INTEL Workshop on "in-core" performance

Code examples

SMatrix

* standaloneMinimal.zip: Several of the original performance test programs for Smatrix and SVector, standalone version.

The other examples (see Attachments) are compiled like this: g++ -I standaloneMinimal/root/ fse_Assign_Sim_M_C_add_Sim_K_V.cxx

The data size can be changed in some of the examples with -DDIM_L=, ex: g++ -I standaloneMinimal/root/ fse_Assign_Sim_M_C_add_Sim_K_V.cxx -DDIM_L=2

-- RuneJensen - 2009-09-07

Geant4

This code shows usage of SSE intrinsics (double precision), but little runtime improvement.

-- RuneJensen - 2009-09-08

CMS geometrical transformations

  • match.tar.bz2: This is an example of matrix-vector multiplication in a real world context. Part of the "project" routine of CMSSW has been rewritten using intel SIMD instrinsics. The performance of the original code has been compared to the SIMD code (read perf_res.txt).

-- VincenzoInnocente - 2009-09-06

  • match_new.tar.bz2: New version of the matrix-multiplication code! SIMD code is unchanged but benchmarking is done more carefully using Rune's trick. Stalls are thus removed and finally SIMD code shows its benefits in performance (perf_res.txt).

  • example.tar.bz2: Yet another version of the match example, including the missing part from the original CMSSW code.
Topic attachments
I Attachment History Action Size Date Who Comment
Unknown file formatcxx C_Assign_A_mul_B.cxx r1 manage 0.9 K 2009-09-07 - 18:41 UnknownUser Basic matrix multiplication (C = A * B;) of two NxN sized matrises.
Unknown file formatcxx C_Assign_A_mul_Trans_B.cxx r1 manage 0.9 K 2009-09-07 - 18:42 UnknownUser Basic matrix multiplication (C = A * Transpose(B); ) of two NxN sized matrises.
Compressed Zip archivezip G4AffineTransform.zip r1 manage 11.4 K 2009-09-08 - 11:30 UnknownUser G4AffineTransform. Several code versions included, also SSE intrinsic.
Unknown file formatbz2 example.tar.bz2 r1 manage 26.3 K 2009-09-23 - 16:26 DanieleFrancescoKruse Yet another version of the match example, including the missing part from the original CMSSW code.
Unknown file formatcxx fse_Assign_Sim_M_C_add_Sim_K_V.cxx r1 manage 1.5 K 2009-09-07 - 18:24 UnknownUser Similatity, as used in the Kalman filter.
Compressed Zip archivezip madeTestCases.zip r1 manage 17.2 K 2009-09-12 - 14:24 UnknownUser A set of simple test cases. Includes several parts of the kalman filter, Cholesky/BK/Fast inversion, vector-vector addition and some matrix multiplications.
Unknown file formatbz2 match.tar.bz2 r1 manage 20.6 K 2009-09-08 - 18:25 DanieleFrancescoKruse This is an example of matrix-vector multiplication in a real world context. Part of the "project" routine of CMSSW has been rewritten using intel SIMD instrinsics. The performance of the original code has been compared to the SIMD code (read perf_res.txt).
Unknown file formatbz2 match_new.tar.bz2 r1 manage 23.0 K 2009-09-14 - 10:17 DanieleFrancescoKruse New version of the matrix-multiplication code! SIMD code is unchanged but benchmarking is done more carefully using Rune's trick. Stalls are thus removed and finally SIMD code shows its benefits in performance (perf_res.txt).
Compressed Zip archivezip modifiedMath.zip r1 manage 177.7 K 2009-09-07 - 18:33 UnknownUser A set of modified/optimized SMatrix and SVector include files. Several versions of some files exists. Intended as a possible basis for optimization.
Unknown file formatcxx similarity55.cxx r1 manage 1.2 K 2009-09-07 - 18:18 UnknownUser Similarity (matrix multiplication), intended as a simple test case.
Compressed Zip archivezip standaloneMinimal.zip r1 manage 80.9 K 2009-09-07 - 18:15 UnknownUser Several of the original performance test programs for Smatrix and SVector, standalone version.
Compressed Zip archivezip vecAddVec_tests.zip r1 manage 3.2 K 2009-09-09 - 17:49 UnknownUser 4 versions of (ROOT) vector-vector addition. 3 are using different tricks to avoid over-optimization, and one is handmade SSE intrinsics.
Edit | Attach | Watch | Print version | History: r11 < r10 < r9 < r8 < r7 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r11 - 2011-03-25 - VincenzoInnocente
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback