.ch
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> Moose-dev(a)iam.unibe.ch
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>> _______________________________________________
>> Moose-dev mailing list
>> Moose-dev(a)iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>
>
>> --
>> www.tudorgirba.com
>
>> "Every thing has its own
flow"
>
>>
_______________________________________________
>> Moose-dev mailing list
>> Moose-dev(a)iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>> _______________________________________________
>> Moose-dev mailing list
>> Moose-dev(a)iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>
>
>> --
>> www.tudorgirba.com
>
>> "Every thing has its own
flow"
>
>>
_______________________________________________
>> Moose-dev mailing list
>> Moose-dev(a)iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>> _______________________________________________
>> Moose-dev mailing list
>> Moose-dev(a)iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
> _______________________________________________
> Moose-dev mailing list
> Moose-dev(a)iam.unibe.ch
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
> _______________________________________________
> Moose-dev mailing list
> Moose-dev(a)iam.unibe.ch
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
_______________________________________________
Moose-dev mailing list
Moose-dev(a)iam.unibe.ch
https://www.iam.unibe.ch/mailman/listinfo/moose-dev