Benoit Steiner
0b9e3dcd06
Added packet primitives to compute exp, log, sqrt and rsqrt on fp16. This improves the performance by 10 to 30%.
2016-05-10 11:05:33 -07:00
Benoit Steiner
6bf8273bc0
Added a test to validate the new non blocking thread pool
2016-05-10 10:49:34 -07:00
Benoit Steiner
4013b8feca
Simplified the reduction code a little.
2016-05-10 09:40:42 -07:00
Benoit Steiner
75bd2bd32d
Fixed compilation warning
2016-05-09 19:24:41 -07:00
Benoit Steiner
4670d7d5ce
Improved the performance of full reductions on GPU:
...
Before:
BM_fullReduction/10 200000 11751 8.51 MFlops/s
BM_fullReduction/80 5000 523385 12.23 MFlops/s
BM_fullReduction/640 50 36179326 11.32 MFlops/s
BM_fullReduction/4K 1 2173517195 11.50 MFlops/s
After:
BM_fullReduction/10 500000 5987 16.70 MFlops/s
BM_fullReduction/80 200000 10636 601.73 MFlops/s
BM_fullReduction/640 50000 58428 7010.31 MFlops/s
BM_fullReduction/4K 1000 2006106 12461.95 MFlops/s
2016-05-09 17:09:54 -07:00
Benoit Steiner
c3859a2b58
Added the ability to use a scratch buffer in cuda kernels
2016-05-09 17:05:53 -07:00
Benoit Steiner
ba95e43ea2
Added a new parallelFor api to the thread pool device.
2016-05-09 10:45:12 -07:00
Benoit Steiner
dc7dbc2df7
Optimized the non blocking thread pool:
...
* Use a pseudo-random permutation of queue indices during random stealing. This ensures that all the queues are considered.
* Directly pop from a non-empty queue when we are waiting for work,
instead of first noticing that there is a non-empty queue and
then doing another round of random stealing to re-discover the non-empty
queue.
* Steal only 1 task from a remote queue instead of half of tasks.
2016-05-09 10:17:17 -07:00
Benoit Steiner
05c365fb16
Pulled latest updates from trunk
2016-05-07 13:39:04 -07:00
Benoit Steiner
691614bd2c
Worked around a bug in nvcc on tegra x1
2016-05-07 13:28:53 -07:00
Benoit Steiner
a2d94fc216
Merged latest updates from trunk
2016-05-06 19:17:57 -07:00
Benoit Steiner
8adf5cc70f
Added support for packet processing of fp16 on kepler and maxwell gpus
2016-05-06 19:16:43 -07:00
Benoit Steiner
1660e749b4
Avoid double promotion
2016-05-06 08:15:12 -07:00
Benoit Steiner
c54ae65c83
Marked a few tensor operations as read only
2016-05-05 17:18:47 -07:00
Benoit Steiner
69a8a4e1f3
Added a test to validate full reduction on tensor of half floats
2016-05-05 16:52:50 -07:00
Benoit Steiner
678a17ba79
Made the testing of contractions on fp16 more robust
2016-05-05 16:36:39 -07:00
Benoit Steiner
e3d053e14e
Refined the testing of log and exp on fp16
2016-05-05 16:24:15 -07:00
Benoit Steiner
9a48688d37
Further improved the testing of fp16
2016-05-05 15:58:05 -07:00
Benoit Steiner
0451940fa4
Relaxed the dummy precision for fp16
2016-05-05 15:40:01 -07:00
Benoit Steiner
910e013506
Relaxed an assertion that was tighter that necessary.
2016-05-05 15:38:16 -07:00
Benoit Steiner
f81e413180
Added a benchmark to measure the performance of full reductions of 16 bit floats
2016-05-05 14:15:11 -07:00
Benoit Steiner
28d5572658
Fixed some incorrect assertions
2016-05-05 10:02:26 -07:00
Benoit Steiner
2aba40d208
Avoid unecessary type promotion
2016-05-05 09:26:57 -07:00
Benoit Steiner
a4d6e8fef0
Strongly hint but don't force the compiler to unroll a some loops in the tensor executor. This results in up to 27% faster code.
2016-05-05 09:25:55 -07:00
Benoit Steiner
7875437ca0
Avoided unecessary type promotion
2016-05-05 09:08:42 -07:00
Benoit Steiner
f363e533aa
Added tests for full contractions using thread pools and gpu devices.
...
Fixed a couple of issues in the corresponding code.
2016-05-05 09:05:45 -07:00
Benoit Steiner
06d774bf58
Updated the contraction code to ensure that full contraction return a tensor of rank 0
2016-05-05 08:37:47 -07:00
Christoph Hertzberg
b300a84989
Fixed some singed/unsigned comparison warnings
2016-05-05 13:36:28 +02:00
Christoph Hertzberg
dacb469bc9
Enable and fix -Wdouble-conversion warnings
2016-05-05 13:35:45 +02:00
Benoit Steiner
62b710072e
Reduced the memory footprint of the cxx11_tensor_image_patch test
2016-05-04 21:08:22 -07:00
Benoit Steiner
dd2b45feed
Removed extraneous 'explicit' keywords
2016-05-04 16:57:52 -07:00
Ola Røer Thorsen
be78aea6b3
fix double-promotion/float-conversion in Core/SpecialFunctions.h
2016-05-04 10:52:08 +02:00
Gael Guennebaud
75a94b9662
Improve documentation of BDCSVD
2016-05-04 12:53:14 +02:00
Benoit Steiner
968ec1c2ae
Use numext::isfinite instead of std::isfinite
2016-05-03 19:56:40 -07:00
Gael Guennebaud
e2ca478485
bug #1214 : consider denormals as zero in D&C SVD. This also workaround infinite binary search when compiling with ICC's unsafe optimizations.
2016-05-03 23:15:29 +02:00
Benoit Steiner
2c5568a757
Added a test to validate the computation of exp and log on 16bit floats
2016-05-03 12:06:07 -07:00
Benoit Steiner
6c3e5b85bc
Fixed compilation error with cuda >= 7.5
2016-05-03 09:38:42 -07:00
Benoit Steiner
aad9a04da4
Deleted superfluous explicit keyword.
2016-05-03 09:37:19 -07:00
Benoit Steiner
da50419df8
Made a cast explicit
2016-05-02 19:50:22 -07:00
Benoit Steiner
73ef5371e4
Pulled latest updates from trunk
2016-05-01 14:48:57 -07:00
Benoit Steiner
8a9228ed9b
Fixed compilation error
2016-05-01 14:48:01 -07:00
Gael Guennebaud
b1bd53aa6b
Fix performance regression: with AVX, unaligned stores were emitted instead of aligned ones for fixed size assignement.
2016-05-01 23:25:06 +02:00
Benoit Steiner
d6c9596fd8
Added missing accessors to fixed sized tensors
2016-04-29 18:51:33 -07:00
Benoit Steiner
17fe7f354e
Deleted trailing commas
2016-04-29 18:39:01 -07:00
Benoit Steiner
e5f71aa6b2
Deleted useless trailing commas
2016-04-29 18:36:10 -07:00
Benoit Steiner
44f592dceb
Deleted unnecessary trailing commas.
2016-04-29 18:33:46 -07:00
Benoit Steiner
2b890ae618
Fixed compilation errors generated by clang
2016-04-29 18:30:40 -07:00
Benoit Steiner
d217217842
Added a few tests to ensure that the dimensions of rank 0 tensors are correctly computed
2016-04-29 18:15:34 -07:00
Benoit Steiner
f100d1494c
Return the proper size (ie 1) for tensors of rank 0
2016-04-29 18:14:33 -07:00
Benoit Steiner
d14105f158
Made several tensor tests compatible with cxx03
2016-04-29 17:22:37 -07:00
