resolution mismatch revisited

Hi all,

I had followed Andrews discussion on resolution mismatch a few weeks back. However, I seem to be running into a different type of resolution mismatch problem right now.

I am venturing into the world of parallel processing on a dual G5 Macintosh, and my first invocation of rpiece works fine.

rpiece -X 5 -Y 5 -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst -af b_b_b.amb -t 600 -vf ../views/plan.vf -dt 0.05 -dc 0.10 -ds 0.01 -dj 0.6 -dr 3 -dp 4096 -st 0.01 -sj 1 -av 0.01 0.01 0.01 -aw 1 -ab 2 ad 2048 -as 1024 --aa 0.05 -ar 256 -aE ../ambient_exclude -x 500 -y 500 -ms 2.7 -lr 12 -lw 0.0005 -ps 1 -pt 0.04 -e b_b_b.out b_b_b.oct &

However, when I try to start the second process

rpiece -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst b_b_b.oct &

I get the error message

rpiece: resolution mismatch on file "b_b_b_1_indirect_plan.pic

I tried adjusting the -X -Y and -x -y parameters, but the same thing happened. Just out of curiosity, I also tried adjusting the -av and -aw to 0, and the same results.

I am certain that this is not a problem with a limit on the available resource. The machine has 1.5GB of memory.

Any hints as to what could be wrong?

Thanks.

John

John,

We use rpiece but probably a little differently in terms of invocation. There are a couple of things to look at:

   1. typo - there is a typo in your first command line you have --aa
      which should be -a
   2. invocations - there is a short form for invoking rpiece after the
      initial invocation, which is what you are trying to do here. I
      would check the man page for rpiece carefully to make sure you are
      doing this correctly, it looks like you are leaving out the -x and
      -y image resolution parameters as well as the viewfile -vf from
      the short version of the invocation. As an alternative to keep
      things more "simple", I would suggest just calling exactly the
      same thing as the initial invocation. This could be coded into a
      simple script or makefile to save yourself some of the headache of
      typing the whole thing our or a potentially incorrect cut and paste.
   3. scene complexity - depending on the simplicity or complexity of
      the scene you are trying to run here there is a another possible
      problem. If the scene is really simple and the image size small
      (500x500 is not that large) then the first rpiece process could
      have completed everything before the next invocation (I seem to
      recall that this may produce the same kind of message?) Are you
      getting a completed image?

-Jack de Valpine

John S. An wrote:

···

Hi all,

I had followed Andrews discussion on resolution mismatch a few weeks back. However, I seem to be running into a different type of resolution mismatch problem right now.

I am venturing into the world of parallel processing on a dual G5 Macintosh, and my first invocation of rpiece works fine.

rpiece -X 5 -Y 5 -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst -af b_b_b.amb -t 600 -vf ../views/plan.vf -dt 0.05 -dc 0.10 -ds 0.01 -dj 0.6 -dr 3 -dp 4096 -st 0.01 -sj 1 -av 0.01 0.01 0.01 -aw 1 -ab 2 ad 2048 -as 1024 --aa 0.05 -ar 256 -aE ../ambient_exclude -x 500 -y 500 -ms 2.7 -lr 12 -lw 0.0005 -ps 1 -pt 0.04 -e b_b_b.out b_b_b.oct &

However, when I try to start the second process

rpiece -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst b_b_b.oct &

I get the error message

rpiece: resolution mismatch on file "b_b_b_1_indirect_plan.pic

I tried adjusting the -X -Y and -x -y parameters, but the same thing happened. Just out of curiosity, I also tried adjusting the -av and -aw to 0, and the same results.

I am certain that this is not a problem with a limit on the available resource. The machine has 1.5GB of memory.

Any hints as to what could be wrong?

Thanks.

John

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

--
# John E. de Valpine
# president
#
# visarc incorporated
# http://www.visarc.com
#
# channeling technology for superior design and construction

Jack,

Thanks for you advice.

The short invocation was what I had found in the radiance-online archives, in which Greg gave a step by step approach. I did take your advice and tried making the -x -y and -vf parameters explicit in the second invocation. I no longer get the resolution mismatch error. Also, although my image is only 500 by 500, the settings seem to be sufficiently high enough that each piece takes at least a couple minutes to render.

However, a few things that I have noticed are:
1) when I use the "top" command, there are 4 rpict processes. 2 seem to be active, and 2 are idle. I had assumed that there should be only 3 rpict processes (2 active, and one waiting fork).
2) when I look over the output log, both processes are working on frame 1, then again, both working on frame 2, etc.
3) the final output .pic file only has the center (where the light fixture is located) rendered, and the rest of the image is completely black.

I'm sure I'm making a dumb newbie mistake, but can't seem to figure out what it is. Could you (or someone) post a detailed printout of which commands with what parameters one would use to get a simple rpiece process going?

Thanks.

(btw, my initial attempt at the rpiece command was adapted from Greg's post a while back)

John

···

_________________________________
Jack de Valpine wrote:

John,

We use rpiece but probably a little differently in terms of invocation.
There are a couple of things to look at:

    1. typo - there is a typo in your first command line you have --aa
       which should be -a
    2. invocations - there is a short form for invoking rpiece after the
       initial invocation, which is what you are trying to do here. I
       would check the man page for rpiece carefully to make sure you are
       doing this correctly, it looks like you are leaving out the -x and
       -y image resolution parameters as well as the viewfile -vf from
       the short version of the invocation. As an alternative to keep
       things more "simple", I would suggest just calling exactly the
       same thing as the initial invocation. This could be coded into a
       simple script or makefile to save yourself some of the headache of
       typing the whole thing our or a potentially incorrect cut and paste.
    3. scene complexity - depending on the simplicity or complexity of
       the scene you are trying to run here there is a another possible
       problem. If the scene is really simple and the image size small
       (500x500 is not that large) then the first rpiece process could
       have completed everything before the next invocation (I seem to
       recall that this may produce the same kind of message?) Are you
       getting a completed image?

-Jack de Valpine

John S. An wrote:

> Hi all,
>
> I had followed Andrews discussion on resolution mismatch a few weeks
> back. However, I seem to be running into a different type of
> resolution mismatch problem right now.
>
> I am venturing into the world of parallel processing on a dual G5
> Macintosh, and my first invocation of rpiece works fine.
>
> rpiece -X 5 -Y 5 -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP
> b_b_b.pst -af b_b_b.amb -t 600 -vf ../views/plan.vf -dt 0.05 -dc 0.10
> -ds 0.01 -dj 0.6 -dr 3 -dp 4096 -st 0.01 -sj 1 -av 0.01 0.01 0.01 -aw
> 1 -ab 2 ad 2048 -as 1024 --aa 0.05 -ar 256 -aE ../ambient_exclude -x
> 500 -y 500 -ms 2.7 -lr 12 -lw 0.0005 -ps 1 -pt 0.04 -e b_b_b.out
> b_b_b.oct &
>
> However, when I try to start the second process
>
> rpiece -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst
> b_b_b.oct &
>
> I get the error message
>
> rpiece: resolution mismatch on file "b_b_b_1_indirect_plan.pic
>
> I tried adjusting the -X -Y and -x -y parameters, but the same thing
> happened. Just out of curiosity, I also tried adjusting the -av and
> -aw to 0, and the same results.
>
> I am certain that this is not a problem with a limit on the available
> resource. The machine has 1.5GB of memory.
>
> Any hints as to what could be wrong?
>
> Thanks.
>
> John

Hey John,

OK, a step in the right direction ;->, some other thought below....

-Jack

John S. An wrote:

Jack,

Thanks for you advice.

The short invocation was what I had found in the radiance-online archives, in which Greg gave a step by step approach. I did take your advice and tried making the -x -y and -vf parameters explicit in the second invocation. I no longer get the resolution mismatch error. Also, although my image is only 500 by 500, the settings seem to be sufficiently high enough that each piece takes at least a couple minutes to render.

However, a few things that I have noticed are:
1) when I use the "top" command, there are 4 rpict processes. 2 seem to be active, and 2 are idle. I had assumed that there should be only 3 rpict processes (2 active, and one waiting fork).

Hmm. Actually, I think for 2 rpiece processes you should expect the following to appear under top:

    * rpiece - 2 instances (minimum), one for each invocation, and you
      may get more as well depending on how radiance was compiled, the
      "extra" ones are around to wait for data to be written out for a
      particular piece before exiting
    * rpict - 5 processes, I think that 1 is a listener that forks off
      processes as needed based on input from rpiece and then a reader
      and writer for each process, perhaps someone else can help clarify.

I suspect that you actually have an old rpict process hanging around. I suggest that you check the process table for rpict and rpiece now (when there is nothing running) and see if there are any processes reported. If so they may be left around from when the second rpiece was dieing and reporting its error. You should be able to use the following to check quickly for what you want:

ps aux | grep rpict
ps aux | grep rpiece

2) when I look over the output log, both processes are working on frame 1, then again, both working on frame 2, etc.

As far as the log is concerned this is ok. Each rpiece process will starting it frame counting at 1.

3) the final output .pic file only has the center (where the light fixture is located) rendered, and the rest of the image is completely black.

I think that you should check the process table as indicated, make sure everything is cleaned up and start fresh. It is also possible that you are looking at an image that is still being rendered.

···

I'm sure I'm making a dumb newbie mistake, but can't seem to figure out what it is. Could you (or someone) post a detailed printout of which commands with what parameters one would use to get a simple rpiece process going?

Thanks.

(btw, my initial attempt at the rpiece command was adapted from Greg's post a while back)

John

_________________________________
Jack de Valpine wrote:

John,

We use rpiece but probably a little differently in terms of invocation.
There are a couple of things to look at:

1. typo - there is a typo in your first command line you have --aa
which should be -a
2. invocations - there is a short form for invoking rpiece after the
initial invocation, which is what you are trying to do here. I
would check the man page for rpiece carefully to make sure you are
doing this correctly, it looks like you are leaving out the -x and
-y image resolution parameters as well as the viewfile -vf from
the short version of the invocation. As an alternative to keep
things more "simple", I would suggest just calling exactly the
same thing as the initial invocation. This could be coded into a
simple script or makefile to save yourself some of the headache of
typing the whole thing our or a potentially incorrect cut and paste.
3. scene complexity - depending on the simplicity or complexity of
the scene you are trying to run here there is a another possible
problem. If the scene is really simple and the image size small
(500x500 is not that large) then the first rpiece process could
have completed everything before the next invocation (I seem to
recall that this may produce the same kind of message?) Are you
getting a completed image?

-Jack de Valpine

John S. An wrote:

> Hi all,
>
> I had followed Andrews discussion on resolution mismatch a few weeks
> back. However, I seem to be running into a different type of
> resolution mismatch problem right now.
>
> I am venturing into the world of parallel processing on a dual G5
> Macintosh, and my first invocation of rpiece works fine.
>
> rpiece -X 5 -Y 5 -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP
> b_b_b.pst -af b_b_b.amb -t 600 -vf ../views/plan.vf -dt 0.05 -dc 0.10
> -ds 0.01 -dj 0.6 -dr 3 -dp 4096 -st 0.01 -sj 1 -av 0.01 0.01 0.01 -aw
> 1 -ab 2 ad 2048 -as 1024 --aa 0.05 -ar 256 -aE ../ambient_exclude -x
> 500 -y 500 -ms 2.7 -lr 12 -lw 0.0005 -ps 1 -pt 0.04 -e b_b_b.out
> b_b_b.oct &
>
> However, when I try to start the second process
>
> rpiece -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst
> b_b_b.oct &
>
> I get the error message
>
> rpiece: resolution mismatch on file "b_b_b_1_indirect_plan.pic
>
> I tried adjusting the -X -Y and -x -y parameters, but the same thing
> happened. Just out of curiosity, I also tried adjusting the -av and
> -aw to 0, and the same results.
>
> I am certain that this is not a problem with a limit on the available
> resource. The machine has 1.5GB of memory.
>
> Any hints as to what could be wrong?
>
> Thanks.
>
> John

--
# John E. de Valpine
# president
#
# visarc incorporated
# http://www.visarc.com
#
# channeling technology for superior design and construction

Jack,

I think it's working now. Actually, after following your initial advice of cleaning up, I think everything was working.

The problem of the black room was that all the surfaces had very low reflectances (somewhere in the order of 7% to 14%), and there was only one 100w incandescent bulb in the scene. When I used the human option, then the space was visible.

So, I guess the moral of the story is clean up after yourself, and know what it is you are doing.

Thanks for you help.

John

···

On Dec 16, 2003, at 6:00 AM, [email protected] wrote:

Hey John,

OK, a step in the right direction ;->, some other thought below....

-Jack

John S. An wrote:

Jack,

Thanks for you advice.

The short invocation was what I had found in the radiance-online
archives, in which Greg gave a step by step approach. I did take your
advice and tried making the -x -y and -vf parameters explicit in the
second invocation. I no longer get the resolution mismatch error.
Also, although my image is only 500 by 500, the settings seem to be
sufficiently high enough that each piece takes at least a couple
minutes to render.

However, a few things that I have noticed are:
1) when I use the "top" command, there are 4 rpict processes. 2 seem
to be active, and 2 are idle. I had assumed that there should be only
3 rpict processes (2 active, and one waiting fork).

Hmm. Actually, I think for 2 rpiece processes you should expect the
following to appear under top:

    * rpiece - 2 instances (minimum), one for each invocation, and you
      may get more as well depending on how radiance was compiled, the
      "extra" ones are around to wait for data to be written out for a
      particular piece before exiting
    * rpict - 5 processes, I think that 1 is a listener that forks off
      processes as needed based on input from rpiece and then a reader
      and writer for each process, perhaps someone else can help clarify.

I suspect that you actually have an old rpict process hanging around. I
suggest that you check the process table for rpict and rpiece now (when
there is nothing running) and see if there are any processes reported.
If so they may be left around from when the second rpiece was dieing and
reporting its error. You should be able to use the following to check
quickly for what you want:

ps aux | grep rpict
ps aux | grep rpiece

2) when I look over the output log, both processes are working on
frame 1, then again, both working on frame 2, etc.

As far as the log is concerned this is ok. Each rpiece process will
starting it frame counting at 1.

3) the final output .pic file only has the center (where the light
fixture is located) rendered, and the rest of the image is completely
black.

I think that you should check the process table as indicated, make sure
everything is cleaned up and start fresh. It is also possible that you
are looking at an image that is still being rendered.

I'm sure I'm making a dumb newbie mistake, but can't seem to figure
out what it is. Could you (or someone) post a detailed printout of
which commands with what parameters one would use to get a simple
rpiece process going?

Thanks.

(btw, my initial attempt at the rpiece command was adapted from Greg's
post a while back)

John

_________________________________
Jack de Valpine wrote:

John,

We use rpiece but probably a little differently in terms of invocation.
There are a couple of things to look at:

1. typo - there is a typo in your first command line you have --aa
which should be -a
2. invocations - there is a short form for invoking rpiece after the
initial invocation, which is what you are trying to do here. I
would check the man page for rpiece carefully to make sure you are
doing this correctly, it looks like you are leaving out the -x and
-y image resolution parameters as well as the viewfile -vf from
the short version of the invocation. As an alternative to keep
things more "simple", I would suggest just calling exactly the
same thing as the initial invocation. This could be coded into a
simple script or makefile to save yourself some of the headache of
typing the whole thing our or a potentially incorrect cut and paste.
3. scene complexity - depending on the simplicity or complexity of
the scene you are trying to run here there is a another possible
problem. If the scene is really simple and the image size small
(500x500 is not that large) then the first rpiece process could
have completed everything before the next invocation (I seem to
recall that this may produce the same kind of message?) Are you
getting a completed image?

-Jack de Valpine

John S. An wrote:

Hi all,

I had followed Andrews discussion on resolution mismatch a few weeks
back. However, I seem to be running into a different type of
resolution mismatch problem right now.

I am venturing into the world of parallel processing on a dual G5
Macintosh, and my first invocation of rpiece works fine.

rpiece -X 5 -Y 5 -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP
b_b_b.pst -af b_b_b.amb -t 600 -vf ../views/plan.vf -dt 0.05 -dc 0.10
-ds 0.01 -dj 0.6 -dr 3 -dp 4096 -st 0.01 -sj 1 -av 0.01 0.01 0.01 -aw
1 -ab 2 ad 2048 -as 1024 --aa 0.05 -ar 256 -aE ../ambient_exclude -x
500 -y 500 -ms 2.7 -lr 12 -lw 0.0005 -ps 1 -pt 0.04 -e b_b_b.out
b_b_b.oct &

However, when I try to start the second process

rpiece -F b_b_b_sync -o b_b_b_1_indirect_plan.pic -PP b_b_b.pst
b_b_b.oct &

I get the error message

rpiece: resolution mismatch on file "b_b_b_1_indirect_plan.pic

I tried adjusting the -X -Y and -x -y parameters, but the same thing
happened. Just out of curiosity, I also tried adjusting the -av and
-aw to 0, and the same results.

I am certain that this is not a problem with a limit on the available
resource. The machine has 1.5GB of memory.

Any hints as to what could be wrong?

Thanks.

John

--
# John E. de Valpine
# president
#
# visarc incorporated
# http://www.visarc.com
#
# channeling technology for superior design and construction

Hi John,

Great glad it worked out. I think the moral of the story (any radiance story), is to wade in, get your hand dirty and see what happens. Good job.

-Jack

John S. An wrote:

···

Jack,

I think it's working now. Actually, after following your initial advice of cleaning up, I think everything was working.

The problem of the black room was that all the surfaces had very low reflectances (somewhere in the order of 7% to 14%), and there was only one 100w incandescent bulb in the scene. When I used the human option, then the space was visible.

So, I guess the moral of the story is clean up after yourself, and know what it is you are doing.

Thanks for you help.

John

--
# John E. de Valpine
# president
#
# visarc incorporated
# http://www.visarc.com
#
# channeling technology for superior design and construction

Jack de Valpine wrote:

Great glad it worked out. I think the moral of the story (any radiance story), is to wade in, get your hand dirty and see what happens.

Hi Jack,

I followed your instructions, here's what happened:
my feet are wet and my hand is dirty. :wink:

···

----

      Rob Guglielmetti

e. [email protected]
w. www.rumblestrip.org