keffer@blake.acs.washington.edu (Thomas Keffer) (06/30/89)
I have written a vector package for scientific computing. Right now
it includes all standard arithmetic operations (+, *, /, -, --, etc.)
for types double, float, and complex, and most math functions (sin,
cos, sum, cumsum, mean, variance, etc.). It includes three FFT
"server" classes:
1) DComplexFFTServer which can transform a complex vector
into a complex vector (and back).
2) DoubleFFTServer which transforms a real vector into a complex
conjugate even vector (and back).
3) DCosineFFTServer which performs sine and cosine transforms. That
is, it can transform either a real even or a real odd vector into a
real even or odd vector (respectively).
In the future, I will be extending the classes to include matrix
operations, including a C++ interface to LINPACK.
The main goal of the vector classes is to be as simple as possible
while not sacrificing (too much) in the way of speed. Others (in
particular, Doug Lea) are working on more exotic vector classes that
avoid the use of temporaries by building expression trees at runtime
and then optimizing them out. No doubt these will be fast, but they
will also be very complicated, big, and hard to extend.
The classes here do not avoid using temporaries, but use reference
counting to minimize the time required for their construction.
They will compile using either the GNU g++ compiler, or the AT&T
"cfront" compiler (no mean trick!).
They can be obtained via anonymous ftp from:
sperm.ocean.washington.edu (128.208.2.7)
(that's sperm as in the whale, not,....oh, never mind). The file you
want is pub/RWVector.tar.Z.
-tk
---
Dr. Thomas Keffer | Internet: keffer@sperm.ocean.washington.edu
School of Oceanography | BITNET: keffer%sperm.ocean.washington.edu@UWAVM
Univ. of Washington, WB-10 | uucp: uw-beaver!sperm.ocean.washington.edu!keffer
Seattle, WA 98195 | Telemail: T.KEFFER/OMNET
(206) 543-6455