- changed status to resolved
Evaluate Squaring Improvement by Marco Bodrato
Issue #12
resolved
Marco Bodrato wrote:
"I've found a sequence that can giv some small improvements (in the linear part, not asymptotically) for matrix squaring. As a test for those sequences, I took your "strassen.c" for the M4RI library and modified it to apply the operations I propose. For the general product, the new code should be "as fast as" the old one (and maybe there is no need for the new sequence). For matrix squaring (C=A*A) it should give a small speed-up (around 1%)."
Code:
http://bodrato.it/software/strassen.html#M4R
Description
http://marco.bodrato.it/papers/Bodrato2008-StrassenLikeMatrixMultiplicationForSquares.pdf
Comments (3)
-
repo owner -
repo owner Opteron Timings
old multiplication
dim min avg med max 10000 2.580 2.616 2.620 2.708 16384 10.097 10.525 10.633 10.753 20000 18.413 18.635 18.677 18.849 32000 67.784 68.484 68.520 69.412 new multiplication
dim min avg med max 10000 2.620 2.630 2.624 2.676 16384 10.245 10.491 10.577 10.737 20000 18.557 18.773 18.817 19.037 32000 67.972 68.885 68.960 69.704 old squaring
dim min avg med max 10000 2.644 2.660 2.652 2.696 16384 10.353 10.738 10.837 11.013 20000 18.205 18.526 18.581 18.973 32000 68.120 69.360 69.760 70.236 new squaring
dim min avg med max 10000 2.568 2.600 2.596 2.652 16384 10.201 10.445 10.573 10.585 20000 18.125 18.534 18.629 18.685 32000 66.856 68.211 68.308 68.952 -
repo owner Opteron Timings
old multiplication
dim min avg med max 10000 2.580 2.616 2.620 2.708 16384 10.097 10.525 10.633 10.753 20000 18.413 18.635 18.677 18.849 32000 67.784 68.484 68.520 69.412 new multiplication
dim min avg med max 10000 2.620 2.630 2.624 2.676 16384 10.245 10.491 10.577 10.737 20000 18.557 18.773 18.817 19.037 32000 67.972 68.885 68.960 69.704 old squaring
dim min avg med max 10000 2.644 2.660 2.652 2.696 16384 10.353 10.738 10.837 11.013 20000 18.205 18.526 18.581 18.973 32000 68.120 69.360 69.760 70.236 new squaring
dim min avg med max 10000 2.568 2.600 2.596 2.652 16384 10.201 10.445 10.573 10.585 20000 18.125 18.534 18.629 18.685 32000 66.856 68.211 68.308 68.952 - Log in to comment
Committed in Wiki macro error: Changeset 297b5e5f9691 not found. with the following results on a Core2Duo CPU:
orig multiplication (geom)
new multiplication (geom)
orig squaring (geom)
new squaring (geom)