[Info-vax] Process memory settings

[Syltrem]

"Ken Fairfield" <ken.fairfield at gmail.com> wrote in message 
news:0938daf2-c7df-4b2e-a14b-1f0b539ad10b at i8g2000pro.googlegroups.com...
On Aug 13, 10:30 am, "Syltrem" <syltremz... at videotron.ca> wrote:
[...]
> I agree with all you're saying and yet, this is what happens.
> If I lower WSEXTENT the process will use its PGFLQUO until the virtual
> memory reaches WSMAX.
> It's not supposed to be like this, but that's what I see.
>
> The process has 5,000,000 PGFLQUO and never gets to use but a fraction of
> it.
> [...]
>
> OK, but would you please post the amount of physical memory
> on the AS4100 on which this is happening, as well as the page
> file size(s)?  Just a SHOW MEMORY would do.  Noting that
> a pgflquota of 5M pagelets is about 2.4 GB, you need to have
> that much (virtual) memory available...
>

I installed the latest version of the compiler I could find V2.9-1453

$ sh me/ph/fil
              System Memory Resources on 13-AUG-2009 14:19:30.05

Physical Memory Usage (bytes):     Total        Free      In Use    Modified
  Main Memory             (GB)      2.50        0.02        2.09        0.37

Swap File Usage (bytes):                       Index        Free        Size
  DISK$ALPHAVMS:[KRONOS]SWAPFILE.SYS;4
                                    (MB)           1       57.18       57.18

Paging File Usage (bytes):                     Index        Free        Size
  DISK$ALPHAVMS:[KRONOS]PAGEFILE2.SYS;1
                                    (GB)         252        1.16        1.25
  DISK$ALPHAVMS:[KRONOS]PAGEFILE1.SYS;2
                                    (GB)         253        1.04        1.25
  DISK$ALPHAVMS:[KRONOS]PAGEFILE.SYS;2
                                    (GB)         254        1.04        1.25

  Total size of all paging files:   (GB)                                3.75
  Total committed paging file usage:(GB)                                2.34

Of the physical memory in use, 475.75 MB are permanently allocated to 
OpenVMS.

>
> The other thing is that SHOW PROC/QUOTA does *not*
> list "Peak working set size:" or "Peak virtual size:".  Are
> these coming from the compiler?

$ SHOW PROC/ACCOUNTING does :

$ sh proc/acco

13-AUG-2009 14:22:10.28   User: ITEST         Process ID:   2336A421
                          Node: KS           Process name: "ITEST"

Accounting information:
 Buffered I/O count:       699  Peak working set size:     872448
 Direct I/O count:         492  Peak virtual size:        1964832
 Page faults:           474682  Mounted volumes:                0
 Images activated:          14
 Elapsed CPU time:          0 00:06:52.16
 Connect time:              0 00:09:38.52


> I'd like to know what you see, from a separate process,
> when you do Show Proc/Quota/Id=xxxxxxxx on the
> process doing the compile.  You should see the value
> for "Paging file quota:" decreasing, right?
>
>    Puzzled, Ken


 PROCESS MONITOR
  Process Name ITEST Process Id 2336A421
  Images Invoked so far 13               Username ITEST

  Current image $1$DGA10:[SYS0.SYSCOMMON.][SYSEXE]COBOL.EXE;1

  Time 13-AUG-2009 14:18:42.43 Process State COM
  Page faults 295146                  Pages Available 2147483647
--------------------------------------------------------------------------------
PROCESS RESOURCE QUOTAS
 Resource           Unused  Quota %Left | Resource           Unused  Quota 
%Left
                                        |
 CPU limit           28995      0    0% | Direct I/O           1000   1000 
100%
 Byte limit         9992961000000   99% | Buff I/O             1000   1000 
100%
 Timer Q              2000   2000  100% | File Lim             2045   2048 
99%
 Page file         36526565000000   73% | Sub-proc's             20     20 
100%
 Enqueue              9999  10000   99% | AST limit            2046   2048 
99%
 WS Current              0 872448    0% |
                                        |
--------------------------------------------------------------------------------


With the new compiler 2.9 Peak virtual size goes up to 1964832 and is no 
longer limited to 847968, but it still fails in the end even when I haven`t 
used up my page file quota.

Now I wonder what this new limit really is. Before it was real close to be = 
wsmax, not anymore.

In the above display, with the old compiler, when WS Current %Left reached 0 
it would immediately terminate.
Now it uses up some of the page file quota before doing so.
I did not change any parameter since my last test.

Puzzled Syltrem