When I call $zgetjpi("","cputim"), I just get back the same number despite repeated calls. Is this expected?

Thanks
Kevin T

yottadb>

yottadb>w $zgetjpi("","cputim")
77
yottadb>w $zgetjpi("","cputim")
77
yottadb>w $zgetjpi("","cputim")
77
yottadb>w $zgetjpi("","cputim")
77
yottadb>w $zver
GT.M V6.3-008 Linux x86_64
yottadb>

I'll add that I was initially looking at GT.M manual, here:
http://tinco.pair.com/bhaskar/gtm/doc/books/pg/UNIX_manual/ch07s36.html

I also see a YottaDB entry here:
https://docs.yottadb.com/ProgrammersGuide/functions.html#zgetjpi
It mentions that there may be a linux-level privileges issues for some information. Could this be the problem?

Kevin

On Sunday, August 7, 2022 at 8:33:40 PM UTC-4, kdtop wrote:
> When I call $zgetjpi("","cputim"), I just get back the same number despite repeated calls. Is this expected?
>
> Thanks
> Kevin T
>
>
> yottadb>
>
> yottadb>w $zgetjpi("","cputim")
> 77
> yottadb>w $zgetjpi("","cputim")
> 77
> yottadb>w $zgetjpi("","cputim")
> 77
> yottadb>w $zgetjpi("","cputim")
> 77
> yottadb>w $zver
> GT.M V6.3-008 Linux x86_64
> yottadb>

More info:

yottadb>w $ZYRELEASE
YottaDB r1.30 Linux x86_64
yottadb>

Kevin

On Sunday, August 7, 2022 at 8:37:04 PM UTC-4, kdtop wrote:
> I'll add that I was initially looking at GT.M manual, here:
> http://tinco.pair.com/bhaskar/gtm/doc/books/pg/UNIX_manual/ch07s36.html
>
> I also see a YottaDB entry here:
> https://docs.yottadb.com/ProgrammersGuide/functions.html#zgetjpi
> It mentions that there may be a linux-level privileges issues for some information. Could this be the problem?
>
> Kevin
> On Sunday, August 7, 2022 at 8:33:40 PM UTC-4, kdtop wrote:
> > When I call $zgetjpi("","cputim"), I just get back the same number despite repeated calls. Is this expected?
> >
> > Thanks
> > Kevin T
> >
> >
> > yottadb>
> >
> > yottadb>w $zgetjpi("","cputim")
> > 77
> > yottadb>w $zgetjpi("","cputim")
> > 77
> > yottadb>w $zgetjpi("","cputim")
> > 77
> > yottadb>w $zgetjpi("","cputim")
> > 77
> > yottadb>w $zver
> > GT.M V6.3-008 Linux x86_64
> > yottadb>

On Sunday, August 7, 2022 at 8:41:06 PM UTC-4, kdtop wrote:
> More info:
>
> yottadb>w $ZYRELEASE
> YottaDB r1.30 Linux x86_64
> yottadb>
>
> Kevin
> On Sunday, August 7, 2022 at 8:37:04 PM UTC-4, kdtop wrote:
> > I'll add that I was initially looking at GT.M manual, here:
> > http://tinco.pair.com/bhaskar/gtm/doc/books/pg/UNIX_manual/ch07s36.html
> >
> > I also see a YottaDB entry here:
> > https://docs.yottadb.com/ProgrammersGuide/functions.html#zgetjpi
> > It mentions that there may be a linux-level privileges issues for some information. Could this be the problem?
> >
> > Kevin
> > On Sunday, August 7, 2022 at 8:33:40 PM UTC-4, kdtop wrote:
> > > When I call $zgetjpi("","cputim"), I just get back the same number despite repeated calls. Is this expected?
> > >
> > > Thanks
> > > Kevin T
> > >
> > >
> > > yottadb>
> > >
> > > yottadb>w $zgetjpi("","cputim")
> > > 77
> > > yottadb>w $zgetjpi("","cputim")
> > > 77
> > > yottadb>w $zgetjpi("","cputim")
> > > 77
> > > yottadb>w $zgetjpi("","cputim")
> > > 77
> > > yottadb>w $zver
> > > GT.M V6.3-008 Linux x86_64
> > > yottadb>
Kevin,

I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.

Mark

So we have one system not working and other system that does work.

Anyone have ideas of how I can test this?

Thanks
Kevin

On Monday, August 8, 2022 at 11:47:45 AM UTC-4, OldMster wrote:
> On Sunday, August 7, 2022 at 8:41:06 PM UTC-4, kdtop wrote:
> > More info:
> >
> > yottadb>w $ZYRELEASE
> > YottaDB r1.30 Linux x86_64
> > yottadb>
> >
> > Kevin
> > On Sunday, August 7, 2022 at 8:37:04 PM UTC-4, kdtop wrote:
> > > I'll add that I was initially looking at GT.M manual, here:
> > > http://tinco.pair.com/bhaskar/gtm/doc/books/pg/UNIX_manual/ch07s36.html
> > >
> > > I also see a YottaDB entry here:
> > > https://docs.yottadb.com/ProgrammersGuide/functions.html#zgetjpi
> > > It mentions that there may be a linux-level privileges issues for some information. Could this be the problem?
> > >
> > > Kevin
> > > On Sunday, August 7, 2022 at 8:33:40 PM UTC-4, kdtop wrote:
> > > > When I call $zgetjpi("","cputim"), I just get back the same number despite repeated calls. Is this expected?
> > > >
> > > > Thanks
> > > > Kevin T
> > > >
> > > >
> > > > yottadb>
> > > >
> > > > yottadb>w $zgetjpi("","cputim")
> > > > 77
> > > > yottadb>w $zgetjpi("","cputim")
> > > > 77
> > > > yottadb>w $zgetjpi("","cputim")
> > > > 77
> > > > yottadb>w $zgetjpi("","cputim")
> > > > 77
> > > > yottadb>w $zver
> > > > GT.M V6.3-008 Linux x86_64
> > > > yottadb>
> Kevin,
>
> I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.
>
> Mark

On Monday, August 8, 2022 at 5:37:52 PM UTC-4, kdtop wrote:
> So we have one system not working and other system that does work.
>
> Anyone have ideas of how I can test this?
>
> Thanks
> Kevin
> On Monday, August 8, 2022 at 11:47:45 AM UTC-4, OldMster wrote:
> > On Sunday, August 7, 2022 at 8:41:06 PM UTC-4, kdtop wrote:
> > > More info:
> > >
> > > yottadb>w $ZYRELEASE
> > > YottaDB r1.30 Linux x86_64
> > > yottadb>
> > >
> > > Kevin
> > > On Sunday, August 7, 2022 at 8:37:04 PM UTC-4, kdtop wrote:
> > > > I'll add that I was initially looking at GT.M manual, here:
> > > > http://tinco.pair.com/bhaskar/gtm/doc/books/pg/UNIX_manual/ch07s36.html
> > > >
> > > > I also see a YottaDB entry here:
> > > > https://docs.yottadb.com/ProgrammersGuide/functions.html#zgetjpi
> > > > It mentions that there may be a linux-level privileges issues for some information. Could this be the problem?
> > > >
> > > > Kevin
> > > > On Sunday, August 7, 2022 at 8:33:40 PM UTC-4, kdtop wrote:
> > > > > When I call $zgetjpi("","cputim"), I just get back the same number despite repeated calls. Is this expected?
> > > > >
> > > > > Thanks
> > > > > Kevin T
> > > > >
> > > > >
> > > > > yottadb>
> > > > >
> > > > > yottadb>w $zgetjpi("","cputim")
> > > > > 77
> > > > > yottadb>w $zgetjpi("","cputim")
> > > > > 77
> > > > > yottadb>w $zgetjpi("","cputim")
> > > > > 77
> > > > > yottadb>w $zgetjpi("","cputim")
> > > > > 77
> > > > > yottadb>w $zver
> > > > > GT.M V6.3-008 Linux x86_64
> > > > > yottadb>
> > Kevin,
> >
> > I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.
> >
> > Mark
Did you try doing a lot of 'work' and then rechecking $zgetjpi? It takes a lot of command line 'work' before it will increment. I ran several 1000 iteration loops doing exponentiation and string manipulation, and only got it up to 2 (which is .02 seconds of cpu time).
Linux is hard to do this with because so much 'stuff' creates a new process..

Mark

So are you saying that this is a measurement of actual CPU working time? I..e. if the mumps environment hasn't had to really do anything, but was just sitting there for 1 minute, that the CPU time wouldn't increase significantly?

OK, then I guess it must be working.

Thanks
Kevin

> > > I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.
> > >
> > > Mark
> Did you try doing a lot of 'work' and then rechecking $zgetjpi? It takes a lot of command line 'work' before it will increment. I ran several 1000 iteration loops doing exponentiation and string manipulation, and only got it up to 2 (which is .02 seconds of cpu time).
> Linux is hard to do this with because so much 'stuff' creates a new process.
>
> Mark

On Monday, August 8, 2022 at 6:11:48 PM UTC-4, kdtop wrote:
> So are you saying that this is a measurement of actual CPU working time? I.e. if the mumps environment hasn't had to really do anything, but was just sitting there for 1 minute, that the CPU time wouldn't increase significantly?
>
> OK, then I guess it must be working.
>
> Thanks
> Kevin
> > > > I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.
> > > >
> > > > Mark
> > Did you try doing a lot of 'work' and then rechecking $zgetjpi? It takes a lot of command line 'work' before it will increment. I ran several 1000 iteration loops doing exponentiation and string manipulation, and only got it up to 2 (which is .02 seconds of cpu time).
> > Linux is hard to do this with because so much 'stuff' creates a new process.
> >
> > Mark

Correct. It is a measurement of CPU Time for the process # passed in the first parameter. If the first parameter is blank, then it is for the current process. It is not a measurement of the entire machine CPU time, from the M programmers guide:

"Returns job or process information of the specified process. The format for the $ZGETJPI function is:"

The privilege issue discussed is if you are trying to retrieve information about process other than the current process - you might not have privileges to get that information for another process.

Mark

On Tuesday, August 9, 2022 at 11:47:47 AM UTC-4, OldMster wrote:
> On Monday, August 8, 2022 at 6:11:48 PM UTC-4, kdtop wrote:
> > So are you saying that this is a measurement of actual CPU working time? I.e. if the mumps environment hasn't had to really do anything, but was just sitting there for 1 minute, that the CPU time wouldn't increase significantly?
> >
> > OK, then I guess it must be working.
> >
> > Thanks
> > Kevin
> > > > > I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor.. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.
> > > > >
> > > > > Mark
> > > Did you try doing a lot of 'work' and then rechecking $zgetjpi? It takes a lot of command line 'work' before it will increment. I ran several 1000 iteration loops doing exponentiation and string manipulation, and only got it up to 2 (which is .02 seconds of cpu time).
> > > Linux is hard to do this with because so much 'stuff' creates a new process.
> > >
> > > Mark
> Correct. It is a measurement of CPU Time for the process # passed in the first parameter. If the first parameter is blank, then it is for the current process. It is not a measurement of the entire machine CPU time, from the M programmers guide:
>
> "Returns job or process information of the specified process. The format for the $ZGETJPI function is:"
>
> The privilege issue discussed is if you are trying to retrieve information about process other than the current process - you might not have privileges to get that information for another process.
>
> Mark
Kevin,

I've done some more playing with this, and it appears $zgetjpi always returns the cpu time for the current process, regardless of what the first parameter is. I'm going to submit an issue for it on GitLab.

Mark

On Tuesday, August 9, 2022 at 4:52:53 PM UTC-4, OldMster wrote:
> On Tuesday, August 9, 2022 at 11:47:47 AM UTC-4, OldMster wrote:
> > On Monday, August 8, 2022 at 6:11:48 PM UTC-4, kdtop wrote:
> > > So are you saying that this is a measurement of actual CPU working time? I.e. if the mumps environment hasn't had to really do anything, but was just sitting there for 1 minute, that the CPU time wouldn't increase significantly?
> > >
> > > OK, then I guess it must be working.
> > >
> > > Thanks
> > > Kevin
> > > > > > I tried it on v 1.28 of YottaDB. The reported value is in hundredths of a second, which is actually a lot of CPU time for a modern processor. I was able to get it to increment by doing some actual work, although after several loop squaring the value of x until it reached a really big number then resetting it to 2, then a 1000 iteration loop taking a string of all Alpha characters, uppercasing it is an odd iteration, and lowercasing it if it is an even iteration, I only got the value of $zgetjpi("","cputim") up to 2.
> > > > > >
> > > > > > Mark
> > > > Did you try doing a lot of 'work' and then rechecking $zgetjpi? It takes a lot of command line 'work' before it will increment. I ran several 1000 iteration loops doing exponentiation and string manipulation, and only got it up to 2 (which is .02 seconds of cpu time).
> > > > Linux is hard to do this with because so much 'stuff' creates a new process.
> > > >
> > > > Mark
> > Correct. It is a measurement of CPU Time for the process # passed in the first parameter. If the first parameter is blank, then it is for the current process. It is not a measurement of the entire machine CPU time, from the M programmers guide:
> >
> > "Returns job or process information of the specified process. The format for the $ZGETJPI function is:"
> >
> > The privilege issue discussed is if you are trying to retrieve information about process other than the current process - you might not have privileges to get that information for another process.
> >
> > Mark
> Kevin,
>
> I've done some more playing with this, and it appears $zgetjpi always returns the cpu time for the current process, regardless of what the first parameter is. I'm going to submit an issue for it on GitLab.
>
> Mark

From https://docs.yottadb.com/ProgrammersGuide/functions.html#zgetjpi

Note that the $ZGETJPI() retrieves process time measurements (CPUTIM, CSTIME, CUTIME, STIME, and UTIME) only of the current process ($JOB). The “child” process time includes ZSYSTEM and PIPE device sub-processes (only after the PIPE CLOSEs), but excludes processes created by JOB commands.

https://gitlab.com/YottaDB/DB/YDB/-/issues/908 has been created suggesting adding the functionality as an improvement to YottaDB.

Regards
– Bhaskar