gmpbench -- how to utilize all CPU cores?

Steffen Brinkmann sitiveni at web.de
Sun Sep 29 15:53:00 CEST 2013


Hello Torbjorn,

I guess I have to clarify my statements above... Before all: I didn't 
mean to complain or being rude. I am aware of the amount of work and 
well-written code that you give away for free because you want to. And I 
appreciate that.

On 29.09.2013 15:31, Torbjorn Granlund wrote:

> Thanks to the latest flurry of messages, we know that "serious users"
> will never use GMP due to GMP's lack of internal parallelism.  Oh my,
> there are lot of un-serious users out there!
My point here was (and I admit that this was not clear from my 
statement) that a library /intending to be parallel/ must be shown to 
scale well in order to be used by serious users. If a library is 
intentionally left serial (or to the largest extend so), it can still be 
a "serious" piece of code. And GMP cerainly is.

In a private discussion, Tor gave me some insight on why GMP should 
remain serial. I see that this is a well-informed decision and respect 
that.

That said, I think it would be nice to have a parallel library /on top 
of GMP/ for scientific purposes. But that is a different project and not 
to be discussed here.

> PS. Please remember that GMP is a gift from us to you.  Complaining
> about it in an quite ignorant way, adding a degree of rudeness, might not
> trigger us to give you a better gift.
>
Again, my apologies, if my comments came across rude.

Cheers, Steffen



More information about the gmp-discuss mailing list