Thank you for the quick answer, I was partly wondering how to get his version without building a new julia but I should have guess that this was how it's done. So now, let's see and bench.
And I'm still open to suggestions, remarks for the future benchmarks that I will do. Thanks Le lundi 31 août 2015 11:55:25 UTC+2, Mauro a écrit : > > If you are enquiring about the git-mechanics to get that branch built, > see below. (Otherwise ignore.) > > You can build that branch. The last rebase is on branch > yyc/dftnew_rebase, but it was not tested. This "should" work: > > git clone [email protected]:JuliaLang/julia.git > git checkout yyc/dftnew_rebase > make testall (the testall will run all the unit-tests after compilation) > > Alternatively you can build Steven's branch (in another folder): > > git clone [email protected]:stevengj/julia.git > git checkout dftnew > make testall > > Mauro > > > On Mon, 2015-08-31 at 11:22, John leger <[email protected] <javascript:>> > wrote: > > Hello everyone, > > > > I am currently playing with Julia and OpenCL, so far everything is good. > I > > would like to bench: FFT in Pure Julia, the fftw wrapper, the opencl FFT > > deployed on GPU and CPU (so 4 benchmarks). > > There is already everything ready except the pure julia fft. I have > found > > some information about it: https://github.com/JuliaLang/julia/pull/6193 > but > > I don't know how I could use it. > > Do you have any ideas how I could test it ? > > > > Also, maybe someone already made some benchmarks, especially on the > opencl > > GPU/CPU part. So if you have any information, thank you. > > > > Jonathan. > >
