Ross Paterson ..... newtype Except αβγ = E (α β (Either String γ)). Define the ..... Arrows were invented as a programming abstraction by John Hughes [8], to deal.
MIT Laboratory for Computer Science, Cambridge, MA 02139. Parallel ...... Of course what we would really like is to have a fully parallel 2-D system, but.
Dec 25, 2013 - LG] 25 Dec 2013. PARALLEL COMPUTATION IS ESS. NABARUN MONDAL AND PARTHA P. GHOSH. Abstract. There are enormous amount ...
In sequential programming environments, common data structures are often .... take two logical terms as inputs, a pattern and a variable-free target, and it produces ... The bisection algorithm for computing the eigenvalues, an algorithm used in ...
photonic crystal waveguide, light is confined in the lattice plane by the ... The fundamental ingredient of the algorithm involves direct discretizations of the.
At this point eigenvalues are corrected for the spectral transformation in order to get the ...... programs. ACM Trans. Math. ... [32] T. S. Elliott and J. Majdalani.
Proof. Note that by the given inequality B is invertible, and Bâ1 + BâT is SPD. ... condition, since it implies that I âMâ1A will damp the (high-frequency) ...
PARALLEL COMPUTATION PLATFORM FOR SOMA. Miroslav Cervenka, Ivan Zelinka. Institute of Process Control and Applied Informatics. Faculty of ...
[email protected] .... Unfortunately, this is not true for a parallel program, ..... for example a high-performance oating-point functional unit. 7 .... chitectures is that, during any step, each of the p processors could generate a communicati
matrix computations, including matrix inversion and solving systems of linear
equations. ... matrix, power matrix, solving a linear system of equations, are very ...
Sci. 1988.3:233-283. Downloaded from arjournals.annualreviews.org by Eidgenossische Technische Hochschule Zurich on 06/08/06. For personal use only.
On the other hand, a number SR methods using CNN [14], .... CNN). Meanwhile, when the parallel algorithm is applied to. Proceedings of APSIPA Annual ...
mingâ, Chartwell-Bratt Ltd, Studentlitteratur, Lund, 1991. [44] âTbug User Guideâ, 3L Ltd, Livingston, Scotland, 1989. [45] âParallel C user Guide (compiler version ...
Parallel Extreme Pathway Computation for Metabolic Networks*. Lie-Quan Lee. Stanford Linear Accelerator Center, Stanford University, Stanford, CA 94309.
temporal parallelism by computing all of the time steps in parallel. Furthermore ..... Without proper termination of the computational grid, outward traveling waves.
1.34 ms. 11.76 ms. Gen. 3.17 ms. 18.63 ms. 70.99 ms. Gen., 2 VP sets. 3.23 ms. 18.75 ms. 93.19 ms. Table 1: Communications time (32-bit integer) for different ...
Jul 17, 2014 - investigate the possibility of parallel quantum computation based on two ...... G. & Long, G. Experimental realization of nonadiabatic Holonomic.
Feb 14, 2014 - We propose efficient parallel algorithms and implementations on shared .... Most numerical dense Gaussian elimination algorithms, like.
Burkill, I.H., A Dictionary of the -Economic Products of the Malay Peninsula, 2 vols
, ...... Walter, C., Pagan Races of the Malay Peninsula, New York, 1965. Wathen ...
Jan 30, 1997 - [email protected] http: icm.mcs.kent.edu pwang. Institute for Computational Mathematics. Kent State University. Kent, Ohio 44242-0001.
We consider parallel prefix computation on processors of dif- ferent and possibly .... one operation â¦; then A is a binary tree with n + 1 leaves. Thus A contains.
Modeling evolutionary principles found in nature in order to make the de- velopment of ... But are these two forces sufficient to explain the wonders of evolution at ...
poses scalability challenges for tool developers. In the field of ... To the best of our knowledge, this paper represents the first parallelization of the MR and. MRE algorithms in .... wordpress.com/2008/08/05/fast-bit-counting-routines/. Based on .
around the center (Conover test, p < 10. â1586. , n â¼ 105; the sample size is ...... Sawicka K, Bushell M, Spriggs KA, Willis AE. Polypyrimidine-tract-binding.
ZU064-05-FPR main. 7th January 2018. 18:35. Under consideration for publication in J. Functional Programming. 1. Arrows for Parallel Computation.
ZU064-05-FPR
main
7th January 2018
18:35
Under consideration for publication in J. Functional Programming
1
Arrows for Parallel Computation Supplementary Material M A RT I N B R AU N University Bayreuth, 95440 Bayreuth, Germany O L E G L O BAC H E V University Bayreuth, 95440 Bayreuth, Germany and P H I L T R I N D E R Glasgow University, Glasgow, G12 8QQ, Scotland
Abstract
This is supplementary material to our paper ‘Arrows for parallel computation’. It shows the speedup and mean overhead plots for all benchmarks.
Contents 1
2
Shared-Memory Plots
2
1.1
Rabin–Miller test
2
1.2
Sudoku
6
1.3 Gentleman Distributed Plots
8 9
2.1
Rabin–Miller test
9
2.2
Jacobi sum test
10
2.3
Gentleman
11
main
7th January 2018
2
18:35
M. Braun, O. Lobachev, and P. Trinder 1 Shared-Memory Plots 1.1 Rabin–Miller test 1.1.1 Speedup Speedup of Rabin–Miller test ‘11213 32’ linear Eden CP GpH Par Monad
16
Speedup
12
8
4
0 0
4
8
12
16
Threads Speedup of Rabin–Miller test ‘11213 64’ linear Eden CP GpH Par Monad
16
12 Speedup
ZU064-05-FPR
8
4
0 0
4
8 Threads
12
16
7th January 2018
18:35
3
Arrows for Parallel Computation 1.1.2 Overhead ·10−2
Mean Overhead for shared-memory Rabin–Miller test 11213 32 vs Eden CP
4 Mean Overhead 3 2 Overhead
main
1 0 −1 −2 −3 −4
0
4
8
12
16
Threads Mean Overhead for shared-memory Rabin–Miller test 11213 64 vs Eden CP −2 ·10 3 Mean Overhead 2.25 1.5 Overhead
ZU064-05-FPR
0.75 0 −0.75 −1.5 −2.25 −3
0
4
8 Threads
12
16
main
7th January 2018
4
18:35
M. Braun, O. Lobachev, and P. Trinder Mean Overhead for shared-memory Rabin–Miller test 11213 32 vs GpH 0.2 Mean Overhead 0.15
Overhead
0.1 5 · 10−2 0 −5 · 10−2 −0.1 −0.15 −0.2
0
·10−2
4
8
12
16
Threads Mean Overhead for shared-memory Rabin–Miller test 11213 64 vs GpH
5 Mean Overhead 3.75 2.5 Overhead
ZU064-05-FPR
1.25 0 −1.25 −2.5 −3.75 −5
0
4
8 Threads
12
16
7th January 2018
18:35
5
Arrows for Parallel Computation Mean Overhead for shared-memory Rabin–Miller test 11213 32 vs Par Monad 0.2 Mean Overhead 0.15 0.1 Overhead
main
5 · 10−2 0 −5 · 10−2 −0.1 −0.15 −0.2
0
4
8 12 16 Threads Mean Overhead for shared-memory Rabin–Miller test 11213 64 vs Par Monad
0.2 Mean Overhead 0.15 0.1 Overhead
ZU064-05-FPR
5 · 10−2 0 −5 · 10−2 −0.1 −0.15 −0.2
0
4
8 Threads
12
16
main
7th January 2018
6
18:35
M. Braun, O. Lobachev, and P. Trinder 1.2 Sudoku 1.2.1 Speedup Parallel speedup of Sudoku ‘1000’ linear Eden CP GpH Par Monad
16
Speedup
12
8
4
0 0
4
8
12
16
Threads
1.2.2 Overhead ·10−2
Mean Overhead for shared-memory Sudoku 1000 vs Eden CP
5 Mean Overhead 3.75 2.5 Overhead
ZU064-05-FPR
1.25 0 −1.25 −2.5 −3.75 −5
0
4
8 Threads
12
16
7th January 2018
18:35
7
Arrows for Parallel Computation
·10−2
Mean Overhead for shared-memory Sudoku 1000 vs GpH
5 Mean Overhead 3.75 2.5 Overhead
main
1.25 0 −1.25 −2.5 −3.75 −5
0
·10−2
4
8
12
16
Threads Mean Overhead for shared-memory Sudoku 1000 vs Par Monad
5 Mean Overhead 3.75 2.5 Overhead
ZU064-05-FPR
1.25 0 −1.25 −2.5 −3.75 −5
0
4
8 Threads
12
16
main
7th January 2018
8
18:35
M. Braun, O. Lobachev, and P. Trinder 1.3 Gentleman 1.3.1 Speedup Parallel speedup of Sudoku ‘512’ linear Eden CP GpH Par Monad
16
Speedup
12
8
4
0 0
4
8
12
16
Threads
1.3.2 Overhead ·10−2
Mean Overhead for shared-memory Gentleman 512 vs Eden CP
8 Mean Overhead 6 4 Overhead
ZU064-05-FPR
2 0 −2 −4 −6 −8
0
4
8 Threads
12
16
7th January 2018
18:35
9
Arrows for Parallel Computation 2 Distributed Plots 2.1 Rabin–Miller test 2.1.1 Speedup Speedup of distributed Rabin–Miller test ‘44497 256’ linear PArrows Eden
256 224 192 Speedup
main
160 128 96 64 32 0
0
32
64
96
128 160 Threads
192
224
256
2.1.2 Overhead ·10−2
Mean Overhead for distributed Rabin–Miller test 44497 256
4 Mean Overhead 3 2 Overhead
ZU064-05-FPR
1 0 −1 −2 −3 −4
0
32
64
96
128
160 Threads
192
224
256
main
7th January 2018
10
18:35
M. Braun, O. Lobachev, and P. Trinder 2.2 Jacobi sum test 2.2.1 Speedup Speedup of distributed Jacobi sum test ‘3217’ 32
Speedup
linear PArrows Eden
0 0
32
64
96
128
160 Threads
192
224
256
2.2.2 Overhead ·10−2
Mean Overhead for distributed Jacobi sum test 3217
4 Mean Overhead 3 2 Overhead
ZU064-05-FPR
1 0 −1 −2 −3 −4
0
32
64
96
128
160 Threads
192
224
256
7th January 2018
18:35
11
Arrows for Parallel Computation 2.3 Gentleman Mean Overhead for distributed Gentleman 4096
·10−2 4
Mean Overhead 3 2 Overhead
main
1 0 −1 −2 −3 −4
0
32
64
96
128
160 Threads
192
224
256
2.3.1 Speedup Speedup of distributed Gentleman ‘4096’ 160 linear PArrows Eden 128
