It may be that I'm missing something, but you appear to be computing different things... Zi = 2.0*Zr*Zi + Ci; Zr = Tr - Ti + Cr; vs Z = Z*Z + C; No doubt the 2.0 makes a difference... There are several minor optimizations also that might speed up your code... if you're interested, email me offline. Brian On 3/3/06, Greg Buchholz <greg@xxxxxxxxxxxxxxxxxxxx> wrote: > Brian Budge wrote: > > Which command line arguments are you using? > > Mostly "-O3", but I've tried lots of combinations of different > options and I haven't yet found the magic incantation. > > >Can you make use of sse2? > > Well, there is a C version that I wrote which uses gcc's vector > extensions and SSE2... > > http://shootout.alioth.debian.org/gp4/benchmark.php?test=mandelbrot&lang=gcc&id=3 > > ...but right now I'm more interested in why normal complex<double>'s are > so slow. > > Greg Buchholz > >
