MtxVec is an object oriented vectorized math library, the core of Dew Lab Studio, featuring a comprehensive set of mathematical and statistical functions executing at impressive speeds (check the testimonies of some of our many customers ).
Having converted our models from Matlab, a speed increase of over 10x has been achieved . This core acceleration finally allows these models to be used in real-time automation control.
Optimized Functions
The base math library uses the LAPACK (Linear Algebra Pack) version optimized for Core Duo and Core i7 CPU’s provided by Intel with their Math Kernel library. Our library is organized into a set of “primitive” highly optimized functions covering all the basic math operations. All higher level algorithms use these basic optimized functions, similar to the way LAPACK uses the Basic Linear Algebra Subprograms (BLAS).
Performance Secrets
Code vectorization
The secret behind our performance is called "code vectorization". We program achieves substantial performance improvements in floating point arithmetic by exploiting the CPU Streaming SIMD Extensions instruction sets. (SIMD = Single Instruction Multiple Data.)
Super conductive memory management
Effective massively parallel execution is achieved with the help of a super conductive memory management, which features zero thread contention and inter-lock problems allowing linear scaling with number of cores while maintaining low memory consumption and no interference with non-computational parts of the project.
Other Features
Linear Algebra Package
MtxVec makes extensive use of the Linear Algebra Package, which today is the de-facto standard for linear algebra and is free from www.netlib.org. Because this package is standard, different CPU makers provide performance-optimized versions to achieve maximum performance. Linear algebra routines are the bottleneck of many frequently used algorithms; therefore, this package is the part of the code that makes the most sense to optimize. MtxVec uses the version optimized for individual CPUs provided by Intel with their Math Kernel library.
Intel Performance Primitives
Our library also makes extensive use of Intel Performance Primitives, which accelerate mainly the non-linear algebra based functions. The Intel Performance Primitives run on all Intel x86-compatible CPUs, old and new, but will achieve highest performance on the Intel Core architecture.
What You Gain With MtxVec
Low-level math functions are wrapped in simple-to-use code primitives.
Write Vector/Matrix expressions with the standard set of *,/,+- operators
Best in class performance when evaluating Vector/Matrix expressions.
All primitives have internal and automatic memory management. This frees you from a wide range of possible errors like: allocating insufficient memory, forgetting to free the memory, keeping too much memory allocated at the same time, and similar errors.
Parameters are explicitly range checked before they are passed to the dll routines. This ensures that all internal dll calls are safe to use.
When calling linear algebra routines, the software automatically compensates for the fact that in FORTRAN the matrices are stored by columns but in other languages are stored by rows.
Many linear algebra functions take several parameters. The software automatically fills in most of the parameters, thus reducing the time to study each function extensively before you can use it.
The software is organized into a set of "primitive" highly optimized functions covering all the basic math operations. All higher-level algorithms use these basic operations, in a way similar to how LAPACK uses the Basic Linear Algebra Subprograms (BLAS).
Although some compilers support native SSE4/AVX2/AVX512 instruction sets, the resulting code can never be as optimal as a hand-optimized version.
Many linear algebra routines are multithreaded, including FFT's and sparse matrix solvers.
All functions must pass very strict automated tests. These tests give the library the highest possible level of reliability, accuracy and error protection.
When you write numerical algorithms, you will find the compactness and readability of code improves noticeably and you achieve significantly shorter development times.
High performance expression parser/scripter that can work with vector and matrix variables.
Debugger Visualizer for faster inspection of variable contents during debugging.
