Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Feature CUDA compatability #228

Closed
wants to merge 75 commits into from
Closed

Feature CUDA compatability #228

wants to merge 75 commits into from

Conversation

RobertPincus
Copy link
Member

This PR introduces interfaces without implementations for the computational kernels. This allows the use of kernels in other languages that follow the C interfaces. We test this with the CUDA kernels developed by @Chiil and his collaborators in the Netherlands.

@RobertPincus
First steps towards adding CI for Fortran frontend + CUDA kernels
c305133
@RobertPincus
Use existing branch in own repo to get CUDA kernels
2134007
@RobertPincus
YML syntax
f76a29d
@RobertPincus
YML syntax
ccef891
@RobertPincus
matrix
c61e9b2
@RobertPincus
No mapping
43f2cb5
@RobertPincus
Where the code at?
9ba1a0e
@RobertPincus
Over there?
fc517f5
@RobertPincus
Maybe it's here...
2c603bd
@RobertPincus
No, I mean here
5311508
@RobertPincus
Syntax
36777eb
@RobertPincus
Over there?
6e7d22b
@RobertPincus
Different checkout actions keyword
10247cc
@RobertPincus
... and now the location
f738782
@RobertPincus
Modules
5a5841f
@RobertPincus
Module, flags
237b11f
@RobertPincus
Flags?
a289673
@RobertPincus
Revisit Fortran flags used in CUDA compilation
24c42fd
@RobertPincus
Build combo and separate libraries by default
e6dfc4b
@RobertPincus
Remove netCDF C dependencies, allow Makefile vars to be overridden by…
2066f70 
… environment variables
@RobertPincus
NCHOME is gone; -L$(NFHOME) in Makefiles remains because of custom in…
401d49b 
…stallation location in containers
@RobertPincus
Get all arrays on device before calling kernels
067d831
@RobertPincus
Arrays in ty_optical_props on device
9d8041a
@RobertPincus
Make that all the arrays
9d7875b
@RobertPincus
Delete device memory on finalization in ty_optical_props
cecd8fd
@RobertPincus
Some inquiry functions use pointers not array copies
b4e88e9
@RobertPincus
Missed one
a2b201e
@RobertPincus
Change build order, use CUDA kernels where possible
ea5b3f9
@RobertPincus
Russa frussa YML syntax
36593d0
@RobertPincus
Using CUDA kernels needs OpenACC enabled for Fortran side
5926693
@RobertPincus
Copy link
Member Author

@Chiil I wonder what you think about merging this to either develop or main when the CUDA kernels don't pass all the tests. I could make the CUDA kernel tests as experimental?

@RobertPincus
Copy link
Member Author

Closing until working

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@Chiil Chiil Awaiting requested review from Chiil

@skosukhin skosukhin Awaiting requested review from skosukhin

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@RobertPincus