GPU Simulation POC - #7018
Draft
multitalentloes wants to merge 3 commits into
Draft
Conversation
multitalentloes
added
the
manual:irrelevant
This was referenced Apr 29, 2026
Closed
multitalentloes
force-pushed
the
full_gpu_sim_v1
branch
2 times, most recently
from
May 6, 2026 08:33
7a1a612 to
249dca8
Compare
This was referenced May 6, 2026
multitalentloes
force-pushed
the
full_gpu_sim_v1
branch
from
May 6, 2026 14:20
27fbe96 to
9f18d9e
Compare
multitalentloes
force-pushed
the
full_gpu_sim_v1
branch
from
May 28, 2026 08:58
49a1089 to
9ff8aa9
Compare
multitalentloes
force-pushed
the
full_gpu_sim_v1
branch
2 times, most recently
from
June 5, 2026 08:26
039d722 to
9635597
Compare
multitalentloes
force-pushed
the
full_gpu_sim_v1
branch
from
August 6, 2026 07:56
9635597 to
f4a4a5d
Compare
misc improvements move files remove unused code support nv and amd gpus simultaneously also remove some dead comments rename gpu version of fvbaseelementcontext deduplicate code in compiled files minor fix GPU assembly support on AMD and CUDA restructure and format reduce tpfalinearizer diff with new headerfiles improve template argument orders use less template arguments clang-format new files and tpfalinearizer ensure compilation on CPU without warnings rename files and structs avoid extra BOIQ copy in kernel avoid extra BOIQ copy in getter extract source terms to separate header Co-authored-by: Atgeirr Flø Rasmussen <atgeirr.rasmussen@sintef.no> improve GPU assembly implementation update gpusparsematrix test create new tpfalinearizerstructs file extract what must be visible from both gpuparams and tpfalinearizer itself remove unused function and move comment to correct place remove unused code Simpler and renamed accessor. protect flow_gpu_main from being compiled on all systems use references instead of ptr improve boiq ctor pattern make refs const refs improve bc computation and change cmake simplify copy_to_gpu and remove alugrid from .cu reduce diff & add extra dune undef fix formatting in newtranfluxmodule make preprocessor statements more precise if you have cuda but do not wish to use the gpu assembly then stuff was included that did not make sense, that is hopefully now resolved fix rebasing issues remove nullfvbaseelementcontext.hh improve cmake structure remove timing code in IQ dispatcher remove more timing code and implement simplifcations remove extra test file make linearize access safer extra if check makes norne work locally add includes remove dead line of code make test compilable w hipcc change default, add tests update tests remove unneeded decorator
use uniqueptr more update docs simplify test docs
multitalentloes
force-pushed
the
full_gpu_sim_v1
branch
from
August 6, 2026 09:08
f4a4a5d to
5f844f4
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
New PR including work on GPU property computation that takes over for #6612.
This PR currently contains the code needed to run properties and matrix assembly for SPE11 cases (Gas+Water+Thermal) on the GPU, the two main components of a basic non-linear solver besides the linear solver which is already implemented.
My plan is to extract parts of the diff gradually in smaller PRs to get it merged.
Marked as irrelevant for the manual as this should be further improved upon and validated to be robust, as well as supporting a broader set of cases first.