Madagascar Code Patterns - Revision history

Nick: /* Loop over ensembles */

2011-10-29T01:20:22Z

Loop over ensembles

← Older revision		Revision as of 01:20, 29 October 2011
Line 5:		Line 5:

	It is necessary in the case of algorithms that act on one ensemble at a time. Parallelizable with OMP.		It is necessary in the case of algorithms that act on one ensemble at a time. Parallelizable with OMP.

	It has the potential bug that the ensemble array is allocated using dimensions that are of type int (usually 4-byte). This will overflow on some rare occasions, and they should either be defined as 8-byte, or there should be a check function in which dimensions are read into a 8-byte int type and checked whether they will overflow the int. A different function than sf_histint needs to be created for this purpose, or the check to be implemented inside sf_histint.

	==I/O-optimized loop over samples==		==I/O-optimized loop over samples==

Nick: /Loop over ensembles / vulnerability

2011-10-29T01:16:15Z

Loop over ensembles: vulnerability

← Older revision		Revision as of 01:16, 29 October 2011
Line 1:		Line 1:
	All patterns are in C, unless otherwise noted.		All patterns are in C, unless otherwise noted.

			==Loop over ensembles==
			This is a simple loop over reading 1-D arrays ("traces"), 2-D ("gathers"), etc.

			It is necessary in the case of algorithms that act on one ensemble at a time. Parallelizable with OMP.

			It has the potential bug that the ensemble array is allocated using dimensions that are of type int (usually 4-byte). This will overflow on some rare occasions, and they should either be defined as 8-byte, or there should be a check function in which dimensions are read into a 8-byte int type and checked whether they will overflow the int. A different function than sf_histint needs to be created for this purpose, or the check to be implemented inside sf_histint.

	==I/O-optimized loop over samples==		==I/O-optimized loop over samples==
Line 37:		Line 44:

	The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.		The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.

	==OMP parallelized loop==		==OMP parallelized loop==
	===Description and usage===		===Description and usage===

Nick: /* Parallelized loop */

2011-07-24T14:49:13Z

Parallelized loop

← Older revision		Revision as of 14:49, 24 July 2011
Line 37:		Line 37:

	The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.		The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.
	==~~Parallelized~~ loop==		==OMP parallelized loop==
	===Description and usage===		===Description and usage===
			Shared-memory parallelization using the [https://secure.wikimedia.org/wikipedia/en/wiki/OpenMP OpenMP] library.
	===Example===		===Example===
	===Potential for improvement===		===Potential for improvement===

Nick: /* Parallelized, I/O-optimized loop over traces */

2011-07-24T14:29:56Z

Parallelized, I/O-optimized loop over traces

← Older revision		Revision as of 14:29, 24 July 2011
Line 37:		Line 37:

	The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.		The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.
	==Parallelized, I/O-optimized loop over ~~traces~~==		==Parallelized loop==
			===Description and usage===
			===Example===
			===Potential for improvement===

			==Parallelized, I/O-optimized loop over samples==
	===Description and usage===		===Description and usage===
	===Example===		===Example===
	===Potential for improvement===		===Potential for improvement===

Nick: /* I/O-optimized loop over samples */

2011-07-24T03:32:48Z

I/O-optimized loop over samples

← Older revision		Revision as of 03:32, 24 July 2011
Line 37:		Line 37:

	The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.		The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.
			==Parallelized, I/O-optimized loop over traces==
			===Description and usage===
			===Example===
			===Potential for improvement===

Nick: Created page with "All patterns are in C, unless otherwise noted. ==I/O-optimized loop over samples== ===Description and usage=== It consists of looping over an entire dataset and applying a given..."

2011-07-24T03:31:38Z

Created page with "All patterns are in C, unless otherwise noted. ==I/O-optimized loop over samples== ===Description and usage=== It consists of looping over an entire dataset and applying a given..."

New page

All patterns are in C, unless otherwise noted.

==I/O-optimized loop over samples==
===Description and usage===
It consists of looping over an entire dataset and applying a given procedure to every single sample in a dataset, regardless of what "trace"/"frame"/"volume" it belongs to. Example: computing the sum of all elements of a dataset; computing a histogram; performing a clip operation; etc. It uses the BUFSIZ macro defined in <tt>stdio.h</tt> to ensure efficient stream I/O. Its occurences can be easily found by grepping for BUFSIZ in the codebase.
===Example===
<c>
int n; /* Total number of elements in dataset */
int nbuf; /* Number of elements in I/O buffer */
float *fbuf; /* I/O array */
sf_file in=NULL; /* Input file. Here is stdin, but this is not compulsory */

in = sf_input("in");

n = sf_filesize(in);

/* This example uses float as data type.
Any other data type (int, sf_complex, etc) can be used, as appropriate */
nbuf = BUFSIZ/sizeof(float);

fbuf = sf_floatalloc(nbuf);

for (; n > 0; n -= nbuf) {

if (nbuf > n) nbuf = n;

sf_floatread(fbuf, nbuf, in);

for (i=0; i < nbuf; i++) {
/* Do computations here */
}
}
</c>

===Potential for improvement===
This pattern should be parallelized using OpenMP.

The [http://www.delorie.com/gnu/docs/glibc/libc_226.html GNU C Library documentation] states that when doing I/O on a file (as opposed to a stream), the <tt>st_blksize</tt> field of the file attributes is a better choice than <tt>BUFSIZ</tt>.