|
On Mon, Jul 6, 2015 at 1:30 PM, Justin Dearing <zippy1981@xxxxxxxxx> wrote:
If you were to do a "pip install numpy" on an IBMi with the C++ compilersI don't know, but it should be easy enough to find out. ;)
installed would it:
1. Produce a working binary version of NumPy
Note that we almost certainly have to be talking about the Unix or
Unix-like compilers for PASE, such as gcc. I doubt it will do you any
good (for this case) to have ILE C++ installed.
2. Would that be the most optimized NumPy for the OS and CPU, or isI think this can largely be translated as "how well optimized is the
there a lot of tweaking that would be possible with a better compiler or
messing with the optomization options like in the case of NumPy on windows
and linux on x64 hardware?
version of gcc which runs in PASE?"
On another level, there is the question of how much overhead there is,
if any, for *everything* that runs in PASE. My mental model of PASE is
that it's an emulation layer that sits on top of the native i layer
(if not directly, then some number of layers above the native i
layer). As such, anything written for PASE is inherently less
optimized than the most optimized native code. My mental model could
well be wrong, though. Maybe there is a common low-level layer for
both what we see as native (where RPG and CL live) and PASE. If
"native" and PASE are peers/siblings, then maybe PASE executables can
indeed be as well optimized as native.
John Y.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2025 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.