Christoph Ruegg
|
1fcd334aaf
|
Release MKL Native Provider v2.2.0 (rev 11)
|
10 years ago |
|
Christoph Ruegg
|
6005ecaac7
|
Compatibility: do not use CharSet.Auto on p/invoke
|
10 years ago |
|
Christoph Ruegg
|
d5d6ef1765
|
Merge pull request #445 from Ksero/patch-1
Fix multivariate Runge-Kutta ODE solver typo
|
10 years ago |
|
Christoph Ruegg
|
9dc7fff20f
|
MKL: indicate in ToString if the provider is ahead or behind the revision it was designed against
|
10 years ago |
Ksero
|
6db52d799a
|
Fix multivariate Runge-Kutta ODE solver typo
The multiplication by the time-step was missing in this method (but not the others)
|
10 years ago |
|
Christoph Ruegg
|
01d2a9ce62
|
FFT: tests to verify real against complex, and real forward to inverse
|
10 years ago |
|
Christoph Ruegg
|
5bdc17afa3
|
FFT-Managed: fix real coonjugate-even impl in odd-N cases
|
10 years ago |
|
Christoph Ruegg
|
e85c2bf137
|
FFT-Managed: throw NotSupportedException on multi-dimensional FFTs
|
10 years ago |
|
Christoph Ruegg
|
e1dae7578b
|
FFT-Managed: temporary real conjugate-even implementation
|
10 years ago |
|
Christoph Ruegg
|
de9e8d101d
|
FFT: real-complex, packed
|
10 years ago |
|
Christoph Ruegg
|
f1f3b93bb9
|
FFT-MKL: real packed transformation support
|
10 years ago |
|
Christoph Ruegg
|
6103b583bf
|
FFT-MKL: use complex complex mode for conjugate even storage, simplify
|
10 years ago |
|
Christoph Ruegg
|
871da47b5b
|
FFT: variation that operates on real/imag arrays instead of a complex number arrray
|
10 years ago |
|
Christoph Ruegg
|
6a808f12c9
|
FFT: Simplify API
|
10 years ago |
|
Christoph Ruegg
|
c168f7d71c
|
FFT: inline docs
|
10 years ago |
|
Christoph Ruegg
|
8a8abb0282
|
FFT: expose 2D, 2D-matrix and multidim FFT
|
10 years ago |
|
Christoph Ruegg
|
d7bc61240c
|
FFT-MKL: generalize 2D case to arbitrary multiple dimensions
|
10 years ago |
|
Christoph Ruegg
|
31106ef3b0
|
FFT-MKL: support for 2D FFT
|
10 years ago |
|
Christoph Ruegg
|
52a857f2c0
|
FFT-MKL: switch back to MKL_LONG instead of MKL_INT64
|
10 years ago |
|
Christoph Ruegg
|
aae1d8b0a2
|
FFT-MKL: cleanup - drop redundant suffixes
|
10 years ago |
|
Christoph Ruegg
|
714d9af14a
|
Special Functions: more robust GammaLowerRegularizedInv and Gamma.InvCDF #442
|
10 years ago |
|
Christoph Ruegg
|
280ee8b0b2
|
LA-MKL: query capability levels once, at initialization
|
10 years ago |
|
Christoph Ruegg
|
3bef6497ca
|
LA-MKL: bind dot-power to MKL vector functions
|
10 years ago |
|
Christoph Ruegg
|
de174200ea
|
Window: periodic versions of Hamming, Hann, Cosine and Lanczos windows
|
10 years ago |
|
Christoph Ruegg
|
6070beba7a
|
LA: Matrix Moore-Penrose Pseudo-Inverse (SVD) #432
|
10 years ago |
|
Christoph Ruegg
|
d1dfb51a3d
|
LA: vector dot-power with vector exponent, bound to provider
|
10 years ago |
|
Christoph Ruegg
|
f8012bca55
|
LA: bind matrix dot-power with matrix exponent to provider
|
10 years ago |
|
Christoph Ruegg
|
42c48a0ce9
|
LA: managed provider experimental dense multiplication variation
|
10 years ago |
|
Christoph Ruegg
|
5a5bc8a6f5
|
LA: Add matrix dot-power with matrix exponent
|
10 years ago |
|
Christoph Ruegg
|
e7cb8d87f3
|
Bench: switch from a * b.Transpose to a.TransposeAndMultiply(b)
|
10 years ago |
|
Christoph Ruegg
|
f242e75e68
|
Bench: Dense Matrix Product, including tall and wide (with transpose)
|
10 years ago |
|
Christoph Ruegg
|
7343f3d848
|
Bench: reinstate matrix product benchmark
|
10 years ago |
|
Christoph Ruegg
|
156faa2b97
|
Tools: drop old NBenchmark
|
10 years ago |
|
Christoph Ruegg
|
36f47eab88
|
FFT: tweak threading of managed radix2 implementation (much faster ion
small N) #440
|
10 years ago |
|
Christoph Ruegg
|
ec91f3bc20
|
FFT-MKL: descriptor reuse in multi-threading scenario
|
10 years ago |
|
Christoph Ruegg
|
12f1c7bb3e
|
FFT-MKL: reuse matching descriptor (cache=1)
|
10 years ago |
|
Christoph Ruegg
|
608fcf4d91
|
FFT-MKL: wip towards reuseable descriptors
|
10 years ago |
|
Christoph Ruegg
|
2e0fca5b72
|
Bench: fft focus on radix2 and smaller N
|
10 years ago |
|
Christoph Ruegg
|
f8557d3f83
|
Bench: add memory and inlining diagnostics
|
10 years ago |
|
Christoph Ruegg
|
0be3925a8c
|
Bench: revive benchmarking project, for FFT
|
10 years ago |
|
Christoph Ruegg
|
7c2d3b1892
|
Providers: test folder structure
|
10 years ago |
|
Christoph Ruegg
|
c836df7219
|
Find Roots: extend bracketing in derivative-free algorithms to also reduce/subdevision after trying to expand
|
10 years ago |
|
Christoph Ruegg
|
b076067402
|
Root Finding: Zero Crossing Bracketing expansion must reset boundaries if it fails
|
10 years ago |
|
Christoph Ruegg
|
9ee9627a09
|
Merge branch 'fft'
|
10 years ago |
|
Christoph Ruegg
|
efcb5d5b5b
|
FFT: verify mapping between all fourier options
|
10 years ago |
|
Christoph Ruegg
|
4bad397c44
|
FFT: activate in Fourier.Forward/Backward, with full option mapping
|
10 years ago |
|
Christoph Ruegg
|
432bf493a5
|
FFT: force 64bit integers with MKL_INT64
|
10 years ago |
|
Christoph Ruegg
|
c472e8d4c8
|
FFT: fix PCL builds
|
10 years ago |
|
Christoph Ruegg
|
25282489f3
|
FFT: cleanup Cuda and OpenBlas providers similar to MKL refactoring
|
10 years ago |
|
Christoph Ruegg
|
037c1043cb
|
FFT: rework provider discovery to throw less exceptions and work better
when FFT and LA use different providers
|
10 years ago |
