-
Notifications
You must be signed in to change notification settings - Fork 130
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Major improvements (discussion) #37
Comments
Hi,
Yes, moving it to Python 3 makes sense, it should be fairly easy, the only
stumbling block is PySparse, which is not being maintained (last time I
checked).
If one can find a solver that works as fast PySparse (which is just a
wrapper for SuperLU and some Iterative solvers) then it should be fairly
quick and painless to do (CHOLMOD?). ToPy's matrices are always Positive
Definite so the solver one chooses should exploit this property.
William
…On Mon, 28 May 2018 at 16:59, V-Kh ***@***.***> wrote:
Hey folks
I've been thinking for a while about possible things that can be done.
However, there are some problems that have to be discussed.
First, I think that the project has to be moved to Python 3 since Python 2
faces end-of-life in 2020. You probably know that supporting legacy
software is sometimes extremely painful.
Second, we need to discuss cross-platform features of the package. I'm a
Linux user so, to be honest, I do not care about other systems. However, I
know that certain researchers use Windows or MacOS and do not want to learn
new stuff. So, something has to be done in this case.
V.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#37>, or mute the thread
<https://github.com/notifications/unsubscribe-auth/ANktUZetwsVsIP9uNbFdJdoUWl5B9M9Eks5t3BC3gaJpZM4UQUIO>
.
|
Any updates on this issue? Pysparse seems to be causing me some issues, when installed using pip, I get the following error:
And when using the conda installer, I get the following error when running
Any ideas? EDIT: This fixed the conda install |
Scipy seems to be the obvious candidate. Together with numpy they form the de facto standard startup for scientific computations in python. According to https://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.sparse.linalg.splu.html then it also uses the same backend: that is SuperLU |
When I started working on ToPy (more than 10 years ago now) I initially did
use SciPy, but the performance was extremely poor compared to PySparse.
It's most proabaly better by now, I'd hope.
You're welcome to make the change, you only need to recode the FEA part :-)
Anaconda is another option I may look into in future.
William
…On Mon, 6 May 2019 at 13:10, Kjetil Andre Johannessen < ***@***.***> wrote:
Hi, Yes, moving it to Python 3 makes sense, it should be fairly easy, the
only stumbling block is PySparse, which is not being maintained (last time
I checked). If one can find a solver that works as fast PySparse (which is
just a wrapper for SuperLU and some Iterative solvers) then it should be
fairly quick and painless to do (CHOLMOD?). ToPy's matrices are always
Positive Definite so the solver one chooses should exploit this property.
William
Scipy seems to be the obvious candidate. Together with numpy they form the
de facto standard startup for scientific computations in python. According
to
https://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.sparse.linalg.splu.html
then it also uses the same backend: that is SuperLU
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#37 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ADMS2UPCQ4YBBNGEPKOOSXDPUAGYRANCNFSM4FCBIIHA>
.
|
Hey folks
I've been thinking for a while about possible things that can be done. However, there are some problems that have to be discussed.
First, I think that the project has to be moved to Python 3 since Python 2 faces end-of-life in 2020. You probably know that supporting legacy software is sometimes extremely painful.
Second, we need to discuss cross-platform features of the package. I'm a Linux user so, to be honest, I do not care about other systems. However, I know that certain researchers use Windows or MacOS and do not want to learn new stuff. So, something has to be done in this case.
Finally, we can add the package to the official repository to make its installation easier.
I would like to discuss this with you before starting anything new.
V.
The text was updated successfully, but these errors were encountered: