Salvatore Filippone
|
188dee6842
|
Add indx_map%inc_lc() method
|
10 months ago |
sfilippone
|
b99aa7a90f
|
Switch off OMP in HASH g2l_ins
|
10 months ago |
sfilippone
|
4e0a9e5db8
|
Merge branch 'repackage' of github.com:sfilippone/psblas3 into repackage
|
10 months ago |
sfilippone
|
e72c0f0bf9
|
Fix OMP impl of sparse-sparse product
|
10 months ago |
Salvatore Filippone
|
d444a12879
|
Condition call to x%sync() in vect_mv
|
10 months ago |
Salvatore Filippone
|
5e2e1e34fd
|
Introduce set_host() in inner_vect_sv
|
10 months ago |
sfilippone
|
025350a361
|
Make sure realloc is always called with size >0
|
11 months ago |
sfilippone
|
ba8c32c507
|
Define merge_nd method
|
11 months ago |
sfilippone
|
aca1848401
|
New timings in CG
|
11 months ago |
sfilippone
|
a2f92e616f
|
Put VOLATILE under ifdef for FLANG
|
11 months ago |
sfilippone
|
0023b8ac78
|
Compile adjcncy_fnd_owner
|
11 months ago |
sfilippone
|
3a25d7b04a
|
Fixes for LLVM compilation
|
12 months ago |
sfilippone
|
373d841bce
|
Don't need renaming of psi_gth and psi_sct
|
12 months ago |
sfilippone
|
472f16f0df
|
Fix compilation with --enable-serial
|
12 months ago |
sfilippone
|
86be8ebcd0
|
New method W%XYZW()
|
1 year ago |
Salvatore Filippone
|
f4c7604f61
|
Fix base implementation of abgdxyz to call set_host
|
1 year ago |
Salvatore Filippone
|
b8f9badf95
|
Fix interface between vect and base_vect%ABGD
|
1 year ago |
Salvatore Filippone
|
2a40b82b58
|
Fix typo in base_vect_mod
|
1 year ago |
Salvatore Filippone
|
4e611bb078
|
Enable psi_abgdxyz
|
1 year ago |
Salvatore Filippone
|
9ced67634d
|
Fix KIND for NR in axpby
|
1 year ago |
Salvatore Filippone
|
3121c43582
|
Silly bug in abgdxyz implementation
|
1 year ago |
Salvatore Filippone
|
5c3d5f0235
|
Silly bug in abgdxyz implementation
|
1 year ago |
Salvatore Filippone
|
29669b56a2
|
Implementation of psb_abgdxyz
|
1 year ago |
Salvatore Filippone
|
a942b47f7c
|
Merge branch 'nond-rep' of github.com:sfilippone/psblas3 into nond-rep
|
1 year ago |
Salvatore Filippone
|
6c53b6ec79
|
Fix typo in interface for psb_abgdxyz
|
1 year ago |
sfilippone
|
83ededd02b
|
Implementatino of abgd_xyz
|
1 year ago |
sfilippone
|
45f00e6e19
|
Fixed comments
|
1 year ago |
Salvatore Filippone
|
14c4ff0f32
|
Added new methd for two combined axpbys
|
1 year ago |
sfilippone
|
3aa3c795e9
|
Refactor assembly and cnv
|
1 year ago |
sfilippone
|
49e99a3e82
|
Fix conversion and product to enable overlap with GPU
|
1 year ago |
sfilippone
|
74cf138a6c
|
Merge branch 'repackage' into non-diag
|
1 year ago |
sfilippone
|
be7571f568
|
Fix missing directive
|
1 year ago |
sfilippone
|
e9d1238b43
|
Add detailed measurements.
|
1 year ago |
sfilippone
|
a6016f00fa
|
Bump PSBLAS version to 3.9
|
1 year ago |
sfilippone
|
6c9ca58282
|
Silly bug in coo insert
|
1 year ago |
sfilippone
|
d3b2b7816d
|
Fix coo insert OpenMP. Fix Make.inc.in
|
1 year ago |
sfilippone
|
ae7fad95d4
|
Merge branch 'development' into non-diag
|
1 year ago |
sfilippone
|
a6ec655a97
|
Prepare merge
|
1 year ago |
sfilippone
|
a2788bdf0b
|
New version with ND product
|
1 year ago |
sfilippone
|
d718ef1e6d
|
Always allocate szs in psb_gather
|
1 year ago |
sfilippone
|
baf18cebd7
|
Further fix for gather.
|
1 year ago |
sfilippone
|
5caee551e5
|
Fixed IN_PLACE option for collectives.
|
1 year ago |
sfilippone
|
d82b090289
|
Fix makefile for psi_acx & friends
|
1 year ago |
sfilippone
|
e31dd52c41
|
Fixed CRITICAL in hash_mod
|
2 years ago |
sfilippone
|
def0635c53
|
More OMP directives in cd_inloc
|
2 years ago |
sfilippone
|
41be1357c3
|
Set defaults for SPSPMM depending on OpenMP compilation.
|
2 years ago |
Salvatore Filippone
|
0d8a5d3dc2
|
New SPSPMM implementation
|
2 years ago |
Salvatore Filippone
|
d0cacda995
|
Moved various modules related to RB around, into auxil, update Makefile.
|
2 years ago |
Salvatore Filippone
|
7b45994b70
|
Setter/getter for SPSPMM algorithm in base_mat_mod
|
2 years ago |
wlthr
|
2322a9ce61
|
using end_idx to copy data from threads in gustavson and gustavson_1d
|
2 years ago |
wlthr
|
0185b79b2a
|
added setter for d_csr_spspmm implementation
|
2 years ago |
wlthr
|
0fe95c3c76
|
added use statement
|
2 years ago |
wlthr
|
979a3da95f
|
merged dev-openmp into omp-walther
|
2 years ago |
wlthr
|
1af76c067c
|
added parallel double precision spspmm implementations
|
2 years ago |
sfilippone
|
f001ebbad3
|
Final fix for COO on OMP
|
2 years ago |
sfilippone
|
26bf4c5d69
|
Fixed COO csput for OMP/not OMP
|
2 years ago |
sfilippone
|
3aa748b0e3
|
Finish dual OMP/notOMP g2lv1_ins
|
2 years ago |
sfilippone
|
08c1ab0cd1
|
Fix tril/triu in COO for non-OMP paths.
|
2 years ago |
sfilippone
|
ca82520b88
|
Reworked CSR TRIL/TRIU for OpenMP
|
2 years ago |
sfilippone
|
5e691d5bff
|
Some improvements for openmp vector updates
|
2 years ago |
sfilippone
|
bb9f213551
|
Define and implement OMP version of TRIL/TRIU
|
2 years ago |
sfilippone
|
2f403e0df7
|
Rework cp_{from|to}_fmt for better OpenMP performance
|
2 years ago |
sfilippone
|
d378266f33
|
Fix synatx error
|
2 years ago |
sfilippone
|
a66778f270
|
Improve coo and merge development
|
2 years ago |
sfilippone
|
347352fe1e
|
Make spins work in OpenMP from either par or serial
|
2 years ago |
sfilippone
|
db0e4db507
|
Minimize debug sttements in hash_ins
|
2 years ago |
sfilippone
|
1941affe7a
|
Exposed error in AMG test when not parallelizing generation loop
|
2 years ago |
sfilippone
|
494e29dd2e
|
Cosmetic adjustments to COO and BSRCH
|
2 years ago |
sfilippone
|
739dc78a75
|
Merge branch 'development' into omp-threadsafe
|
2 years ago |
sfilippone
|
7e5dc20e03
|
Define new options for BSRCH, clean interface
|
2 years ago |
sfilippone
|
40cc78854a
|
Improve implementation of fix_coo using exscan
|
2 years ago |
sfilippone
|
91d3e66547
|
Merge branch 'omp-threadsafe' of github.com:sfilippone/psblas3 into omp-threadsafe
|
2 years ago |
sfilippone
|
74a8217520
|
Fixed silly bug in EXSCAN and usage in CSR_IMPL
|
2 years ago |
Salvatore Filippone
|
5bc02fb2e6
|
Take out redundant statements in SPINS
|
2 years ago |
sfilippone
|
f3efea0a89
|
Take out IBASE from exscan, makes no sense.
|
2 years ago |
sfilippone
|
05b684ddbb
|
Updated use of exscan in CSC
|
2 years ago |
sfilippone
|
9c248a31e2
|
Refactored EXSCAN and its OpenMP usage.
|
2 years ago |
sfilippone
|
02dd204351
|
Implement psi_exscan and use in _from_coo
|
2 years ago |
sfilippone
|
dbd55321f8
|
Fixed CSR mv and cp _from_coo with OpenMP.
|
2 years ago |
sfilippone
|
6ba7d93159
|
Fix CRITICAL in LIST%G2L_INS
|
2 years ago |
sfilippone
|
5a5712b4f0
|
Rely on CRITICAL inside G2L_INS implementation
|
2 years ago |
sfilippone
|
f068d73ef1
|
First working version
|
2 years ago |
sfilippone
|
8459ea28f5
|
Modified matrix build procedures with OpenMP
|
2 years ago |
sfilippone
|
eb11e5e053
|
Put CRITICAL(name) in G2L_INS
|
2 years ago |
sfilippone
|
0f1603a2e9
|
The current version of test/omp seems to be working. To be completed
|
2 years ago |
sfilippone
|
98945f36b5
|
Fix nrm2 with overlap
|
2 years ago |
sfilippone
|
c05b32c202
|
Reset status for csr_impl.
|
2 years ago |
Salvatore Filippone
|
ed7862a848
|
Fix OpenMP g2lv1_ins
|
2 years ago |
Salvatore Filippone
|
bb4e80f647
|
Bit of cleanup in psb_hash_map_mod
|
2 years ago |
Salvatore Filippone
|
49d37911ca
|
Work on psb_hash_map_mod
|
2 years ago |
Salvatore Filippone
|
0480610822
|
Merge branch 'dev-openmp' of github.com:sfilippone/psblas3 into dev-openmp
|
2 years ago |
Salvatore Filippone
|
784cc65e51
|
Temporarily revert hash_map_mod waiting for a proper fix
|
2 years ago |
Salvatore Filippone
|
fd0b1482e5
|
Merge branch 'dev-openmp' of github.com:sfilippone/psblas3 into dev-openmp
|
2 years ago |
Salvatore Filippone
|
afdbac6727
|
Swicth csr_impl to F90
|
2 years ago |
Salvatore Filippone
|
86b8a261ef
|
Fixed conversion bug, changed SPASB interface
|
2 years ago |
Salvatore Filippone
|
f09e25524e
|
Create ECSR format and use it for A%AND
|
2 years ago |
Salvatore Filippone
|
00cc83cde8
|
First version of AD/AND with memory duplication
|
2 years ago |
Salvatore Filippone
|
de37e3602a
|
Fix SV with CONJG
|
2 years ago |
Salvatore Filippone
|
d4b6d4dfa1
|
Fix reinit
|
2 years ago |
Salvatore Filippone
|
7028cb656a
|
Fix trim never to reallocate to sizes <=0
|
2 years ago |