To my benchmark proposal with the determinant programs
Juergen Pfitzenmaier <[EMAIL PROTECTED]> writes
> The setting for the suggested benchmark needs some clarification. For
> a dense matrix of sufficient size haskell should be able to beat C.
> I think Sergey knows this and therefor didn't mention a benchmark
> involving a sparse matrix ;)
No. I am quite naive about this. Just curious to know the average
performance ratio of Haskell to other languages.
The DoCon-2 manual (section 'pe') contains certain comparison to the
strict, non-functional system MuPAD-1.3.
But here i suggest to program in C the same mathematical method
(1 - Gauss, 2 - expansion by row) as in Haskell and compare.
I expect C to be about 8 times faster. But if it occurs 100 times, i
would have to think, what is the matter.
I also want to show that arrays are not so necessary.
The Haskell program was given.
But it is harder for me to provide the C program, do not want to
recall the C programming.
The program is for the matrices over Int-s modulo prime p.
But, probably, type C = Float will do as well.
And why the dense matrix representation is better for Haskell?
Rather i would expect it is the sparse one.
------------------
Sergey Mechveliani
[EMAIL PROTECTED]