Skip to content

[WIP][DISCUSS] Single source QFunction proposal #288

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

Closed
YohannDudouit wants to merge 131 commits into from
Closed

[WIP][DISCUSS] Single source QFunction proposal #288

Changes from 4 commits
Commits
Show all changes
131 commits
Select commit Hold shift + click to select a range
13681b3
Rescaled NS example
May 1, 2019
f7c73ca
Rescaled NS example
May 1, 2019
dca4711
Merge branch 'rescaled-ns-example' of https://github.com/valeriabarra…
May 3, 2019
9ccb6b9
Add PetscObjectSetName to set solution's vec name
May 9, 2019
70cdef1
Merge branch 'rescaled-ns-example' of github.com:valeriabarra/libCEED…
May 9, 2019
5cf694b
Rename field name Total Energy -> EnergyDensity
May 10, 2019
47868d5
Rescale time-stepper options with units of time
May 10, 2019
0cdf8f0
Some tidying
May 13, 2019
9ad4535
Add a reference non-tensor BasisApply for cuda backends.
YohannDudouit May 16, 2019
469f022
Remove useless function declaration.
YohannDudouit May 16, 2019
abe33e5
make style
YohannDudouit May 16, 2019
be9261b
Add ElemRestrictionApplyBlock
Mar 28, 2019
a765294
Add restriction by block to /cpu/self/*/blocked
Mar 28, 2019
045b9c4
Include full evec blocked backend
Mar 29, 2019
89c6efa
Use blocking in optimized serial backends
May 3, 2019
9fbf56a
Update t208-f to use offset
May 14, 2019
1f37b40
Add block paramenter example in doc
May 14, 2019
d4fd279
Merge pull request #228 from CEED/rstr-block
jeremylt May 18, 2019
a7a2e80
Use function pointers in CeedSetBackendFunction
Apr 1, 2019
6e79d47
Dynamically allocate Ceed function table
Apr 1, 2019
7e95bb9
Add all backends to Doxygen (#251)
jeremylt May 18, 2019
a17ac2e
Brief description of the cuda-shared backend in README.md
YohannDudouit May 24, 2019
db777ff
Merge pull request #255 from CEED/yohann/cuda-shared-readme
jeremylt May 24, 2019
8ffbcc6
Revert "Brief description of the cuda-shared backend in README.md" (#…
jeremylt May 24, 2019
854b039
Add description of cuda-shared backend in README.md
YohannDudouit May 24, 2019
bdc3149
Check for devel Valgrind
May 24, 2019
cb37edd
Improved documentation for SetBackendFunction
May 24, 2019
f8902d9
VecCreate -> VectorCreate
May 24, 2019
a7724da
Merge pull request #232 from CEED/offsetof-fix
jeremylt May 24, 2019
b8fb44e
Makefile: avoid leaving -.o file in valgrind check
jedbrown May 24, 2019
7cfafa9
Merge pull request #258 from CEED/valgrind-fix
jeremylt May 24, 2019
683faae
make tidy: static analysis using clang-tidy
jedbrown May 26, 2019
5904c99
travis: add clang-tidy
jedbrown May 26, 2019
d9e86a0
Merge pull request #259 from CEED/jed/clang-tidy
jeremylt May 26, 2019
aefd837
Add delegates for specific objects
Apr 29, 2019
a4999ed
Update Ceed Delegate refrencing
May 24, 2019
f29ffe7
Update Ceed delegate documentation
May 25, 2019
265be9c
Makefile: Add libceed_test, two stage testing
May 26, 2019
686de4a
Makefile: add search for testing
jedbrown May 27, 2019
1048f70
Nek examples: work around some error messages due to unset env
jedbrown May 27, 2019
c4216f2
Makefile: enable jobserver for example builds
jedbrown May 27, 2019
7e68d26
Makefile: make CEED_LIBS a deferred variable
jedbrown May 27, 2019
567e4d0
Makefile: libceed_test links to libceed; does not need source objects
jedbrown May 27, 2019
76af460
Makefile: fix libceed_test
jedbrown May 27, 2019
68ba631
README: Drop tmpl backend from list
May 28, 2019
ad9dbbe
Merge pull request #240 from valeriabarra/rescaled-ns-example
jedbrown May 29, 2019
1856ee7
Merge pull request #239 from CEED/decorator
jeremylt May 29, 2019
e183437
Update top level .gitignore
May 29, 2019
4febc4f
Merge pull request #261 from CEED/valeria/update-gitignore
jedbrown May 29, 2019
0c59ef1
Add PETSc BP2, BP4, BP5, BP6 (#243)
jeremylt May 29, 2019
f5ef5ec
OCCA Backend clang-tidy fixes
May 29, 2019
79f7945
Update postprocessing scripts
May 29, 2019
d1f4761
Re-add test_required_examples for proper title and pdf file name
May 29, 2019
bec1c03
NOLINT for OCCA tensor contract false positive
May 30, 2019
4c77ad0
Travis: clang-tidy only on linux
May 30, 2019
103dcb4
OCCA backend update note
May 31, 2019
e113e56
Merge pull request #262 from CEED/occa-clang-tidy
jeremylt May 31, 2019
918855f
Merge pull request #263 from CEED/valeria/update-postprocessingscripts
jedbrown May 31, 2019
c263cd5
Add delegation to CeedGetPreferedMemType
May 31, 2019
4bee41b
Merge pull request #264 from CEED/get-memtype-fix
jeremylt May 31, 2019
1571c99
Add -include directive to Makefile to allow import of custom build co…
Jun 2, 2019
d3c6b40
make configure: cache user variables in config.mk
jedbrown Jun 2, 2019
4b5b4ec
Typo: Error -> Error3
Jun 4, 2019
c74a5ab
Merge pull request #268 from CEED/petsc-bp-bugfix
jeremylt Jun 5, 2019
5721344
Added append option to tee for CC, FC, and NVCC
Jun 10, 2019
196a75e
Makefile: distclean removes config.mk; NVCC ?=
jedbrown Jun 10, 2019
9ef2071
Start the optimization of the Cuda restriction operator.
YohannDudouit May 17, 2019
d3232bb
Optimization of cuda-reg restriction.
YohannDudouit May 30, 2019
074be16
Optimization of weight kernel and dynamic allocation of shared memory.
YohannDudouit Jun 3, 2019
717ff8a
Minor bug fix
YohannDudouit Jun 3, 2019
4247ecf
Optimization of 2D kernels for cuda-shared backend.
YohannDudouit Jun 3, 2019
d94769d
Optimization of 1D kernels for cuda-shared backend.
YohannDudouit Jun 3, 2019
698ebc3
Optimization of 3D kernels for cuda-shared backend.
YohannDudouit Jun 3, 2019
3f63d31
Remove dead code. Cuda-reg restriction optimization.
YohannDudouit Jun 4, 2019
df4cfd6
Remove dead or unnecessary code.
YohannDudouit Jun 4, 2019
c3d0bf4
Minor documentation fix in examples/README.md
Jun 10, 2019
56f30de
Merge pull request #267 from CEED/valeria/Makefile-include
jeremylt Jun 10, 2019
a7bd39d
Fix underinterpolation mode for /cpu/self backends
Jun 10, 2019
76174be
Update 'make junit'
Jun 11, 2019
3ce8654
Improve petsc-bps benchmark output (#266)
jeremylt Jun 11, 2019
55542c5
Update README for PETSc BPs
Jun 17, 2019
59d7a4c
Update Benchmark documentation
Jun 17, 2019
650cb9e
Merge pull request #278 from CEED/petsc-bp-docfix
jeremylt Jun 17, 2019
29187ef
Merge pull request #274 from CEED/underintegrate-basis-fix
jeremylt Jun 19, 2019
194c25f
Update qextra to reflect BP paper
Jun 24, 2019
e07e9dd
Convert BP5 BP6 to GLL
Jun 24, 2019
b629e3f
Merge pull request #282 from CEED/bp5-bp6-fix
jeremylt Jun 24, 2019
afd75a3
double pointer format for the qFunction.
YohannDudouit Jun 25, 2019
3dd0010
Add CUDA_LIB_DIR_STUBS for systems that don't have CUDA drivers insta…
Jun 2, 2019
0109ba8
Minor: styling
YohannDudouit Jun 4, 2019
4a6d4bb
namespace cuda backends functions.
YohannDudouit Jun 17, 2019
961116e
make style.
YohannDudouit Jun 17, 2019
ab7ab56
ceed-cuda: resolve -Wsign-compare for CUresult (unsigned enum) in Cee…
jedbrown Jun 23, 2019
9d77422
Merge branch 'yohann/cuda-non-tensor' [PR #249]
jedbrown Jun 26, 2019
77538c7
Makefile: AVX test cannot match -mno-avx
jedbrown Jun 26, 2019
c0ea104
Makefile: MKL link flags for non-Intel compilers
jedbrown Jun 26, 2019
ece74ea
Merge pull request #284 from CEED/jed/blas-lib
jeremylt Jun 27, 2019
1226057
Merge branch 'master' into yohann/cuda-restr-opt
YohannDudouit Jun 27, 2019
0f9724d
Fix derivatives order in du
Jun 27, 2019
f51b1ba
navier-stokes: use du[3][3] to clean up indexing
jedbrown Jun 28, 2019
45dba41
Change the .qf files to have the same code as the C functions.
YohannDudouit Jul 1, 2019
09382c9
Make previous Cuda backends use .qf files.
YohannDudouit Jul 2, 2019
6620acc
Add a return value to qFunctions.
YohannDudouit Jul 2, 2019
4c01945
Make cpu backends use .qf files.
YohannDudouit Jul 2, 2019
17f141b
Minor: clean commented code.
YohannDudouit Jul 2, 2019
ec3b0e5
Add guarded math.h for petsc examples.
YohannDudouit Jul 2, 2019
1f6d0b3
Merge branch 'yohann/cuda-restr-opt' [PR #269]
jedbrown Jul 2, 2019
ea03cb9
Update CUDA/OCCA CEED_COPY_VALUES logic
Jul 3, 2019
f405f80
Merge pull request #289 from CEED/cuda-occa-copy-vals
jeremylt Jul 4, 2019
1469ee4
Add ElemRestrGetMult and tests
Jul 10, 2019
56cf2fb
Merge pull request #291 from CEED/rstr-mult
jeremylt Jul 10, 2019
e0fc044
Fix resource strcmp in xsmm backends
Jul 12, 2019
6f7d248
Update CPU backends to give default for /cpu/self/***
Jul 12, 2019
04ad072
Merge pull request #295 from CEED/cpu_priority_fix
jeremylt Jul 12, 2019
6278642
examples/navier-stokes: formatting cleanup
Jul 11, 2019
e24ee97
Merge branch 'valeria/fix-ns-derivatives' [PR #292]
jedbrown Jul 14, 2019
00742fa
examples/petsc: fix max reduction type and address
jedbrown Jul 14, 2019
01b3627
Merge pull request #296 from CEED/jed/fix-error-reduction
jeremylt Jul 15, 2019
deafd6b
tests: fix typo in t209-f
Jul 18, 2019
706bc5e
backends: fix ref backend priorities
Jul 18, 2019
3a6bb9f
Merge pull request #297 from CEED/jeremy/209-f-fix
jeremylt Jul 18, 2019
bd2096d
Merge pull request #298 from CEED/jeremy/ref-priority-fix
jeremylt Jul 18, 2019
86a4271
Update nek5000 example scripts (#241)
thilinarmtb Jul 22, 2019
01a7252
README: Update corectness for examples
Jul 22, 2019
6f67eec
Merge pull request #300 from CEED/jeremy/doc-fix
jeremylt Jul 23, 2019
241a4b8
Full jit compiled operator: cuda-gen backend (#275)
YohannDudouit Jul 25, 2019
87e11ce
Merge branch 'master' into yohann/double-ptr
YohannDudouit Jul 27, 2019
0b902cb
Remove previous nek qf files.
YohannDudouit Jul 29, 2019
f442366
Remove .cu files.
YohannDudouit Jul 31, 2019
40d6fce
Remove .qf files.
YohannDudouit Aug 2, 2019
1175c3d
Remove dead code in the tests.
YohannDudouit Aug 2, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
138 changes: 67 additions & 71 deletions examples/navier-stokes/densitycurrent.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -217,88 +217,84 @@ static int DC(void *ctx, CeedInt Q,
for (CeedInt i=0; i<Q; i++) {
// Setup
// -- Interp in
const CeedScalar rho = q[i+0*Q];
const CeedScalar u[3] = { q[i+1*Q] / rho,
q[i+2*Q] / rho,
q[i+3*Q] / rho
};
const CeedScalar E = q[i+4*Q];
const CeedScalar rho = q[i+0*Q];
const CeedScalar u[3] = { q[i+1*Q] / rho,
q[i+2*Q] / rho,
q[i+3*Q] / rho
};
const CeedScalar E = q[i+4*Q];
// -- Grad in
const CeedScalar drho[3] = { dq[i+(0+5*0)*Q],
dq[i+(0+5*1)*Q],
dq[i+(0+5*2)*Q]
};
const CeedScalar du[9] = { (dq[i+(1+5*0)*Q] - drho[0]*u[0]) / rho,
(dq[i+(1+5*1)*Q] - drho[1]*u[0]) / rho,
(dq[i+(1+5*2)*Q] - drho[2]*u[0]) / rho,
(dq[i+(2+5*0)*Q] - drho[0]*u[1]) / rho,
(dq[i+(2+5*1)*Q] - drho[1]*u[1]) / rho,
(dq[i+(2+5*2)*Q] - drho[2]*u[1]) / rho,
(dq[i+(3+5*0)*Q] - drho[0]*u[2]) / rho,
(dq[i+(3+5*1)*Q] - drho[1]*u[2]) / rho,
(dq[i+(3+5*2)*Q] - drho[2]*u[2]) / rho
};
const CeedScalar dE[3] = { dq[i+(4+5*0)*Q],
dq[i+(4+5*1)*Q],
dq[i+(4+5*2)*Q]
};
const CeedScalar drho[3] = { dq[i+(0+5*0)*Q],
dq[i+(0+5*1)*Q],
dq[i+(0+5*2)*Q]
};
const CeedScalar du[3][3] = {{(dq[i+(1+5*0)*Q] - drho[0]*u[0]) / rho,
(dq[i+(1+5*1)*Q] - drho[1]*u[0]) / rho,
(dq[i+(1+5*2)*Q] - drho[2]*u[0]) / rho},
{(dq[i+(2+5*0)*Q] - drho[0]*u[1]) / rho,
(dq[i+(2+5*1)*Q] - drho[1]*u[1]) / rho,
(dq[i+(2+5*2)*Q] - drho[2]*u[1]) / rho},
{(dq[i+(3+5*0)*Q] - drho[0]*u[2]) / rho,
(dq[i+(3+5*1)*Q] - drho[1]*u[2]) / rho,
(dq[i+(3+5*2)*Q] - drho[2]*u[2]) / rho}
};
const CeedScalar dE[3] = { dq[i+(4+5*0)*Q],
dq[i+(4+5*1)*Q],
dq[i+(4+5*2)*Q]
};
// -- Interp-to-Interp qdata
const CeedScalar wJ = qdata[i+ 0*Q];
const CeedScalar wJ = qdata[i+ 0*Q];
// -- Interp-to-Grad qdata
// Symmetric 3x3 matrix
const CeedScalar wBJ[9] = { qdata[i+ 1*Q],
qdata[i+ 2*Q],
qdata[i+ 3*Q],
qdata[i+ 4*Q],
qdata[i+ 5*Q],
qdata[i+ 6*Q],
qdata[i+ 7*Q],
qdata[i+ 8*Q],
qdata[i+ 9*Q]
};
const CeedScalar wBJ[9] = { qdata[i+ 1*Q],
qdata[i+ 2*Q],
qdata[i+ 3*Q],
qdata[i+ 4*Q],
qdata[i+ 5*Q],
qdata[i+ 6*Q],
qdata[i+ 7*Q],
qdata[i+ 8*Q],
qdata[i+ 9*Q]
};
// -- Grad-to-Grad qdata
const CeedScalar wBBJ[6] = { qdata[i+10*Q],
qdata[i+11*Q],
qdata[i+12*Q],
qdata[i+13*Q],
qdata[i+14*Q],
qdata[i+15*Q]
};
const CeedScalar wBBJ[6] = { qdata[i+10*Q],
qdata[i+11*Q],
qdata[i+12*Q],
qdata[i+13*Q],
qdata[i+14*Q],
qdata[i+15*Q]
};
// -- gradT
const CeedScalar gradT[3] = { (dE[0]/rho - E*drho[0]/(rho*rho) -
(u[0]*du[0+3*0] + u[1]*du[1+3*0] +
u[2]*du[2+3*0])) / cv,
(dE[1]/rho - E*drho[1]/(rho*rho) -
(u[0]*du[0+3*1] + u[1]*du[1+3*1] +
u[2]*du[2+3*1])) / cv,
(dE[2]/rho - E*drho[2]/(rho*rho) -
(u[0]*du[0+3*2] + u[1]*du[1+3*2] +
u[2]*du[2+3*2]) - g) / cv
const CeedScalar gradT[3] = {( dE[0]/rho - E*drho[0]/(rho*rho) -
( u[0]*du[0][0] + u[1]*du[1][0] + u[2]*du[2][0])) / cv,
( dE[1]/rho - E*drho[1]/(rho*rho) -
( u[0]*du[0][1] + u[1]*du[1][1] + u[2]*du[2][1])) / cv,
( dE[2]/rho - E*drho[2]/(rho*rho) -
( u[0]*du[0][2] + u[1]*du[1][2] + u[2]*du[2][2]) - g) / cv
};
// -- Fuvisc
// Symmetric 3x3 matrix
const CeedScalar Fu[6] = { mu * (du[0+3*0] * (2 + lambda) +
lambda * (du[1+3*1] + du[2+3*2])),
mu * (du[0+3*1] + du[1+3*0]),
mu * (du[0+3*2] + du[2+3*0]),
mu * (du[1+3*1] * (2 + lambda) +
lambda * (du[0+3*0] + du[2+3*2])),
mu * (du[1+3*2] + du[2+3*1]),
mu * (du[2+3*2] * (2 + lambda) +
lambda * (du[0+3*0] + du[1+3*1]))
};

const CeedScalar Fu[6] = { mu * (du[0][0] * (2 + lambda) +
lambda * (du[1][1] + du[2][2])),
mu * (du[0][1] + du[1][0]),
mu * (du[0][2] + du[2][0]),
mu * (du[1][1] * (2 + lambda) +
lambda * (du[0][0] + du[2][2])),
mu * (du[1][2] + du[2][1]),
mu * (du[2][2] * (2 + lambda) +
lambda * (du[0][0] + du[1][1]))
};
// -- Fevisc
const CeedScalar Fe[3] = { u[0]*Fu[0] + u[1]*Fu[1] + u[2]*Fu[2] +
k * gradT[0],
u[0]*Fu[1] + u[1]*Fu[3] + u[2]*Fu[4] +
k * gradT[1],
u[0]*Fu[2] + u[1]*Fu[4] + u[2]*Fu[5] +
k * gradT[2]
};
const CeedScalar Fe[3] = { u[0]*Fu[0] + u[1]*Fu[1] + u[2]*Fu[2] +
k * gradT[0],
u[0]*Fu[1] + u[1]*Fu[3] + u[2]*Fu[4] +
k * gradT[1],
u[0]*Fu[2] + u[1]*Fu[4] + u[2]*Fu[5] +
k * gradT[2]
};
// -- P
const CeedScalar P = (E - (u[0]*u[0] + u[1]*u[1] + u[2]*u[2])*rho/2 -
rho*g*x[i+Q*2]) * (gamma - 1);
const CeedScalar P = ( E - (u[0]*u[0] + u[1]*u[1] + u[2]*u[2])*rho / 2 -
rho*g*x[i+Q*2] ) * (gamma - 1);

// The Physics

Expand Down