|
Short Description:
Dew Lab Studio is an object oriented, multithreaded, vectorized, 64bit, Open CL enabled numerical library for Delphi/C++ 2007/9/10/XE1-8/10/10.1/10.2/10.3. Supports complex numbers, math parser, optimization, signal processing, statistics...
Long Description 1:
Dew Lab Studio is an object oriented, multithreaded, vectorized, 64bit, Open CL enabled numerical matrix library for Delphi/C++ 2007/9/10/XE1,2,3,4,5,6,7,8/Seattle 10/Berlin 10.1/Tokyo 10.2. Features: complete support for complex numbers, math parser , polynomial handling , probabilities, unconstrained and bounded minimization, SVD, QR, LQ, and LU solvers, sparse matrices, sparse matrices PLUS signal processing and statistical analysis.
Long Description 2:
Dew Lab Studio includes MtxVec math library and additional signal analysis (DSP Master) and statistical analysis (Stats Master) add-on packages.
Features include:
- fast object oriented numerical/matrix library
- Intel AVX 1/2, AVX-512, SSE4, Open CL and 64bit support enabled.
- optional full source build path for cross-platform support.
- super conductive memory manager allows linear scaling with number of CPU cores.
- LAPACK 3.4 based
- Complete support for complex numbers for all functions.
- Rich set of matrix and vector functions and operators.
- Solver for banded sparse matrices, Iterative sparse solvers, Direct sparse solver for non-symmetric matrices. Conversions support Banded, Dense, Sparse and Triplets matrix formats
- 23 probability distributions (PDF, CDF and inverse CDF function), special functions, optimization routines, random generators and distribution parameter estimation.
- fast vectorized math expression parser and evaluator
- 1D and 2D FFT's, correlations and autocorrelations.
- arbitrary functions of matrix including square root of matrix and logarithm of matrix
- optimized integer math library.
- generalized Eig and SVD values
- sparse matrix eigenvalues for symmetric and generalized symmetric matrices
Statistics
- Descriptive statistics
- Hypothesis testing
- Weighted/unweigted linear and multiple linear regression.
- Weighted/unweighted non-linear regression
- Time series analysis (ARMA, ARIMA, ARAR, MA, AR, ...)
- PCA, multivariante analysis, ...
Signal processing
- IIR filters (Cheybshev I-II, Butterworth and Elliptic).
- FIR filters (Remez and Kaiser) with complete filter design support.
- Signal forecasting
- Standard and extended Kalman filters.
- Spectral analysis, cross/bi spectral analysis
- Conversions between forms of transfer function
- Audio rec/playback with ASIO/WASAPI.
- interpolated FIR filters, DCT/IDCT
- real factor high accuracy fast rate conversion
|