Mark Stock's benchmark on native Windows Radiance

radiance-dev
1

Has anyone got this going?

(Yes! I have managed to get it built and installed. No, I have no idea if it is actually working. More, later, when I've tested the thing.)

Randolph

2

Randolph M. Fritz <randolph+LD@...> writes:

Has anyone got this going?

(Yes! I have managed to get it built and installed. No, I have no idea
if it is actually working. More, later, when I've tested the thing.)

I am pleased to be able to say that I have been able to get the simulation,
though not the timing part, of this benchmark running. There is no "time"
command in the native Windows environment, and I don't yet know how to use
similar commands in the Windows Power Shell. Looking hopeful, though of
course I have not yet seen the output.

Mark Stock, if you're reading, would you like my cross-platform changes to
the .rad files? Windows, of course, does not have a "cat" command, which
means that "!cat" has to be replaced with "!xform".

Randolph

3

Side note on the use of "!cat" I often see things like this:

!cat input1.rad input2.rad | xform -ry 20

Which is completely silly. Why not just use:

!xform -ry 20 input1.rad input2.rad

Even if you are going to a program that expects input on stdin, you can redirect it instead:

!fussy_program < input | second_program etc...

I've seen almost no one use cat for a reason in a Radiance file. (No offense Mark, as I haven't even looked at yours!)

Cheers,
-Greg

···

From: "Randolph M. Fritz" <[email protected]>
Date: June 27, 2013 5:04:11 PM PDT

Randolph M. Fritz <randolph+LD@...> writes:

Has anyone got this going?

(Yes! I have managed to get it built and installed. No, I have no idea
if it is actually working. More, later, when I've tested the thing.)

I am pleased to be able to say that I have been able to get the simulation,
though not the timing part, of this benchmark running. There is no "time"
command in the native Windows environment, and I don't yet know how to use
similar commands in the Windows Power Shell. Looking hopeful, though of
course I have not yet seen the output.

Mark Stock, if you're reading, would you like my cross-platform changes to
the .rad files? Windows, of course, does not have a "cat" command, which
means that "!cat" has to be replaced with "!xform".

Randolph

4

I am pleased to be able to say that, in fact, I ran a successful simulation.
My hasty hack compilations of libtiff 3.9.7 (the last version 3 libtiff)
and zlib 1.2.8 also worked, which was nice to see.

Performance, well, not so good. It took 4219.9 wall clock seconds on a
fairly capable Xeon, which is 1.5-2x slower than I expect based on the
benchmark table. I suspect some tweaking of compilation options would
considerably improve the performance, even with the relatively blah
Microsoft VC10 compiler.

Randolph

5

Greg, Randolph,

I admit that my use of "cat" is simply habit---I saw it once (probably over 15 years ago) and never let go of it. I am actually not as savvy with Unix as I am on TV.

By all means make the benchmark as portable as you can and send me the changes. I will incorporate them as soon as I can.

Mark

···

On Thu, 27 Jun 2013, Gregory J. Ward wrote:

Side note on the use of "!cat" I often see things like this:

!cat input1.rad input2.rad | xform -ry 20

Which is completely silly. Why not just use:

!xform -ry 20 input1.rad input2.rad

Even if you are going to a program that expects input on stdin, you can redirect it instead:

!fussy_program < input | second_program etc...

I've seen almost no one use cat for a reason in a Radiance file. (No offense Mark, as I haven't even looked at yours!)

Cheers,
-Greg

From: "Randolph M. Fritz" <[email protected]>
Date: June 27, 2013 5:04:11 PM PDT

Randolph M. Fritz <randolph+LD@...> writes:

Has anyone got this going?

(Yes! I have managed to get it built and installed. No, I have no idea
if it is actually working. More, later, when I've tested the thing.)

I am pleased to be able to say that I have been able to get the simulation,
though not the timing part, of this benchmark running. There is no "time"
command in the native Windows environment, and I don't yet know how to use
similar commands in the Windows Power Shell. Looking hopeful, though of
course I have not yet seen the output.

Mark Stock, if you're reading, would you like my cross-platform changes to
the .rad files? Windows, of course, does not have a "cat" command, which
means that "!cat" has to be replaced with "!xform".

Randolph

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

6

I am silly. This vestigial code is still in OpenStudio, for "Radiance Classic" calcs with a continuous sky:

catCommand = "cat"
if /mswin/.match(RUBY_PLATFORM) or /mingw/.match(RUBY_PLATFORM)
catCommand = "type"
end

system("#{catCommand} ./numeric/#{space_name}.map | rtrace #{rtrace_args} …")

In this case, this really seemed to be the most direct way in. In any event, !type would be the true one-for-one replacement for !cat when porting Mark's scene to windoze.

- Rob

···

On Jun 27, 2013, at 6:21 PM, Gregory J. Ward <[email protected]> wrote:

Side note on the use of "!cat" I often see things like this:

!cat input1.rad input2.rad | xform -ry 20

Which is completely silly. Why not just use:

!xform -ry 20 input1.rad input2.rad

Even if you are going to a program that expects input on stdin, you can redirect it instead:

!fussy_program < input | second_program etc...

I've seen almost no one use cat for a reason in a Radiance file. (No offense Mark, as I haven't even looked at yours!)

Cheers,
-Greg

From: "Randolph M. Fritz" <[email protected]>
Date: June 27, 2013 5:04:11 PM PDT

Randolph M. Fritz <randolph+LD@...> writes:

Has anyone got this going?

(Yes! I have managed to get it built and installed. No, I have no idea
if it is actually working. More, later, when I've tested the thing.)

I am pleased to be able to say that I have been able to get the simulation,
though not the timing part, of this benchmark running. There is no "time"
command in the native Windows environment, and I don't yet know how to use
similar commands in the Windows Power Shell. Looking hopeful, though of
course I have not yet seen the output.

Mark Stock, if you're reading, would you like my cross-platform changes to
the .rad files? Windows, of course, does not have a "cat" command, which
means that "!cat" has to be replaced with "!xform".

Randolph

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

7

Mark,

I've seen lots of people use it, probably because it fits what people expect to see in an inline command more than the stdin redirection. I think there are even some examples in "Rendering with Radiance" using it, so you're in good company...

Cheers,
-Greg

···

From: Mark Stock <[email protected]>
Date: June 28, 2013 7:58:43 AM PDT

Greg, Randolph,

I admit that my use of "cat" is simply habit---I saw it once (probably over 15 years ago) and never let go of it. I am actually not as savvy with Unix as I am on TV.

By all means make the benchmark as portable as you can and send me the changes. I will incorporate them as soon as I can.

Mark

On Thu, 27 Jun 2013, Gregory J. Ward wrote:

Side note on the use of "!cat" I often see things like this:

!cat input1.rad input2.rad | xform -ry 20

Which is completely silly. Why not just use:

!xform -ry 20 input1.rad input2.rad

Even if you are going to a program that expects input on stdin, you can redirect it instead:

!fussy_program < input | second_program etc...

I've seen almost no one use cat for a reason in a Radiance file. (No offense Mark, as I haven't even looked at yours!)

Cheers,
-Greg

From: "Randolph M. Fritz" <[email protected]>
Date: June 27, 2013 5:04:11 PM PDT

Randolph M. Fritz <randolph+LD@...> writes:

Has anyone got this going?

(Yes! I have managed to get it built and installed. No, I have no idea
if it is actually working. More, later, when I've tested the thing.)

I am pleased to be able to say that I have been able to get the simulation,
though not the timing part, of this benchmark running. There is no "time"
command in the native Windows environment, and I don't yet know how to use
similar commands in the Windows Power Shell. Looking hopeful, though of
course I have not yet seen the output.

Mark Stock, if you're reading, would you like my cross-platform changes to
the .rad files? Windows, of course, does not have a "cat" command, which
means that "!cat" has to be replaced with "!xform".

Randolph

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

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

8

I favor "xform" with no options for this job; if the system has Radiance, I
expect it will have "xform", regardless of what other commands are or are
not available. But, YMMV.

Randolph

9

Rob Guglielmetti <rob.guglielmetti@...> writes:

I am silly. This vestigial code is still in OpenStudio, for "Radiance

Classic" calcs with a continuous sky:

catCommand = "cat"
if /mswin/.match(RUBY_PLATFORM) or /mingw/.match(RUBY_PLATFORM)
catCommand = "type"
end

system("#{catCommand} ./numeric/#{space_name}.map | rtrace #{rtrace_args} …")

boy, was that last remark off-target.

I don't understand how that code can work? Does something in that
implementation of "system" convert slashes to backslashes?

Anyhow, in this case, could one just perhaps divert rtrace's input?

  system("rtrace #{rtrace_args} < ./numeric/#{space_name}.map")

Or would that fail?

10

Redirecting the input works better on both Windows and Unix than cat, to my earlier point.

-Greg

···

Sent from my iPad

On Jun 28, 2013, at 4:58 PM, "Randolph M. Fritz" <[email protected]> wrote:

Rob Guglielmetti <rob.guglielmetti@...> writes:

I am silly. This vestigial code is still in OpenStudio, for "Radiance

Classic" calcs with a continuous sky:

catCommand = "cat"
if /mswin/.match(RUBY_PLATFORM) or /mingw/.match(RUBY_PLATFORM)
catCommand = "type"
end

system("#{catCommand} ./numeric/#{space_name}.map | rtrace #{rtrace_args} …")

boy, was that last remark off-target.

I don't understand how that code can work? Does something in that
implementation of "system" convert slashes to backslashes?

Anyhow, in this case, could one just perhaps divert rtrace's input?

system("rtrace #{rtrace_args} < ./numeric/#{space_name}.map")

Or would that fail?

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

11

I have benchmarked Relux with the Radiance (or 'Raytracer', as they call it now) engine under Windows. This was done with Process Monitor, which is free download from the M$ site:

It's not quite the same as the UNIX 'time' command, but saves you sitting there with the old stop watch.

What you do is you launch PM and set up a filter for 'rpict', and/or anything else you wish to benchmark. Now run the benchmark. PM will capture the start and end time of this process. You can export the result to a text file. You still need to work out the runtime:
runtime = end - start.

It's not terribly elegant, but does work.

PM tends to be a little temperamental. I never got it to work on certain machines. It also might crash unexpectedly.

Cheers

Axel

12

Measure-Command in PowerShell would be a much better solution -> http://technet.microsoft.com/en-us/library/ee176899.aspx

I was planning on making the necessary tweaks to Mark's benchmarks so they run on Windows via PowerShell but alas I never have any time for anything interesting.

Palbinder Sandher
Software Platform Engineer
T: +44 (0) 141 945 8500
F: +44 (0) 141 945 8501
http://www.iesve.com

**Design, Simulate + Innovate with the <Virtual Environment>**
Integrated Environmental Solutions Limited. Registered in Scotland No. SC151456
Registered Office - Helix Building, West Of Scotland Science Park, Glasgow G20 0SP
Email Disclaimer

···

-----Original Message-----
From: Axel Jacobs [mailto:[email protected]]
Sent: 29 June 2013 16:48
To: [email protected]
Subject: Re: [Radiance-dev] Mark Stock's benchmark on native Windows Radiance

I have benchmarked Relux with the Radiance (or 'Raytracer', as they call it now) engine under Windows. This was done with Process Monitor, which is free download from the M$ site:

It's not quite the same as the UNIX 'time' command, but saves you sitting there with the old stop watch.

What you do is you launch PM and set up a filter for 'rpict', and/or anything else you wish to benchmark. Now run the benchmark. PM will capture the start and end time of this process. You can export the result to a text file. You still need to work out the runtime:
runtime = end - start.

It's not terribly elegant, but does work.

PM tends to be a little temperamental. I never got it to work on certain machines. It also might crash unexpectedly.

Cheers

Axel

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