commits

SHA	Message	Author	Date
e5f6a303	Update link composyx-full	MARAIT Gilles <g**t@i**r>	9 months ago
00468d60	Remove bash-completion file generation at configure time	MARAIT Gilles <g**t@i**r>	9 months ago
d100da00	Updates for release 1.4.1	MARAIT Gilles <g**t@i**r>	9 months ago
2024be63	Add test files	MARAIT Gilles <g**t@i**r>	9 months ago
2cdfc60b	Fortran interface: can now use CSR as input local matrix	MARAIT Gilles <g**t@i**r>	9 months ago
0e56b18c	Refactor ZFP_compressor to manage bitstream lifecycle and improve memory...	SYLVAND Guillaume <g**d@ir> Committed by: MARAIT Gilles <gt@i**r>	9 months ago
20f7d4e4	Add .editorconfig for indentation	MARAIT Gilles <g**t@i**r>	9 months ago
3974fef9	Add assertions for safety when using threads	MARAIT Gilles <g**t@i**r>	9 months ago
116f63bd	Context: n_cpus = 1 by default. Ensure n_cpus = 1 even if the context is not ...	Florent Pruvost <f**t@i**r>	9 months ago
fb0b4c4b	Minor fixes in Paddle and PartDenseMatrix tests	MARAIT Gilles <g**t@i**r>	9 months ago
eeddc6f1	Set CI slurm version to 23	MARAIT Gilles <g**t@i**r>	9 months ago
6dc7069e	Use latests channels	MARAIT Gilles <g**t@i**r>	9 months ago
a449a1b5	Fixes	MARAIT Gilles <g**t@i**r>	9 months ago
334a473a	Add traits to ease AbstractSchwarz type detection	MARAIT Gilles <g**t@i**r>	9 months ago
1d48dfbf	Remove a hack in AbstractSchwarz and DiagonalPrecond pcd	MARAIT Gilles <g**t@i**r>	9 months ago
2b87cbad	omp parallelization of PartDenseMatrix initialisation and of copy B->X in LQ ...	Florent Pruvost <f**t@ir> Committed by: MARAIT Gilles <gt@i**r>	9 months ago
425922a1	CastOperator: keeping an internal copy of the operator setup	MARAIT Gilles <g**t@i**r>	9 months ago
b1d1c26c	Revert commits b321026345293339c955743a2ba3724333c68362, 495828df56663f0c671d...	Florent Pruvost <f**t@ir> Committed by: MARAIT Gilles <gt@i**r>	9 months ago
556267e8	Forgot tangle	MARAIT Gilles <g**t@i**r>	9 months ago
4865c3dc	IRSolver: set reset cumul_iter at each solve (fix last commit)	MARAIT Gilles <g**t@i**r>	9 months ago
fbd1fe25	IRSolver: add flag to avoid multiple setup on the same matrix	MARAIT Gilles <g**t@i**r>	9 months ago
6e283bb0	IterativesSolver: store separately R and the true residual	MARAIT Gilles <g**t@i**r>	9 months ago
f820e3d6	Fix: forgot to tangle file	MARAIT Gilles <g**t@i**r>	9 months ago
f54d0848	Now using shared_ptr for SchurSolver	MARAIT Gilles <g**t@i**r>	9 months ago
f6a4973a	Clarify this is a copy and not a view. We cannot use a view in the context of...	PRUVOST Florent <f**t@ir> Committed by: MARAIT Gilles <gt@i**r>	10 months ago
495828df	dist_rsvd: no more need to initialize the part matrix, done in the gemm if si...	Florent Pruvost <f**t@i**r>	10 months ago
3ca52fa9	Now use new cuDSS Schur computation to driver cg ddm cuda	MARAIT Gilles <g**t@i**r>	10 months ago
b9df2b2b	Cuda Schur: add sym/herm test + factorize unit test code	MARAIT Gilles <g**t@i**r>	10 months ago
4fb1c04a	Forget test file	MARAIT Gilles <g**t@i**r>	10 months ago
59df0e6a	cuDSS add new Schur complement functions	MARAIT Gilles <g**t@i**r>	10 months ago
0581b2c8	Fix Jacobi with Eigen and Armadillo matrices	MARAIT Gilles <g**t@i**r>	10 months ago
a9bb6302	Fix Jacobi, can now be used with sparse matrices	MARAIT Gilles <g**t@i**r>	11 months ago
48c426ea	Update Changelog	MARAIT Gilles <g**t@i**r>	11 months ago
f0e00a4a	Add codemeta	MARAIT Gilles <g**t@i**r>	11 months ago
c7afc3e7	Moving ComposyxKernels -> experimental_RSVD	MARAIT Gilles <g**t@i**r>	11 months ago
b8cb60f9	Installing composyx_cuda_kernels	MARAIT Gilles <g**t@i**r>	11 months ago
70baa84a	Fix warnings GCC + update submodules	MARAIT Gilles <g**t@i**r>	11 months ago
104263f1	Update README	MARAIT Gilles <g**t@i**r>	11 months ago
d19f0d62	ci: fix gcovr call, due to Catch2 source file location mismatch	Florent Pruvost <f**t@i**r>	11 months ago
b3210263	PartDenseMatrix: wrap existing Chameleon matrix with a view.	Florent Pruvost <f**t@i**r>	11 months ago
84578556	Clean repository and documentation	MARAIT Gilles <g**t@i**r>	11 months ago
ca42e12f	SchurSolver / ImplSchur -> can now use Sparse Kgg for precond	MARAIT Gilles <g**t@i**r>	11 months ago
4b5fec17	driver cuda cg ddm: can now read params from file	MARAIT Gilles <g**t@i**r>	11 months ago
ef5a7f55	Fixes solve multi-RHS Mumps&Pastix, + Schur Computation	MARAIT Gilles <g**t@i**r>	11 months ago
4eb90197	merge Merge	MARAIT Gilles <g**t@i**r>	11 months ago
03d9ab9f	Putting back files removed by error	MARAIT Gilles <g**t@i**r>	11 months ago
8bba2fe5	Finish code factorization	MARAIT Gilles <g**t@i**r>	11 months ago
645bba8e	Fixes + finish factorizing Schur functions	MARAIT Gilles <g**t@i**r>	11 months ago
3f816e1f	Add slicing for CudaVec and Vector, factorize some Schur code	MARAIT Gilles <g**t@i**r>	11 months ago
91d9d2fa	Adding AS PCD on Kgg for cuda cg ddm with impl. Schur	MARAIT Gilles <g**t@i**r>	12 months ago
a16e7e46	ImplicitSchur cg ddm cuda (pcd 0 for the moment)	MARAIT Gilles <g**t@i**r>	12 months ago
6b0b66bd	Chameleon: fix call to CHAMELEON_InitParComm	Florent Pruvost <f**t@i**r>	12 months ago
86727cda	Remove debug output	MARAIT Gilles <g**t@i**r>	12 months ago
92c66d96	Fixes to use ImplicitSchur with geneo (driver_cg)	MARAIT Gilles <g**t@i**r>	12 months ago
1683436e	restore chameleon.initialize with comm to see if it fixes test	Florent Pruvost <f**t@i**r>	12 months ago
851e6fee	Chameleon: CHAMELEON_InitParComm gives a wrong MPI binding, restore to simple...	Florent Pruvost <f**t@i**r>	12 months ago
573a94c3	Factorize some code + add ImplictSchur back in driver_cg	MARAIT Gilles <g**t@i**r>	12 months ago
e46e81df	CudaSparseSolver -> multirhs solve now solves rhs one by one	MARAIT Gilles <g**t@i**r>	12 months ago
217bc88f	Clarify interface for schur computation (CPU and CUDA)	MARAIT Gilles <g**t@i**r>	12 months ago
08ccfdd9	Fix CudaSchur (cuDSS multi-RHS unreliable)	MARAIT Gilles <g**t@i**r>	12 months ago
e1baa95b	Add a C++ driver dist_rsvd to test and time rsvd distributed, with chameleon ...	Florent Pruvost <f**t@i**r>	12 months ago
af05f162	TestMatrix: upgrade matrix generation	MARAIT Gilles <g**t@i**r>	12 months ago
c742208c	PartDenseMatrix: add a method to return the transposed matrix, only for fixed...	Florent Pruvost <f**t@i**r>	12 months ago
f65c9b44	Add Composyx kernels' RSVD and its benchmark	FOUSSAC Mathis <m**c@ir> Committed by: MARAIT Gilles <gt@i**r>	12 months ago
0e5edae0	Various fixes for CUDA Schur version	MARAIT Gilles <g**t@i**r>	12 months ago
d2bab0e9	PartDenseMatrix: add a check if blocks size is regular, required with chameleon	Florent Pruvost <f**t@i**r>	12 months ago
9c1ce466	PartDenseMatrix: glob_m -> get_n_rows, glob_n -> get_n_cols	Florent Pruvost <f**t@i**r>	12 months ago
9774dfef	restore demo	Florent Pruvost <f**t@i**r>	12 months ago
53e84604	Fix: I/O domain decomposition was missing an index	MARAIT Gilles <g**t@i**r>	about 1 year ago
9ef12838	Fix: I/O domain decomposition was missing an index	MARAIT Gilles <g**t@i**r>	about 1 year ago
f4b908a9	Fix/hack for Z gemv bug	MARAIT Gilles <g**t@i**r>	about 1 year ago
274abb2c	ChameleonKernels: support PartDenseMatrix	Florent Pruvost <f**t@i**r>	about 1 year ago
d04933df	CUDA: ASG+ / S (and ASGd / S wip)	MARAIT Gilles <g**t@i**r>	about 1 year ago
7c144a68	Add generalized SVD (smallest eval) with cuSOLVER	MARAIT Gilles <g**t@i**r>	about 1 year ago
15c2ae40	Chameleon: fix chameleon_getldd, we do not need to depend on PartDenseMatrix ...	Florent Pruvost <f**t@i**r>	about 1 year ago
986e903b	Revert "Distributed QR/LQ work only with -DCHAMELEON_COPY_DIAG=OFF"	Florent Pruvost <f**t@i**r>	about 1 year ago
5e230c21	Add operations for RR/S	MARAIT Gilles <g**t@i**r>	about 1 year ago
65941dd6	Distributed QR/LQ work only with -DCHAMELEON_COPY_DIAG=OFF	Florent Pruvost <f**t@i**r>	about 1 year ago
e17d644c	Add AS/S in cuda_cg_ddm	MARAIT Gilles <g**t@i**r>	about 1 year ago
c3127d7a	Update driver cuda_cg_ddm to handle Schur (no pcd yet)	MARAIT Gilles <g**t@i**r>	about 1 year ago
849f850e	Fix CudaPartSchurSolver: Subdomain::partofunity	MARAIT Gilles <g**t@i**r>	about 1 year ago
81f94d4b	Fixes for CudaPartSchurSolver + many cuda impovements	MARAIT Gilles <g**t@i**r>	about 1 year ago
3483b4bd	Add sym facto for CudaDenseSolver	MARAIT Gilles <g**t@i**r>	about 1 year ago
e19dea10	fix: add correct template defaults for the dense lq and qr facto	asolmele <a**c@g**m>	about 1 year ago
1e8a4338	Adding CudaDenseSolver + fixes	MARAIT Gilles <g**t@i**r>	about 1 year ago
6e63c4fe	CudaSchurSolver -> externalize functions	MARAIT Gilles <g**t@i**r>	about 1 year ago
c0935efe	Adding permutation kernel files	MARAIT Gilles <g**t@i**r>	about 1 year ago
cc335300	composyx_cuda_cg -> can now work on Schur complement	MARAIT Gilles <g**t@i**r>	about 1 year ago
448c7dd9	Add kernels for vector permutation	MARAIT Gilles <g**t@i**r>	about 1 year ago
1bb588d4	Fix CudaSchurSolver	MARAIT Gilles <g**t@i**r>	about 1 year ago
c297a055	Add CUDA solver on the Schur complement (WIP)	MARAIT Gilles <g**t@i**r>	about 1 year ago
3ff70130	Move cuda cg ddm to driver/C++ dir	MARAIT Gilles <g**t@i**r>	about 1 year ago
139a62f9	cuda_ddm: add ASGd	MARAIT Gilles <g**t@i**r>	about 1 year ago
268f7c4f	Fixes 2lvlAS, driver_cg, CudaVec ET	MARAIT Gilles <g**t@i**r>	about 1 year ago
4c5547ef	Fixes	MARAIT Gilles <g**t@i**r>	about 1 year ago
f32c191f	Adding CUDA gemm expression templates + fixes	MARAIT Gilles <g**t@i**r>	about 1 year ago
7863fb3e	Expression templates cuda GEMV (dense)	MARAIT Gilles <g**t@i**r>	about 1 year ago
530c325b	Pastix distributed: matrix A is centralized on rank 0 on entry. Hence we requ...	Florent Pruvost <f**t@i**r>	about 1 year ago
2c6c98bc	Fix centralized solver for ASG+	MARAIT Gilles <g**t@i**r>	about 1 year ago
ba019526	Fixes in CudaDense and PartMatrix centralize CSR	MARAIT Gilles <g**t@i**r>	about 1 year ago

← Back to repository

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Commits

gitlab.inria.fr / composyx/composyx / commits