TCE-IP

From HPCRL Wiki
(Difference between revisions)
Jump to: navigation, search
(KfQqzIRBLgPiHXbq)
 
(990 intermediate revisions by 4 users not shown)
Line 1: Line 1:
http://goed4367.fusedtree.com/ hunter story maple
+
=== Introduction ===
http://foru3227.sitebooth.com/ youtube downloader how
+
The IP-TCE library is a set of C functions designed to provide
http://tero5704.freehostplace.com/ beef tenderloin recipe
+
high-performance array index permutation (index sorting), an important
http://tere1408.phreesite.com/ microsoft officexp
+
group of kernels used in many scientific applications and compilation
http://daqo5139.freewhost.com/ orbit downloader 2.1
+
techniques. The generation of the library code combines analytical and
http://refi0181.110mb.com/ avatar the last airbender book 3
+
empirical approaches. The details of how to generate the efficient 2-D IP
http://kare6429.700megs.com/ download windows xp cd key
+
(Matrix Transposition) code can be found in [1]. For high-dimensional arrays,
http://goed4367.fusedtree.com/maple-story-pet.html maple story warrior guide
+
with challenges such as high indexing costs and short dimensions, we employ
http://foru3227.sitebooth.com/psp-youtube-downloader.html youtube downloader add
+
optimizations such as restricting code versions, using one-level tiling and
http://tero5704.freehostplace.com/mongolian-beef-recipe.html shredded beef recipe
+
generating indexing code to achieve high performance without code size
http://tere1408.phreesite.com/instal-microsoft-office.html microsoft office exel 2007
+
explosion.
http://daqo5139.freewhost.com/orbit-downloader-version.html orbit downloader 1.5 7
+
 
http://refi0181.110mb.com/avatar-book-3-videos.html avatar complete book 2
+
The source code we provide does index permutation on 2-D,4-D and 6-D
http://kare6429.700megs.com/download-service-pack-2-for-windows-xp.html windows xp sp3 rc2 download
+
64-bit floating point arrays on IA-32 machines. Except for the 2-D functions that
http://goed4367.fusedtree.com/maple-story-mage-guide.html maple story fly
+
have more complicated optimizations than the others, the library employs SSE2
http://foru3227.sitebooth.com/youtube-downloader-work.html em youtube video downloader
+
instructions when the fastest varying dimensions of source and destination
http://tero5704.freehostplace.com/beef-marinade-recipe.html beef chuck steak recipe
+
arrays are multiples of vector sizes in elements, otherwise a scalar version
http://tere1408.phreesite.com/microsoft-visual-studio-tools-for-the-microsoft-office-system.html www microsoft officelive
+
is chosen. The user should be able to modify our code to get IP code working for
http://daqo5139.freewhost.com/orbit-downloader-for-linux.html orbit downloader add on
+
other dimension numbers.
http://refi0181.110mb.com/avatar-book-iii.html avatar book 3 dvds
+
 
http://kare6429.700megs.com/buy-and-download-windows-xp.html windows xp service pack 3 download iso
+
The IP-TCE library implements two variants of index permutation. One variant
http://goed4367.fusedtree.com/maple-story-level-hacks.html maple story mesos hack
+
implements B = factor * Permute(A, permutation) and the other one is the
http://foru3227.sitebooth.com/tech-crunch-youtube-downloader.html using youtube downloader
+
accumulative version B = B + factor * Permute(A, permutation). Doing so makes
http://tero5704.freehostplace.com/beef-and-gravy-recipe.html italian roast beef recipe
+
our code compatible with the index permutation routines in nwchem/tce. We have
http://tere1408.phreesite.com/microsoft-office-proffesional-plus-2007.html error 25003 microsoft office
+
been able to plug our code into nwchem/tce and obtained overall performance
http://daqo5139.freewhost.com/orbit-downloader-wiki.html orbit downloader doesnt
+
improvements ranging from 74% to 253% with different methods and inputs.
http://refi0181.110mb.com/avatar-book-3-schedule.html book 3 of avatar
+
 
http://kare6429.700megs.com/download-of-windows-xp-service-pack-2.html how to download windows xp service pack 3
+
=== Compilation ===
http://goed4367.fusedtree.com/maple-story-free-hacks.html maple story hacks forums
+
The library code can be compiled by either the Intel C compiler or the GNU C
http://foru3227.sitebooth.com/youtube-google-video-downloader-for-zune.html video downloader not working on youtube
+
compiler. We used icc 10.1 and gcc 4.1.2 when testing the code. If other
http://tero5704.freehostplace.com/sirloin-roast-beef-recipe.html roast beef gravy recipe
+
compilers such as pgcc is used, the user is responsible for finding the
http://tere1408.phreesite.com/microsoft-officeexcel.html microsoft office outlook not implemented
+
alignment directives used by the compiler and changing the library code
http://daqo5139.freewhost.com/to-use-orbit-downloader-with.html use orbit downloader to download
+
correspondingly.
http://refi0181.110mb.com/avatar-book-3-on-dvd.html book 3 of the avatar
+
 
http://kare6429.700megs.com/free-download-windows-xp-upgrade.html windows xp drivers free download
+
=== Library Usage ===
 +
The interface of the IP-TCE library is compatible with the index permutation
 +
routines in nwchem/tce and is mainly for Fortran used. For example,
 +
the prototype of the 4-D non-accumulative permutation routine is:
 +
 
 +
                  tce_sort_4_(double* unsorted,double* sorted,
 +
                              int* a_in, int* b_in, int* c_in, int* d_in,
 +
                              int* i_in, int* j_in, int* k_in, int* l_in,
 +
                              double* factor_in)
 +
 
 +
Where all the arguments are pointers and function names are in the lowercase
 +
and end with "_". If needed,  the user can write a wrapper or directly modify
 +
the function (which is actually a wrapper to the SIMD and scalar IP functions)
 +
to obtain a desired interface.
 +
 
 +
=== Contact Info ===
 +
Please contact Qingda Lu(luq@cse.ohio-state.edu) for questions.
 +
 
 +
=== Reference ===
 +
[1] Qingda Lu, Sriram Krishnamoorthy, P. Sadayappan: Combining analytical and
 +
empirical approaches in tuning matrix transposition. 15th International
 +
Conference on Parallel Architecture and Compilation Techniques(PACT 2006):233-242

Latest revision as of 23:34, 20 February 2009

Contents

Introduction

The IP-TCE library is a set of C functions designed to provide high-performance array index permutation (index sorting), an important group of kernels used in many scientific applications and compilation techniques. The generation of the library code combines analytical and empirical approaches. The details of how to generate the efficient 2-D IP (Matrix Transposition) code can be found in [1]. For high-dimensional arrays, with challenges such as high indexing costs and short dimensions, we employ optimizations such as restricting code versions, using one-level tiling and generating indexing code to achieve high performance without code size explosion.

The source code we provide does index permutation on 2-D,4-D and 6-D 64-bit floating point arrays on IA-32 machines. Except for the 2-D functions that have more complicated optimizations than the others, the library employs SSE2 instructions when the fastest varying dimensions of source and destination arrays are multiples of vector sizes in elements, otherwise a scalar version is chosen. The user should be able to modify our code to get IP code working for other dimension numbers.

The IP-TCE library implements two variants of index permutation. One variant implements B = factor * Permute(A, permutation) and the other one is the accumulative version B = B + factor * Permute(A, permutation). Doing so makes our code compatible with the index permutation routines in nwchem/tce. We have been able to plug our code into nwchem/tce and obtained overall performance improvements ranging from 74% to 253% with different methods and inputs.

Compilation

The library code can be compiled by either the Intel C compiler or the GNU C compiler. We used icc 10.1 and gcc 4.1.2 when testing the code. If other compilers such as pgcc is used, the user is responsible for finding the alignment directives used by the compiler and changing the library code correspondingly.

Library Usage

The interface of the IP-TCE library is compatible with the index permutation routines in nwchem/tce and is mainly for Fortran used. For example, the prototype of the 4-D non-accumulative permutation routine is:

                  tce_sort_4_(double* unsorted,double* sorted,
                              int* a_in, int* b_in, int* c_in, int* d_in,
                              int* i_in, int* j_in, int* k_in, int* l_in,
                              double* factor_in)

Where all the arguments are pointers and function names are in the lowercase and end with "_". If needed, the user can write a wrapper or directly modify the function (which is actually a wrapper to the SIMD and scalar IP functions) to obtain a desired interface.

Contact Info

Please contact Qingda Lu(luq@cse.ohio-state.edu) for questions.

Reference

[1] Qingda Lu, Sriram Krishnamoorthy, P. Sadayappan: Combining analytical and empirical approaches in tuning matrix transposition. 15th International Conference on Parallel Architecture and Compilation Techniques(PACT 2006):233-242