The Experts below are selected from a list of 36 Experts worldwide ranked by ideXlab platform
Hisao Kanda - One of the best experts on this subject based on the ideXlab platform.
-
an annealing study of nickel point defects in high pressure synthetic diamond
Journal of Applied Physics, 1993Co-Authors: Simon C Lawson, Hisao KandaAbstract:Results of an annealing study, carried out in the temperature range 1500–1900 °C, of nickel‐related optical centers in high‐pressure synthetic diamond are presented. It is established that the well‐known 1.883 and 2.51 eV systems anneal out during the temperature regime of nitrogen aggregation and the concurrent growth of an array of structure, which extends throughout the visible region of the absorption spectrum, and gives the previously bright‐yellow‐colored diamonds a rich golden‐yellow color, is observed. By carrying out the annealing sequence on diamonds grown using various solvent catalysts a correlation is found between preanneal 1.883 eV absorption and maximized absorption of some of the annealed structure. From the results it is proposed that centers associated with nickel and nitrogen are produced and their possible natures are speculated on. It is found that the defect‐induced one‐phonon spectra of the diamonds examined may be satisfactorily decomposed into three components. To account for cer...
-
an annealing study of nickel point defects in high pressure synthetic diamond
Journal of Applied Physics, 1993Co-Authors: Simon C Lawson, Hisao KandaAbstract:Results of an annealing study, carried out in the temperature range 1500–1900 °C, of nickel‐related optical centers in high‐pressure synthetic diamond are presented. It is established that the well‐known 1.883 and 2.51 eV systems anneal out during the temperature regime of nitrogen aggregation and the concurrent growth of an array of structure, which extends throughout the visible region of the absorption spectrum, and gives the previously bright‐yellow‐colored diamonds a rich golden‐yellow color, is observed. By carrying out the annealing sequence on diamonds grown using various solvent catalysts a correlation is found between preanneal 1.883 eV absorption and maximized absorption of some of the annealed structure. From the results it is proposed that centers associated with nickel and nitrogen are produced and their possible natures are speculated on. It is found that the defect‐induced one‐phonon spectra of the diamonds examined may be satisfactorily decomposed into three components. To account for certain changes in the infrared spectra during the annealing sequences, and using previously reported results, it is proposed that one of these components may result from nitrogen in a positive charge state.
Simon C Lawson - One of the best experts on this subject based on the ideXlab platform.
-
an annealing study of nickel point defects in high pressure synthetic diamond
Journal of Applied Physics, 1993Co-Authors: Simon C Lawson, Hisao KandaAbstract:Results of an annealing study, carried out in the temperature range 1500–1900 °C, of nickel‐related optical centers in high‐pressure synthetic diamond are presented. It is established that the well‐known 1.883 and 2.51 eV systems anneal out during the temperature regime of nitrogen aggregation and the concurrent growth of an array of structure, which extends throughout the visible region of the absorption spectrum, and gives the previously bright‐yellow‐colored diamonds a rich golden‐yellow color, is observed. By carrying out the annealing sequence on diamonds grown using various solvent catalysts a correlation is found between preanneal 1.883 eV absorption and maximized absorption of some of the annealed structure. From the results it is proposed that centers associated with nickel and nitrogen are produced and their possible natures are speculated on. It is found that the defect‐induced one‐phonon spectra of the diamonds examined may be satisfactorily decomposed into three components. To account for cer...
-
an annealing study of nickel point defects in high pressure synthetic diamond
Journal of Applied Physics, 1993Co-Authors: Simon C Lawson, Hisao KandaAbstract:Results of an annealing study, carried out in the temperature range 1500–1900 °C, of nickel‐related optical centers in high‐pressure synthetic diamond are presented. It is established that the well‐known 1.883 and 2.51 eV systems anneal out during the temperature regime of nitrogen aggregation and the concurrent growth of an array of structure, which extends throughout the visible region of the absorption spectrum, and gives the previously bright‐yellow‐colored diamonds a rich golden‐yellow color, is observed. By carrying out the annealing sequence on diamonds grown using various solvent catalysts a correlation is found between preanneal 1.883 eV absorption and maximized absorption of some of the annealed structure. From the results it is proposed that centers associated with nickel and nitrogen are produced and their possible natures are speculated on. It is found that the defect‐induced one‐phonon spectra of the diamonds examined may be satisfactorily decomposed into three components. To account for certain changes in the infrared spectra during the annealing sequences, and using previously reported results, it is proposed that one of these components may result from nitrogen in a positive charge state.
Michael Bader - One of the best experts on this subject based on the ideXlab platform.
-
Vectorization and Minimization of Memory Footprint for Linear High-Order Discontinuous Galerkin Schemes
arXiv: Mathematical Software, 2020Co-Authors: Jean-matthieu Gallard, Leonhard Rannabauer, Anne Reinarz, Michael BaderAbstract:We present a sequence of optimizations to the performance-critical compute kernels of the high-order discontinuous Galerkin solver of the hyperbolic PDE engine ExaHyPE -- successively tackling bottlenecks due to SIMD operations, cache hierarchies and restrictions in the software design. Starting from a generic scalar implementation of the numerical scheme, our first optimized variant applies state-of-the-art optimization techniques by vectorizing loops, improving the data layout and using Loop-over-GEMM to perform tensor contractions via highly optimized matrix multiplication functions provided by the LIBXSMM library. We show that memory stalls due to a memory footprint exceeding our L2 cache size hindered the vectorization gains. We therefore introduce a new kernel that applies a sum factorization approach to reduce the kernel's memory footprint and improve its cache locality. With the L2 cache bottleneck removed, we were able to exploit additional vectorization opportunities, by introducing a hybrid array-of-structure-of-array data layout that solves the data layout conflict between matrix multiplications kernels and the point-wise functions to implement PDE-specific terms. With this last kernel, evaluated in a benchmark simulation at high polynomial order, only 2\% of the floating point operations are still performed using scalar instructions and 22.5\% of the available performance is achieved.
-
IPDPS Workshops - Vectorization and Minimization of Memory Footprint for Linear High-Order Discontinuous Galerkin Schemes
2020 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), 2020Co-Authors: Jean-matthieu Gallard, Leonhard Rannabauer, Anne Reinarz, Michael BaderAbstract:We present a sequence of optimizations to the performance-critical compute kernels of the high-order discontinuous Galerkin solver of the hyperbolic PDE engine ExaHyPE – successively tackling bottlenecks due to SIMD operations, cache hierarchies and restrictions in the software design.Starting from a generic scalar implementation of the numerical scheme, our first optimized variant applies state-of the-art optimization techniques by vectorizing loops, improving the data layout and using Loop-over-GEMM to perform tensor contractions via highly optimized matrix multiplication functions provided by the LIBXSMM library. We show that memory stalls due to a memory footprint exceeding our L2 cache size hindered the vectorization gains. We therefore introduce a new kernel that applies a sum factorization approach to reduce the kernel’s memory footprint and improve its cache locality. With the L2 cache bottleneck removed, we were able to exploit additional vectorization opportunities, by introducing a hybrid array-of-structure-of-array data layout that solves the data layout conflict between matrix multiplications kernels and the point-wise functions to implement PDE-specific terms.With this last kernel, evaluated in a benchmark simulation at high polynomial order, only 2% of the floating point operations are still performed using scalar instructions and 22.5% of the available performance is achieved.
Jean-matthieu Gallard - One of the best experts on this subject based on the ideXlab platform.
-
Vectorization and Minimization of Memory Footprint for Linear High-Order Discontinuous Galerkin Schemes
arXiv: Mathematical Software, 2020Co-Authors: Jean-matthieu Gallard, Leonhard Rannabauer, Anne Reinarz, Michael BaderAbstract:We present a sequence of optimizations to the performance-critical compute kernels of the high-order discontinuous Galerkin solver of the hyperbolic PDE engine ExaHyPE -- successively tackling bottlenecks due to SIMD operations, cache hierarchies and restrictions in the software design. Starting from a generic scalar implementation of the numerical scheme, our first optimized variant applies state-of-the-art optimization techniques by vectorizing loops, improving the data layout and using Loop-over-GEMM to perform tensor contractions via highly optimized matrix multiplication functions provided by the LIBXSMM library. We show that memory stalls due to a memory footprint exceeding our L2 cache size hindered the vectorization gains. We therefore introduce a new kernel that applies a sum factorization approach to reduce the kernel's memory footprint and improve its cache locality. With the L2 cache bottleneck removed, we were able to exploit additional vectorization opportunities, by introducing a hybrid array-of-structure-of-array data layout that solves the data layout conflict between matrix multiplications kernels and the point-wise functions to implement PDE-specific terms. With this last kernel, evaluated in a benchmark simulation at high polynomial order, only 2\% of the floating point operations are still performed using scalar instructions and 22.5\% of the available performance is achieved.
-
IPDPS Workshops - Vectorization and Minimization of Memory Footprint for Linear High-Order Discontinuous Galerkin Schemes
2020 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), 2020Co-Authors: Jean-matthieu Gallard, Leonhard Rannabauer, Anne Reinarz, Michael BaderAbstract:We present a sequence of optimizations to the performance-critical compute kernels of the high-order discontinuous Galerkin solver of the hyperbolic PDE engine ExaHyPE – successively tackling bottlenecks due to SIMD operations, cache hierarchies and restrictions in the software design.Starting from a generic scalar implementation of the numerical scheme, our first optimized variant applies state-of the-art optimization techniques by vectorizing loops, improving the data layout and using Loop-over-GEMM to perform tensor contractions via highly optimized matrix multiplication functions provided by the LIBXSMM library. We show that memory stalls due to a memory footprint exceeding our L2 cache size hindered the vectorization gains. We therefore introduce a new kernel that applies a sum factorization approach to reduce the kernel’s memory footprint and improve its cache locality. With the L2 cache bottleneck removed, we were able to exploit additional vectorization opportunities, by introducing a hybrid array-of-structure-of-array data layout that solves the data layout conflict between matrix multiplications kernels and the point-wise functions to implement PDE-specific terms.With this last kernel, evaluated in a benchmark simulation at high polynomial order, only 2% of the floating point operations are still performed using scalar instructions and 22.5% of the available performance is achieved.
Bader Michael - One of the best experts on this subject based on the ideXlab platform.
-
Vectorization and Minimization of Memory Footprint for Linear High-Order Discontinuous Galerkin Schemes
2020Co-Authors: Gallard Jean-matthieu, Rannabauer Leonhard, Reinarz Anne, Bader MichaelAbstract:We present a sequence of optimizations to the performance-critical compute kernels of the high-order discontinuous Galerkin solver of the hyperbolic PDE engine ExaHyPE -- successively tackling bottlenecks due to SIMD operations, cache hierarchies and restrictions in the software design. Starting from a generic scalar implementation of the numerical scheme, our first optimized variant applies state-of-the-art optimization techniques by vectorizing loops, improving the data layout and using Loop-over-GEMM to perform tensor contractions via highly optimized matrix multiplication functions provided by the LIBXSMM library. We show that memory stalls due to a memory footprint exceeding our L2 cache size hindered the vectorization gains. We therefore introduce a new kernel that applies a sum factorization approach to reduce the kernel's memory footprint and improve its cache locality. With the L2 cache bottleneck removed, we were able to exploit additional vectorization opportunities, by introducing a hybrid array-of-structure-of-array data layout that solves the data layout conflict between matrix multiplications kernels and the point-wise functions to implement PDE-specific terms. With this last kernel, evaluated in a benchmark simulation at high polynomial order, only 2\% of the floating point operations are still performed using scalar instructions and 22.5\% of the available performance is achieved.Comment: PDSEC 202