CATS : Documention .. Vs .. ERROR

For discussion of how to use the Cycles Research Institute' own Cycles Analysis & Time Series software.

CATS : Documention .. Vs .. ERROR

Postby Steve » 04 Nov 2010, 01:00

Hi Ray,

Thank you for indicating to me that The CATS S/W can be Terminated via Ctrl-C .

Below, I Specifically indicate Each and All
Commands I issued to the CATS S/W

The Pro-log .... that led up to the CATS S/W Crash

a) I was Looking at http://img253.imageshack.us/img253/5614 ... ct2010.jpg
and
b) I arbitrarily ( subjectively ) Selected to use the 70 Cycle Lengths shown in
Columns B and J in the Picture =
http://img192.imageshack.us/img192/7606 ... 4oct10.jpg
[ which is a subset of the result of The CATS' S command ]

c) Corresponding, to the 70 Cycle Lengths indicated above, I modified the cats.ini file
to ... that depicted in the Picture = http://img560.imageshack.us/img560/8355/catsdayini.jpg
{ so that the UscK Command would Effect the 70 Usc Commands corresponding
to the above Subjectively Selected Cycle Lengths }

The Event ... I issued the following Commands to the CATS S/W ..

1) day
2) F902 #
3) L1
4) UscK ................................. It took near 50 minutes for My Machine to Compete that Command
5) ? .............................................. Which resulted with 142 Items on the STACK
6) L1
7) R140

The Resultant ERROR Message
issued by that CATS S/W
was
" The Stack is : 140 by 15647
error calling C:\PROGRAM FILES\CATS\catsreg.exe : -1
ABORTING Command
"

So,
What do I need to do so that The CATS S/W does not CRASH( as
described in the above sequence) ?

Regards,
Steve

PS : CATS default Set-Up was only Modified via the content of
http://img560.imageshack.us/img560/8355/catsdayini.jpg

..
Steve
 
Posts: 16
Joined: 13 Oct 2010, 21:25

Re: CATS : Documention .. Vs .. ERROR

Postby RayTomes » 04 Nov 2010, 09:45

Steve wrote:Hi Steve

1) day
2) F902 #
3) L1
4) UscK ................................. It took near 50 minutes for My Machine to Compete that Command
5) ? .............................................. Which resulted with 142 Items on the STACK
6) L1
7) R140

The Resultant ERROR Message
issued by that CATS S/W
was
" The Stack is : 140 by 15647
error calling C:\PROGRAM FILES\CATS\catsreg.exe : -1
ABORTING Command
"


Without having the data in my computer I cannot replicate.

I would recommend putting a space between commands such as Usc and K (perhaps you did). The U command will take scK as the macro name otherwise.

The K command should have a number on it that is greater than zero. It may be that this is not tested for in CATS. I will note it as a possible bug.

The sequence of commands does not make sense to me.
F902 # will load a series and take logs
L1 will duplicate that line making a second line
Usc does not operate on a time series. It requires a single constant period before it and makes two series which are a sine and a cosine wave of that period (for use in regressions). e.g. 5.9 Usc

Hope that this helps.
Ray
User avatar
RayTomes
 
Posts: 128
Joined: 02 Aug 2010, 23:24
Location: Auckland, New Zealand

Re: CATS : Documention .. Vs .. ERROR

Postby Steve » 04 Nov 2010, 14:12

The cats.ini file defines Macros ... that .ini file is depicted in
the Picture http://img560.imageshack.us/img560/8355/catsdayini.jpg

Please look at the above Picture .


Specifically ... the macro named scA is defined as
OscA 9198.23535 Usc 48865.62500 Usc 2.51563 Usc 2.92773 Usc 4.92199 Usc 4.30185 Usc 3.66468 Usc

Using such marco effectively executes 7 sc commands ...

The macro named scK is defined as
OscK UscA UscB UscC UscD UscE UscF UscG UscH UscI UscJ

Using such macro effectively executes 70 sc commands

From the cats-manual.pdf file ( that came with the cats s/w ) it states under
Section Heading " LOOPS AND MACROS " ... under the blurb about Uname
specifically " Macro names may consist of letters only " .

F902 # ....... load Log of Series ..
L1 ....... duplicate the above ( the Stack now contains 2 Items )
UscK ....... execute marco scK, effectively executing 70 sc commands
and as result there are then 142 items on the Stack
L1 ....... load the Log of the Data Series to Position 143 of the Stack
R140 ....... run the Regression

That Regression command CRASHED the S/W ... and gave error message
" " The Stack is : 140 by 15647
error calling C:\PROGRAM FILES\CATS\catsreg.exe : -1
ABORTING Command "
"

The culprit appears to have something to do with
the catsreg.exe process ( whatever its purpose ) .

I hope the above better explains the Sequence of Events.

Regards,
Steve
Steve
 
Posts: 16
Joined: 13 Oct 2010, 21:25

Re: CATS : Documention .. Vs .. ERROR

Postby RayTomes » 05 Nov 2010, 03:23

Hi Steve

My apology. I have a macro called sc which makes sine and cosine waves and I thought that your UscK was Usc fllowed by K.
That invalidates my remarks.

OK, so you are creating 140 waves which are 70 pairs of cycles. Got it now.

The problem is probably that you are set to a stack size of 100. You need to change your cats.ini file in the day folder. The default file contains:

18000101,20200101,1
100


and you need to increase that 100 on the second line to something over 140. Suggest maybe 200 from teh look of what you are doing. :-)

Best of luck
Ray
User avatar
RayTomes
 
Posts: 128
Joined: 02 Aug 2010, 23:24
Location: Auckland, New Zealand

Re: CATS : Documention .. Vs .. ERROR

Postby Steve » 05 Nov 2010, 07:01

Hi Ray,

Thanks for taking the time to respond.

Please Note that I have already modified The STACK Length Limit to 400
Please re-look at the picture of the cats.ini file ( in the day folder ) I used ...
http://img560.imageshack.us/img560/8355/catsdayini.jpg

Hmmm, Now I wonder if there may be a Limit on the number of Days Between Dates
as per the first line of that cats.ini file ?

What is the meaning of the reported ERROR Message
" The Stack is : 140 by 15647
error calling C:\PROGRAM FILES\CATS\[color=#FF4000]catsreg.exe
: -1
ABORTING Command[/color] "

NOTE : that it is known that there are 143 Items on The Stack
before the F140 Command was Issued

What does the number 15647 in the reported Error Message
" The Stack is : 140 by [color=#FF4000]15647 ...[/color] "
refer to and what is its associated Limit and where is that Limit Defined ?

Regards,
Steve.
Steve
 
Posts: 16
Joined: 13 Oct 2010, 21:25

Re: CATS : Documention .. Vs .. ERROR

Postby Steve » 05 Nov 2010, 07:17

Steve wrote:Hi Ray,

Thanks for taking the time to respond.

Please Note that I have already modified The STACK Length Limit to 400
Please re-look at the picture of the cats.ini file ( in the day folder ) I used ...
http://img560.imageshack.us/img560/8355/catsdayini.jpg

Hmmm, Now I wonder if there may be a Limit on the number of Days Between Dates
as per the first line of that cats.ini file ?

What is the meaning of the reported ERROR Message
" The Stack is : 140 by 15647
error calling C:\PROGRAM FILES\CATS\[color=#FF4000]catsreg.exe
: -1
ABORTING Command[/color] "

NOTE : that it is known that there are 143 Items on The Stack
before the F140 Command was Issued

What does the number 15647 in the reported Error Message
" The Stack is : 140 by 15647 ... "
refer to and what is its associated Limit and where is that Limit Defined ?

Regards,
Steve.
Steve
 
Posts: 16
Joined: 13 Oct 2010, 21:25

CATS S/W & catsreg.exe MALFUNCTION

Postby Steve » 05 Nov 2010, 23:24

Hi Ray,

I am aware that you are using a version of the CATS S/W
which is other than that which I have access to.

Now with specific regard to all that I have already written ( above ) in this thread ...

I did try again
and
the result remains the same
the Regression Command FAILS

Does the catsreg.exe Executable perform the Regressions ?
If so then it appears to me to have " an opportunity for enhancement ", because
it does not work under conditions I have witnessed.

Please refresh your memory with that which I have previously stated in this Thread.

Then

SPECIFICALLY what I did on my re-try was

a) modify the cats.ini file to Limit STACK Length to 250
b) modify the INPUT Data Series to have Identically 15651 Data Points ( i.e. :
Updated for Current Values, not contained in Previous Data Set )

c) issued the following Exact Sequence of Commands to the CATS S/W ....
day
F902 #
L1
UscK
?
.... which showed that there were then 142 items on The STACK
L1 .... put the initial Series onto STACK Top
R140 ..... which results with ERROR

The ERROR is Identified in the DOS Application ( CATS ) Window with the message =
" The Stack is : 140 by 15651
error calling C:\PROGRAM FILES\CATS\catsreg.exe : -1
ABORTING Command
"

This ERROR Message indicates that the above F140 Command is the culprit and
further that F140 Command FAILED because of some malfunction in the catsreg.exe Executable

as Evidenced with the following 2 Pictures taken of the ERROR MESSAGE
1) http://img574.imageshack.us/img574/7061 ... errora.jpg
2) http://img829.imageshack.us/img829/5516 ... errorb.jpg

Symptoms:
a) 70 Cycle Lengths Issued to sc Commands generated 140 Items on the STACK
b) The Data Series had Length 15651
c) The ERROR message stated " The Stack is : 140 by 15651 "
which Might Imply a STACK Size of 2191140 ( with no units Defined that I am aware of )

So,
Ray
what are the undocumented Limits that seem to exist on
the catsreg.exe Executable ?

Regards,
Steve
Steve
 
Posts: 16
Joined: 13 Oct 2010, 21:25

Re: CATS : Documention .. Vs .. ERROR

Postby Steve » 06 Nov 2010, 02:23

Where in my above post I made the statement
" This ERROR Message indicates that the above F140 Command is the culprit and
further that F140 Command FAILED because of some malfunction in the catsreg.exe Executable
"
Please replace that statement with
" This ERROR Message indicates that the above R140 Command is the culprit and
further that R140 Command FAILED because of some malfunction in the catsreg.exe Executable
"
Thank you ..
Steve
 
Posts: 16
Joined: 13 Oct 2010, 21:25

Re: CATS S/W & catsreg.exe MALFUNCTION

Postby RayTomes » 09 Nov 2010, 13:18

Hi Steve

The regression, factor analysis and spectral analysis are all performed by external programs to CATS.

There probably are some limits on the arrays, but they are intended to be very large. However these limits may not be fully tested. So it is possible that you hit one.

I need to get my IT man to look at the code because I can't find it. Too many files arranged in weird way. We have certainly tested for E100 and R100, but R142 may be failing in some way. I will get back to you when we find the answer. Logging this as a CATS bug.

Regards
Ray
User avatar
RayTomes
 
Posts: 128
Joined: 02 Aug 2010, 23:24
Location: Auckland, New Zealand

Re: CATS : Documention .. Vs .. ERROR

Postby RayTomes » 09 Nov 2010, 23:02

Hi Steve

I can confirm that the programs for Factor Analysis and Regression both have array sizes limited to n=100. Sorry about that. These should definitely be increased in the next version.

Regards
Ray
User avatar
RayTomes
 
Posts: 128
Joined: 02 Aug 2010, 23:24
Location: Auckland, New Zealand

Next

Return to CATS Software

Who is online

Users browsing this forum: No registered users and 1 guest

cron