Re: memory error

To: Jonathan Callahan <callahan@xxxxxxxxxxxxx>, John C Cartwright <John.C.Cartwright@xxxxxxxx>
Subject: Re: memory error
From: Joe Sirott <sirott@xxxxxxxxxxxxx>
Date: Tue, 11 Jun 2002 14:12:38 -0700
Cc: LAS Users Mailing List <las_users@xxxxxxxxxxxxxxxxxxx>
In-reply-to: <3D066172.3DA44D43@pmel.noaa.gov>
References: <3D065A8E.7020308@ngdc.noaa.gov>
Sender: owner-las_users@xxxxxxxxxxxxxxxxxxx

Hi John,

A couple of other suggestions:

Upgrade to ferret v5.40. Previous versions of Ferret had a fixed limit on the amount of memory that could be allocated for graphics. That was removed in v5.40

That's the quickest fix. However, it would also make sense to decimate the dataset before plotting it. If you are plotting a dataset in, say, a 800x600 window, there are only about 500K pixels available. The message from Ferret indicates that it want to plot 2331719 points, which is much higher resolution than the plot window and much of that higher resolution info is lost when creating the 800x600 plot.

At 01:45 PM 6/11/2002 -0700, Jonathan Callahan wrote:

John,

It sounds like you're doing something that is extremely graphics
intensive. This "program limit" error is not a memory error. You can
run the following script (extracted from your log) directly in Ferret to
see where (and perhaps why) it's crashing. The "/tmp/lasgo..." files
are not deleted if Ferret fails to generate a plot. You will want to
set up the Ferret environment variables to match those the web server is
using.

Suggestions why graphics might be intensive:

1) noise about zero with a contour line at zero
--> remove fill level at zero, remove contour line at zero

-- Jon

cancel mode verify
cancel mode interp
canc data/all
go std_initialize "bt.nc" "1" "1" "btt"
set
region/x="180":"540"/y="-90.0":"89.8333333333333"/k=1/t="08-Apr-1989":"08-Apr-1989"
set win/size=0.5
set win/asp=0.749305555555556
GO "/tmp/lasgo18572669910.jnl"

John C Cartwright wrote:
>
> Hello All,
>
> I seem to still be having memory errors trying to map a large netCDF
> dataset.
>
> I set the MEMSIZE ferret property in the las.xml as follows:
> <properties>
> <ferret>
> <memsize>127</memsize>
> </ferret>
> </properties>
>
> However, this doesn't seem to make a difference. I can see in the log
> file (attached) where ferret is getting the property, but seems to
> ignore it. I finally modified the server/Ferret_config.pl to set
> "-memsize 64" on the ferret command line arguments and that did seem to
> increase the memory. Now, however, ferret is complaining that:
>
> **ERROR: a program limit has been reached: 2331719 points to plot
> Plot buffer size is: 990000
>
> Is this something that I can configure? Is there a problem with the way
> I was originally trying to increase the memory?
>
> Thanks!
>
> -- john
>
> ------------------------------------------------------------------------
> Browser = Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020605
>
> Dumping config file
> cache : 1
> error_file : log/errors
> ferret_ps_args : -server -batch
> log_file : log/access
> output_directory : output/
> custom_include : custom
> which_ferret : /usr/local/ferret/bin/ferret
> ferret_args : -gif -server -memsize 64
> debug_file : log/GenericLAS_debug.txt
> output_alias : /las-output/
> Got XML request:
> <?xml version="1.0"?><lasRequest href="file:las2.xml"><link match="/lasdata/operations/shade" /><properties><ferret><format>shade</format><size>0.5</size><imgWidth>800</imgWidth><imgHeight>600</imgHeight></ferret><ui><varname>SST</varname><dsetname>Brown AVHRR SST</dsetname><view>xy</view><configid>1</configid></ui></properties><args><link match="/lasdata/datasets/bt/variables/btt"></link><region><range type="x" low="-180.0" high="180.0"/><range type="y" low="-90.0" high="89.8333333333333"/><point type="t" v="08-Apr-1989"/></region></args></lasRequest>
> Region: -180.0 180.0 x
> Region: -90.0 89.8333333333333 y
> Region: 08-Apr-1989 08-Apr-1989 t
> ferret prop: refmap_xhi = 540
> ferret prop: view = xy
> ferret prop: t_lo = 08-Apr-1989
> ferret prop: refmap_yhi = 89.8333333333333
> ferret prop: x_hi = 540
> ferret prop: y_hi = 89.8333333333333
> ferret prop: land_type = contour
> ferret prop: format = shade
> ferret prop: imgHeight = 600
> ferret prop: size = 0.5
> ferret prop: refmap_xlo = 180
> ferret prop array: jnl_t = (08-Apr-1989,08-Apr-1989)
> ferret prop: refmap_view = box
> ferret prop: refmap_ylo = -90.0
> ferret prop: memsize = 127
> ferret prop: x_lo = 180
> ferret prop: fill_type = fill
> ferret prop: t_hi = 08-Apr-1989
> ferret prop array: jnl_x = (180,540)
> ferret prop: view_centered = no
> ferret prop: variable_name = btt
> ferret prop array: jnl_y = (-90.0,89.8333333333333)
> ferret prop: y_lo = -90.0
> ferret prop array: jnl_z = ()
> ferret prop: HASH(0x88008d0) =
> ferret prop: diag_direction = up
> ferret prop: imgWidth = 800
> ferret prop: dataset_name = bt.nc
> ferret prop: rank = 2
> Starting:/usr/local/ferret/bin/ferret -gif -server -memsize 64
> Got reply: NOAA/PMEL TMAP
> Program FERRET
> Version 5.22 - 07/27/00
> 11-Jun-02 13:03
>
> yes? Sent command: cancel mode verify
> Got reply: yes? Sent command: cancel mode interp
> Got reply: yes? Sent command: canc data/all
> Got reply: yes? Sent command: go std_initialize "bt.nc" "1" "1" "btt"
> Got reply: yes? Sent command: set region/x="180":"540"/y="-90.0":"89.8333333333333"/k=1/t="08-Apr-1989":"08-Apr-1989"
> Got reply: yes? Sent command: set win/size=0.5
> Got reply: yes? Aspect: 0.749305555555556
> Sent command: set win/asp=0.749305555555556
> Got reply: yes?
> Searching for template/journal file: custom/std_gif.tmpl
>
> Searching for template/journal file: custom/std_gif_xyref.jnl
>
> Searching for template/journal file: custom/std_gif.tmpl
>
> Searching for template/journal file: custom/std_gif_xyref.jnl
>
> Searching for template/journal file: jnls/std_gif.tmpl
>
> Running template file jnls/std_gif.tmpl
> template: props is: HASH(0x88c35f0) arg is: 'dataset_name' val is: bt.nc
> template: props is: HASH(0x88c35f0) arg is: 'variable_name' val is: btt
> template: props is: HASH(0x88c35f0) arg is: 'fill_type' val is: fill
> template: props is: HASH(0x88c35f0) arg is: 'view' val is: xy
> template: props is: HASH(0x88c35f0) arg is: 'fill_levels' val is:
> template: props is: HASH(0x88c35f0) arg is: 'contour_levels' val is:
> template: props is: HASH(0x88c35f0) arg is: 'magnify' val is: 0
> template: props is: HASH(0x88c35f0) arg is: 'palette' val is:
> template: props is: HASH(0x88c35f0) arg is: 'draw_reference' val is: 0
> template: props is: HASH(0x88c35f0) arg is: 'do_shade' val is:
> template: props is: HASH(0x88c35f0) arg is: 'do_contour' val is: 1
> template: props is: HASH(0x88c35f0) arg is: 'do_overlay' val is:
> template: props is: HASH(0x88c35f0) arg is: 'rank' val is: 2
> template: props is: HASH(0x88c35f0) arg is: 'title' val is:
> template: props is: HASH(0x88c35f0) arg is: 'overlay_variable_name' val is:
> Sent command: GO "/tmp/lasgo18572669910.jnl"
> Got reply: **ERROR: a program limit has been reached: 2331719 points to plot
> Plot buffer size is: 990000
> CONTOUR/FILL/levels=30/set btt
> Command file, command group, or REPEAT execution aborted
> Error: transcript is
> Adding an acceptable error string: "*** NOTE: ".
> NOAA/PMEL TMAP
> Program FERRET
> Version 5.22 - 07/27/00
> 11-Jun-02 13:03
>
> yes? cancel mode verify
> yes? cancel mode interp
> yes? canc data/all
> yes? go std_initialize "bt.nc" "1" "1" "btt"
> *** NOTE: Coordinates out of order or missing on axis slat at subscript 2
> *** NOTE: A dummy axis of subscripts will be used
> *** NOTE: Coordinates out of order or missing on axis slon at subscript 2
> *** NOTE: A dummy axis of subscripts will be used
> yes? set region/x="180":"540"/y="-90.0":"89.8333333333333"/k=1/t="08-Apr-1989":"08-Apr-1989"
> yes? set win/size=0.5
> yes? set win/asp=0.749305555555556
> yes? GO "/tmp/lasgo18572669910.jnl"
> **ERROR: a program limit has been reached: 2331719 points to plot
> Plot buffer size is: 990000
> CONTOUR/FILL/levels=30/set btt
> Command file, command group, or REPEAT execution aborted
>
> killing...Waiting for program to exit...
> program exited...
> killing...Waiting for program to exit...
> program exited...

Follow-Ups:
- Ignored Parameters [was Re: memory error]
  - From: Mark A Ohrenschall

References:
- memory error
  - From: John C Cartwright
- Re: memory error
  - From: Jonathan Callahan

Previous by thread: Re: memory error
Next by thread: Ignored Parameters [was Re: memory error]

[Thread Prev][Thread Next][Index]

Dept of Commerce / NOAA / OAR / PMEL / TMAP
Contact Us | Privacy Policy | Disclaimer | Accessibility Statement