William P. Risk
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
William P. Risk has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Andreopoulos
IBM Research, San Jose, CA, USA
1 shared publication - Andrew S. Cassidy
IBM Research Almaden 95120 San Jose CA
1 shared publication - Arnon Amir
IBM Research, San Jose, CA, USA
1 shared publication Ben Shaw
IBM Research, San Jose, CA, USA
1 shared publication- Brian Taba
IBM Research, San Jose, CA, USA
1 shared publication Carlos Ortega Otero
IBM Research, San Jose, CA, USA
1 shared publication
Latest Publications
- 20 October 2023
- Published by American Association for the Advancement of Science (AAAS) in Science
- Vol. 382 (6668) , 329-335
Neural inference at the frontier of energy, space, and time
Computing, since its inception, has been processor-centric, with memory separated from compute. Inspired by the organic brain and optimized for inorganic silicon, NorthPole is a neural inference architecture that blurs this boundary by eliminating off-chip memory, intertwining compute with memory on-chip, and appearing externally as an active memory chip. NorthPole is a low-precision, massively parallel, densely interconnected, energy-efficient, and spatial computing architecture with a co-optimized, high-utilization programming model. On the ResNet50 benchmark image classification network, relative to a graphics processing unit (GPU) that uses a comparable 12-nanometer technology process, NorthPole achieves a 25 times higher energy metric of frames per second (FPS) per watt, a 5 times higher space metric of FPS per transistor, and a 22 times lower time metric of latency. Similar results are reported for the Yolo-v4 detection network. NorthPole outperforms all prevalent architectures, even those that use more-advanced technology processes.
- 20 October 2023
- Published by American Association for the Advancement of Science (AAAS) in Science
- Vol. 382 (6668) , 329-335
Neural inference at the frontier of energy, space, and time
Computing, since its inception, has been processor-centric, with memory separated from compute. Inspired by the organic brain and optimized for inorganic silicon, NorthPole is a neural inference architecture that blurs this boundary by eliminating off-chip memory, intertwining compute with memory on-chip, and appearing externally as an active memory chip. NorthPole is a low-precision, massively parallel, densely interconnected, energy-efficient, and spatial computing architecture with a co-optimized, high-utilization programming model. On the ResNet50 benchmark image classification network, relative to a graphics processing unit (GPU) that uses a comparable 12-nanometer technology process, NorthPole achieves a 25 times higher energy metric of frames per second (FPS) per watt, a 5 times higher space metric of FPS per transistor, and a 22 times lower time metric of latency. Similar results are reported for the Yolo-v4 detection network. NorthPole outperforms all prevalent architectures, even those that use more-advanced technology processes.
- 1 June 1999
- Published by American Physical Society (APS) in Physical Review B
- Vol. 59 (22) , 14259-14264
Studies of periodic ferroelectric domains inusing high-resolution x-ray scattering and diffraction imagingA periodically domain-inverted crystal, in which inverted ferroelectric domains were produced by electric-field poling, has been studied using high-resolution x-ray scattering and diffraction imaging techniques. The degree of perfection of domain inversion as well as the periodic structure has been revealed in high-spatial-resolution images, and the periodic contrast is shown to arise largely from lattice distortions produced in the poled regions. The dependence of intensity scattered from the domain-inverted structure can be described by a two-component line shape, one with a sharp nearly-pure-Gaussian line, the other with a broad pseudo-Voigt-function profile. A high-angular-resolution method has been combined to elucidate the origin of the two-component profile. The diffuse scattering of the crystal truncation rod along is observed to follow a power law with after poling compared with an average value of before poling. Discussions of the electric-field poling and its consequence for the domain-inverted structure are presented in terms of the results of x-ray measurements.- 1 June 1999
- Published by American Physical Society (APS) in Physical Review B
- Vol. 59 (22) , 14259-14264
Studies of periodic ferroelectric domains inusing high-resolution x-ray scattering and diffraction imagingA periodically domain-inverted crystal, in which inverted ferroelectric domains were produced by electric-field poling, has been studied using high-resolution x-ray scattering and diffraction imaging techniques. The degree of perfection of domain inversion as well as the periodic structure has been revealed in high-spatial-resolution images, and the periodic contrast is shown to arise largely from lattice distortions produced in the poled regions. The dependence of intensity scattered from the domain-inverted structure can be described by a two-component line shape, one with a sharp nearly-pure-Gaussian line, the other with a broad pseudo-Voigt-function profile. A high-angular-resolution method has been combined to elucidate the origin of the two-component profile. The diffuse scattering of the crystal truncation rod along is observed to follow a power law with after poling compared with an average value of before poling. Discussions of the electric-field poling and its consequence for the domain-inverted structure are presented in terms of the results of x-ray measurements.- 1 August 1990
- Published by SPIE-Intl Soc Optical Eng
- Vol. 1316, 194-199
Compact blue lasers for optical recording applications
Compact blue laser sources with output powers of 10 and 40 mW at 473 and 428 nm have been developed using nonlinear frequency upconversion of GaAlAs diode lasers. These devices are attractive for application to high-density optical data storage.
- 1 August 1990
- Published by SPIE-Intl Soc Optical Eng
- Vol. 1316, 194-199
Compact blue lasers for optical recording applications
Compact blue laser sources with output powers of 10 and 40 mW at 473 and 428 nm have been developed using nonlinear frequency upconversion of GaAlAs diode lasers. These devices are attractive for application to high-density optical data storage.