Skip to content

Commit bcf2855

Browse files
committed
first commit
0 parents  commit bcf2855

File tree

1 file changed

+34
-0
lines changed

1 file changed

+34
-0
lines changed

README.md

Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
# PSort: Parallel Sorting using Sampling with Randomized and Deterministic Approaches
2+
3+
Previous schemes for sorting on general-purpose parallel machines have
4+
had to choose between poor load balancing and irregular communication
5+
or multiple rounds of all-to-all personalized communication. This code
6+
provides a sample sort which uses only two rounds of regular
7+
all-to-all personalized communication in a scheme that yields very
8+
good load balancing with virtually no overhead. Moreover, unlike
9+
previous variations, our algorithm efficiently handles the presence of
10+
duplicate values without the overhead of tagging each element with a
11+
unique identifier. This algorithm was implemented in Split-C and run
12+
on a variety of platforms, including the Thinking Machines CM-5, the
13+
IBM SP-2, and the Cray Research T3D. This implementation of our new
14+
algorithm seems to outperform all similar algorithms known to the
15+
authors on these platforms, and its performance is invariant over the
16+
set of input distributions unlike previous efficient algorithms. Our
17+
results also compare favorably with those reported for the simpler
18+
ranking problem posed by the NAS Integer Sorting (IS) Benchmark.
19+
20+
The code also provides a new deterministic parallel sorting algorithm
21+
based on the regular sampling approach. The performance compares
22+
closely to that of our random sample sort algorithm, which seems to
23+
outperform all similar algorithms known to the authors on these
24+
platforms. Together, their performance is nearly invariant over the
25+
set of input distributions, unlike previous efficient
26+
algorithms. However, unlike our randomized sorting algorithm, the
27+
performance and memory requirements of our regular sorting algorithm
28+
can be deterministically guaranteed.
29+
30+
References:
31+
32+
D. R. Helman, D. A. Bader, and J. Já Já . "A Randomized Parallel Sorting Algorithm With an Experimental Study," Journal of Parallel and Distributed Computing, 52(1):1-23, 1998.
33+
34+
D.R. Helman, J. Já Já , D.A. Bader. "A New Deterministic Parallel Sorting Algorithm With an Experimental Evaluation," ACM Journal of Experimental Algorithmics, 3(4):1-24, 1998.

0 commit comments

Comments
 (0)