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

DO NOT SQUASH Absorb FInAT and GEM #104

Merged
merged 923 commits into from
Dec 6, 2024
Merged

DO NOT SQUASH Absorb FInAT and GEM #104

merged 923 commits into from
Dec 6, 2024

Conversation

connorjward
Copy link

@connorjward connorjward commented Dec 4, 2024 •
edited
Loading

TODOs

  • Fix testing
  • Fix docs
  • Fix install

Goes with firedrakeproject/firedrake#3904

sv2518 and others added 30 commits April 30, 2020 16:19
@sv2518 @wence-
gem: New Inverse and Solve nodes
fbaf67a 
These will be used for shape-based operations on tensors to allow
compilation of Slate via GEM. Additionally, expose FlexiblyIndexed in
the public API.
@sv2518 @wence-
gem: Allow FlexiblyIndexed to apply to any shaped thing
82ed173
@wence-
gem: Avoid complex warnings when initialising Literal array
47e2357
@FAznaran
AWc had been failing to chop off the AWnc constraint dofs.
5b60d80
@FAznaran
AWc now gets constructed, but diverges.
afcf70a
@wence-
gem: Teach refactoriser about conj/real/imag
3876090 
Fixes #166.
@wence-
Merge pull request #212 from firedrakeproject/wence/complex
4ae1782 
wence/complex
@rckirby
WIP: aw.py is still wrong.
9fbef27
@rckirby
WIP: AW elements
dc21ed8
@rckirby
update scaling in W for aWc
e1c96f6
@wence-
impero: Add option to avoid ReturnAccumulate nodes
aade71d 
Will be used in new pointwise expressions compiler. This avoids a case
where an indexsum accumulates directly into an output variable. This
is only safe if the output tensor is zero on entry (which need not
be the case for pointwise expressions).
@wence-
Merge pull request #216 from firedrakeproject/wence/feature/return-ac…
bfde63d 
…cumulate

wence/feature/return accumulate
@sv2518
Allow for ComponentTensor(NotAFlexiblyIndexed()) and Inverse/Solve in…
fd14cbf 
… decompose_variable_view, so that this can be reused for generating Slate blocks.
@rckirby
Update scaling in AWnc element
aa8ac4f
@wence-
Merge pull request #214
05f3501 
Allow for ComponentTensor(NotAFlexiblyIndexed())
@rckirby
WIP AW
ca0244a
@FAznaran
The AWc transformation now seems to be working.
dd85cfd
@rckirby
Plausible rescaling of aw dofs
5476708
@rckirby
Update AW dof scaling to improve conditioning
a1ec31e
@rckirby
Fix vertex transformations in AW
bf9b644
@rckirby
put back AW
5c3b55b
@rckirby
Merge remote-tracking branch 'origin/master' into directSerendipity
1b78e90
@rckirby
WIP: general order
e615fcf
@rckirby
Generalize direct serendipity to general order
784ef00
@rckirby
Remove unneeded file
cc0a129
@rckirby
Clean up some flake8
75056d4
@rckirby
Address most of Lawrence's comments. Need to figure out geometric_dim…
1145909 
…ension still.
@wence-
enriched: Return subelement for single argument
700f167 
Catches a case where someone built an enriched element with only a
single subelement, which subsequently dies in tree_map(max, ...)
because that implicitly relies on there being at least two
arguments (such that mapping max works).

Fixes firedrakeproject/firedrake#1762.
@rckirby
fix a bug
48e8926
@rckirby
Add helper function; fix typos
6a06450
pbrubeck and others added 16 commits July 17, 2024 14:10
@pbrubeck
HCT: basis transformation for hierarchical edge functions (#131)
0b60daa
@rckirby @pbrubeck
Add QuadraticPowellSabin{6|12} (#132)
660bda0 
Co-authored-by: Pablo Brubeck <[email protected]>
@pbrubeck
RestrictedElement: fixes for prisms (#133)
58e4c5c
@pbrubeck
finat.ufl.RestrictedElement: support restriction_domain="reduced" (#135)
1990cb6
@JDBetteridge
Use md5 to hash FiniteElementBase (#134)
673f743
@pbrubeck
sympy2gem: if-then-else (#137)
014f246
@pbrubeck
Add physically mapped Stokes elements (#139)
2c5b662
@pbrubeck
Guzman-Neilan H1(div) macroelement (#140)
0381989
@ksagiyam
Ksagiyam/hex interior facet (#138)
c37a93d
@ksagiyam
generalise VariableIndex and FlexiblyIndexed (#317)
348569f 
hex: enable interior facet integration
@mscroggs @pbrubeck
Update UFL element interface (#118)
e3b9bbe 
Co-authored-by: Pablo Brubeck <[email protected]>
@pbrubeck @FAznaran
Allow HHJ in 3D (#142)
50bc36f 
Co-authored-by: Francis Aznaran <[email protected]>
@ksagiyam
gem: attach dtype to every node (#327)
f4c4866
@pbrubeck
Use static Literal for faster code generation (#146)
14b757f
@connorjward
Merge remote-tracking branch 'finat/master' into connorjward/add-fina…
358aab4 
…t-gem
@connorjward
Merge remote-tracking branch 'gem/master' into connorjward/add-finat-gem
4746c81
@connorjward connorjward mentioned this pull request Dec 4, 2024
Merged
@connorjward connorjward marked this pull request as ready for review December 5, 2024 11:04
@connorjward connorjward force-pushed the connorjward/add-finat-gem branch from 67fa35e to 7c725d8 Compare December 5, 2024 13:25
JDBetteridge
JDBetteridge reviewed Dec 5, 2024
View reviewed changes
.github/workflows/docs.yml Outdated Show resolved Hide resolved
pyproject.toml Show resolved Hide resolved
@pbrubeck
Copy link

pbrubeck commented Dec 5, 2024

This looks great! Only one request: Could CI test FIAT and FInAT separately?

@connorjward
Copy link
Author

This looks great! Only one request: Could CI test FIAT and FInAT separately?

It could, but why is that important to have? Then we need to duplicate various bits of installation logic. The CI isn't exactly slow.

@pbrubeck
Copy link

pbrubeck commented Dec 5, 2024

This looks great! Only one request: Could CI test FIAT and FInAT separately?

It could, but why is that important to have? Then we need to duplicate various bits of installation logic. The CI isn't exactly slow.

I am mostly concerned about the ordering of the tests. Right now, FInAT is tested before FIAT. Since FInAT depends on both GEM and FIAT, could we not have it at the end?

@connorjward
Copy link
Author

This looks great! Only one request: Could CI test FIAT and FInAT separately?

It could, but why is that important to have? Then we need to duplicate various bits of installation logic. The CI isn't exactly slow.

I am mostly concerned about the ordering of the tests. Right now, FInAT is tested before FIAT. Since FInAT depends on both GEM and FIAT, could we not have it at the end?

No problem. I will change it.

@connorjward connorjward force-pushed the connorjward/add-finat-gem branch from 18e30f3 to 39247cb Compare December 5, 2024 20:17
@JDBetteridge
Copy link
Member

You can always add an if: always to the GH action script to force the tests to run even if the step before fails

@connorjward connorjward merged commit acbd449 into master Dec 6, 2024
8 checks passed
@connorjward connorjward deleted the connorjward/add-finat-gem branch December 6, 2024 13:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@JDBetteridge JDBetteridge JDBetteridge left review comments

@dham dham dham approved these changes

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

Successfully merging this pull request may close these issues.