Fixed mat_seqaij copy down in matassemblyend_seqaij for cuda. Also changed an archive flag in builder.py so that it works on sections of the library with a different rootdir.

Hg-commit: 2009a1c3989

Fixed mat_seqaij copy down in matassemblyend_seqaij for cuda. Also changed an archive flag in builder.py so that it works on sections of the library with a different rootdir.

Hg-commit: 2009a1c398929c2204f8c6dd468ee4de9c33219d

show more ...

made changes to make VecView use CUDA copy up, a few Vec_Seq functions use correct form of VecGetArray. Added function to 'allocate' GPU memory without copying down for Vec_seq

Hg-commit: 90375c35f

made changes to make VecView use CUDA copy up, a few Vec_Seq functions use correct form of VecGetArray. Added function to 'allocate' GPU memory without copying down for Vec_seq

Hg-commit: 90375c35fa92e90e4031e89ec6ff43c897caa6d6

show more ...

edited the new example for ease of use

Hg-commit: 4cd6ddbc026e6dda37e3349ec73b009058e0d28e

moved the CUDA matrix multiplication to the correct conditional in aij.c
made it so the cu->c extension converter simply toggles

Hg-commit: 85b4866b0a0baa2466af3973b9ab74d7f23b41b9

Added script to change .cu files to .c and tweaked rules/variables so .cu files can be "make"-d.

Hg-commit: a7d7557e0e1f2445697081949b503850aa82b9e7

removed many unused variables
small amount of reformating
added use of all arguments in MPIUNI call to prevent complaints about unused values

Hg-commit: f1e85e77153d4207c8ab2522ae8e9ec72b7358ec

commit after merge

Hg-commit: 7a886a3525f2eaae91d3e29d0db757ab8fe14c71

added CUDA matrix on GPU for seqaij matrices

Hg-commit: 3fa202ee7087184dca6e7950c52921bbebbf0f38

Added seq dense support for new matload interface

Hg-commit: 32aa5dfc4949422f382f057cd2f3139a16dd4d6e

Added SeqBAIJ implementation for new MatLoad interface

Hg-commit: 08365da5a42c5aa8527c17effdb8cb0f6ef4604a

commit after merge

Hg-commit: ac1adcc66e9cf4bb0adb5d3d9b72b96bef9970a0

Added MPIAIJ implementation for new MatLoad interface.
Small bug fix in VecLoad interface.

Hg-commit: 468674f220b2f370b25aa95c5be79f344669779f

switched c++ style comments to c-style comments

Hg-commit: cf0e387fae4d78d9e295ea00246c5277a3d6ae8d

Started work on upgrading MatLoad()

Hg-commit: e759be04090b2041f169a94773bc54b62c24ff11

commit after merge

Hg-commit: 8f08ad9f64d125fb53316ec5474fbfc0e8ec6eb9

created cubuilder.py file to make symlinks with extension .cu to all .c files for use with nvcc for cuda

Hg-commit: e5465f2536b8efde45ebfa6ff6388d80c6d0faa6

three variables with bad const issues

Hg-commit: 167b35b9f2cf6d95f1d3c227aeaa3ca038b67448

changed many VecGetArray() to VecGetArrayRead()

Hg-commit: d7f64a5632750c1ca905eea86db48747b7367a8a

more cholmod makefile fixes

Hg-commit: 4c2ad6802e5a7c5e3d51dd2ee592927cccaa41e2

Add CHOLMOD to standard build

Hg-commit: aec05d2a4ae0545087e791194271bc4898edebef

SeqAIJ implementation for CHOLMOD

Hg-commit: 8ae4884ae75bffc4c810c58d11d51aa9b024e337

fixed MatOrdering_xxx to MatGetOrdering_xxx to match the routine MatGetOrdering()

Hg-commit: 94d7f334ca880f2933e9eff3f44ef72291bfc5c7

make AIJ matrices be able to turn off use of inodes when inodes no longer exist. Previously once an AIJ matrix used Inodes it always used inodes.

Hg-commit: 3dfe5fed95605badef5e0b0e882202c04fe873e0

commit after merge

Hg-commit: d5d6271688b0e7608ca3883cc177d4942a399f4c

undid the MatSeqXAIJReallocation conversion from macro to function because it was broken.

Hg-commit: dea7891b2ca43e5c3492cc1ddf826db987b36c10