Re: Threads vs. OpenMP in PSPP.

[Ben Pfaff]

"John McCabe-Dansted" <address@hidden> writes:

> On 4/2/07, Ben Pfaff <address@hidden> wrote:
>> If we implemented all of these, or even just through #3, I bet
>> PSPP would be one of the best things out there for dealing with
>> large amounts of data.
>
> Another approach would be to split of the code that thats the majority
> of the cpu time and any or all of the following
> 1) Compile that code with -march="" and link in the code for the
> appropriate CPU.
> 2) Create a benchmark and
> 2a)  Tune the compiler optimisations for each algorithm using Acova.
> 2b)   Pick the compiler that generates the fastest code for that algorithm.

Sure.  This is a lot easier than the algorithmic and structural
improvements than I'm suggesting.  Of course, we can do both.

We don't have any good benchmarks (yet).  Perhaps we can ask
users to submit analyses that in their opinions run too slowly.

I don't think compiler optimizations will help much for analyses
that bottleneck on disk activity.  For big data sets that's
likely to occur.
-- 
Ben Pfaff 
http://benpfaff.org