31106ef3b0
FFT-MKL: support for 2D FFT
10 years ago
52a857f2c0
FFT-MKL: switch back to MKL_LONG instead of MKL_INT64
10 years ago
aae1d8b0a2
FFT-MKL: cleanup - drop redundant suffixes
10 years ago
714d9af14a
Special Functions: more robust GammaLowerRegularizedInv and Gamma.InvCDF #442
10 years ago
280ee8b0b2
LA-MKL: query capability levels once, at initialization
10 years ago
3bef6497ca
LA-MKL: bind dot-power to MKL vector functions
10 years ago
de174200ea
Window: periodic versions of Hamming, Hann, Cosine and Lanczos windows
10 years ago
6070beba7a
LA: Matrix Moore-Penrose Pseudo-Inverse (SVD) #432
10 years ago
d1dfb51a3d
LA: vector dot-power with vector exponent, bound to provider
10 years ago
f8012bca55
LA: bind matrix dot-power with matrix exponent to provider
10 years ago
42c48a0ce9
LA: managed provider experimental dense multiplication variation
10 years ago
5a5bc8a6f5
LA: Add matrix dot-power with matrix exponent
10 years ago
e7cb8d87f3
Bench: switch from a * b.Transpose to a.TransposeAndMultiply(b)
10 years ago
f242e75e68
Bench: Dense Matrix Product, including tall and wide (with transpose)
10 years ago
7343f3d848
Bench: reinstate matrix product benchmark
10 years ago
c1d2b4bfeb
Merge pull request #439 from scottstephens/optimization_finite_difference
Optimization: add finite difference objective function adapter
10 years ago
156faa2b97
Tools: drop old NBenchmark
10 years ago
36f47eab88
FFT: tweak threading of managed radix2 implementation (much faster ion
small N) #440
10 years ago
ec91f3bc20
FFT-MKL: descriptor reuse in multi-threading scenario
10 years ago
12f1c7bb3e
FFT-MKL: reuse matching descriptor (cache=1)
10 years ago
608fcf4d91
FFT-MKL: wip towards reuseable descriptors
10 years ago
2e0fca5b72
Bench: fft focus on radix2 and smaller N
10 years ago
35137adef5
Optimization: add finite difference objective function adapter
10 years ago
f8557d3f83
Bench: add memory and inlining diagnostics
10 years ago
0be3925a8c
Bench: revive benchmarking project, for FFT
10 years ago
7c2d3b1892
Providers: test folder structure
10 years ago
c836df7219
Find Roots: extend bracketing in derivative-free algorithms to also reduce/subdevision after trying to expand
10 years ago
b076067402
Root Finding: Zero Crossing Bracketing expansion must reset boundaries if it fails
10 years ago
9ee9627a09
Merge branch 'fft'
10 years ago
efcb5d5b5b
FFT: verify mapping between all fourier options
10 years ago
4bad397c44
FFT: activate in Fourier.Forward/Backward, with full option mapping
10 years ago
432bf493a5
FFT: force 64bit integers with MKL_INT64
10 years ago
c472e8d4c8
FFT: fix PCL builds
10 years ago
25282489f3
FFT: cleanup Cuda and OpenBlas providers similar to MKL refactoring
10 years ago
037c1043cb
FFT: rework provider discovery to throw less exceptions and work better
when FFT and LA use different providers
10 years ago
ea3b57c4b1
FFT: include MKL version in MKL provider description (ToString)
10 years ago
d7440d5125
FFT: MKL provider refactoring due to sharing between FFT and LA - 2
10 years ago
a4f5b618e4
FFT: MKL provider refactoring due to sharing between FFT and LA
10 years ago
6a42fb7c83
FFT: fix PCL builds, cleanup
10 years ago
df3264c173
FFT: do not mask MKL error status
10 years ago
3b92b38afd
FFT: forward and backward scaling support, drop manual correction in parseval test
10 years ago
31c140cf80
FFT: MKL backward transformation
10 years ago
2ab73ac781
FFT: verify parseval theorem
10 years ago
8639df292f
FFT: experimental implementation with test for complex-complex inplace forward FFT
10 years ago
0486971c4a
FFT: wire provider to Control class
10 years ago
7f5b414bbe
FFT: fft provider mvp interface
10 years ago
50203b2c1e
Merge pull request #435 from scottstephens/unified_optimization_better_tests
Better tests for optimization
10 years ago
de2ce192bb
Optimization: Add MGH tests to Nelder-Mead algo
10 years ago
05a0805194
Optimization: Ignore tests that fail
* I suspect they fail because the algorithm doesn't work well
for those test cases, not that there is an implementation error.
But this suspicion has not been verified.
10 years ago
f621a940d1
Optimization: revamp testing
* Changed file and class names from TestClassName to
ClassNameTests to better conform with other parts of Math.Net
* Added some organizational abstractions to simplify adding
test cases.
* Added test cases based on several functions from
Testing Unconstrained Optimization Software
Jorge J. Moré, Burton S. Garbow, Kenneth E. Hillstrom
ACM Transactions on Mathematical Software, Vol 7, No. 1, March 1981, Pages 17-41.
* At this point, most test functions are low-dimensional,
well-scaled, unimodal functions.
* There are test failures for Conjugate Gradient and Newton
unconstrained minimizers. Currently I believe these to be
failures of the algorithm, rather than failures of
implementation, but it hasn't been investigated thoroughly.
10 years ago
Christoph Ruegg
Scott Stephens