2.5 speed comparison: exw vs. eu.ex vs. compiled eu.ex

new topic     » topic index » view thread      » older message » newer message

Just playing around a little with the new 2.5.

I was curious how much slower the Euphoria-based backend
interpreter is than the C-based backend.

On my first test, eu.ex is about 212 times slower than exw.exe.
Translating eu.ex, and then compiling with Borland C, it is only
190 times slower.  This first test is my entry for Derek's
contest.  It is both sequence and integer intensive.

My second test was a simple prime number generator.  Sequences
are not used, only integers.  In this test, eu.ex is about 17
times slower than exw.exe.  The compiled eu.ex is only about 3
times slower.

constant START = time()
integer file_no
procedure primes()
    integer n, lim, k, p, highest

    highest = 0
    n = 1
    lim = 350000   -- adjust as needed to get acceptable times

    while n < lim do
        k = 3
        p = 1
        n = n + 2
        while k * k <= n and p do
            p = floor(n / k) * k != n
            k = k + 2
        end while
        if p then
            highest = n
        end if
    end while
    ? highest
end procedure

primes()
printf(1, "Elapsed time: %f\n", time() - START)
file_no=getc(0)


The final test was sieve8k.exw, found in the demo\bench
directory.  On this test, eu.ex is about 1284 times slower than
exw.exe, and the compiled eu.ex is 1088 times slower.

new topic     » topic index » view thread      » older message » newer message

Search



Quick Links

User menu

Not signed in.

Misc Menu