Matthijs Douze
df2edbe3ab
fix issue 320
2018-01-31 07:00:24 -07:00
matthijs
9933892ec9
sync with FB version 2017-01-09
...
- adding HNSW indexing method
- simultaneous search and reconstruction for IndexIVFPQ
2018-01-09 06:42:06 -08:00
matthijs
fe446e41a2
bugfixes
2017-11-23 06:34:53 -08:00
matthijs
250a3d3f18
sync with FB version 2017-11-22
...
various bugfixes from github issues
kmean with some frozen centroids
GPU better tiling for large flat datasets
default AVX for vector ops
2017-11-22 05:11:28 -08:00
matthijs
71335194f6
sync with FB version 2017-11-22
...
various bugfixes from github issues
kmean with some frozen centroids
GPU better tiling for large flat datasets
default AVX for vector ops
2017-11-22 05:10:35 -08:00
matthijs
d8eeceb883
sync with FB version 2017-09-08
...
fix index_cpu_to_gpu_multiple
fix PCA dimension
2017-09-08 04:39:41 -07:00
matthijs
54a2abb30d
sync with FB version 2017-08-09 (2)
2017-08-09 11:21:50 -07:00
matthijs
a5ef16db89
sync with FB version 2017-08-09
2017-08-09 11:13:51 -07:00
matthijs
8e3dc6f2b0
changed license
2017-07-30 00:18:45 -07:00
matthijs
2882e89397
forgotten
2017-07-26 03:12:49 -07:00
matthijs
ca2ea5097b
Sync with FB version 2017-07-16
...
- access GPU resources
- all Hamming distances of all sizes supported
2017-07-26 03:11:01 -07:00
matthijs
13754a74a3
fix cmakelists
2017-06-21 09:01:45 -07:00
matthijs
2d59917d33
fix GPU compile
2017-06-21 09:01:44 -07:00
matthijs
784e2facd8
Synchronization with FB version 2017-06-21
...
* moved most FAISS_ASSERT calls to C++ exceptions, and adjusted
memory allocation to avoid mem leaks
* added an IndexIVFScalarQuantizer type that offers an
intermediate compression between IVFFlat and IVFPQ
* support removal of indices in IndexIDMap / IndexFlat combination
* various fixes in GPU code
2017-06-21 09:01:06 -07:00
Boris Mansencal
678750113e
add support for Nvidia GPUs with CUDA compute capability 6.1 ( #110 )
2017-05-31 15:52:39 +02:00
BorisMansencal
5b92eeb285
correct utils::nextHighestPowerOf2
...
The bit shift was done as int and not as current type (size_t for example), thus the computed next highest power of 2 was wrong on large numbers.
2017-05-06 20:36:16 +02:00
Tianwei Shen
80314d9f07
add initial cmake support ( #75 )
...
* add initial cmake support
* update cmake, add cmake instructions to INSTALL
* update findopenmp and INSTALL
* change FindOpenBLAS.cmake to cater for macports
- change cblas.h to openblas_config.h since macports does not ship
cblas.h with openblas.
* revise INSTALL for cmake
2017-05-02 11:04:50 +02:00
matthijs
7abe81b4f6
Better support for low-mem GPUs
...
avoid reading beyond the end of an array in fvec_L2sqr and related functions
2017-04-06 04:33:41 -07:00
matthijs
c507707098
sync with FB version. Added:
...
- better selection of training sets for PQ and preprocessing
- GPU parameter object
- IndexIDMap fixed
- fixed redo bug in clustering
2017-03-20 10:48:35 -07:00
matthijs
1114358311
make GPU headers C++03 compatible
2017-03-02 22:31:32 -08:00
matthijs
e636a6e499
fix gpu makefile
2017-02-24 09:01:49 -08:00
matthijs
00c288c0d9
added benchmarking scripts
2017-02-24 08:37:57 -08:00
matthijs
a67190c69c
added benchmarking scripts
2017-02-24 08:32:08 -08:00
Matthijs Douze
c670118a4e
Initial commit
2017-02-22 23:26:44 +01:00
