[GSoC] 3rd coding week blog post

[Sarah Berkemer]

Hey,
I will write this in more detail in the blog, too.
For the plotting, I use a tool called hp2pretty. I used hp2ps but I think
that there is an error in the hp2ps tool, since I got plots with negative
memory values which doesn't make sense in this case. So plotting the same
.hp file with hp2pretty looked much better.
You can also ask Christian about the error in the hp2ps, he may know more
about it than me.


On Tue, Jun 3, 2014 at 1:28 PM, Ketil Malde <ketil at malde.org> wrote:

>
> > And yes, I did the plots with this program and the Haskell profiling
> library :) .
>
> >> This looks very cool so far. Can you point me to the source code for the
> >> script transalign_prof, and was that the same program that generated
> >> those nice plots on the wiki?
>
> Just to fill in a bit - the memory profiling (i.e. plots) is a feature
> of the GHC compiler, which generates executables that can dump run-time
> statistics on heap use at various time points to a text file.  Then,
> there's a tool (called 'hp2ps') that graphs this.
>
> (The figures on the blog looks slightly different from standard hp2ps,
> so either Sarah is using a newer version, or she found an alternative
> way to plot them.)
>
> -k
> --
> If I haven't seen further, it is by standing in the footprints of giants
>