An Extact Hardware Implementation of the Boltzmann Machine

نویسنده

  • Marcin Skubiszewski
چکیده

We present a faithful hardware implementation of the Boltzmann machine. The prototype performs 505 megasynapses (millionof additionsand multiplications)per second, using 16-bit fixed-point weights. It can emulate fully connected instances of the Boltzmann machine containing up to 1438 variables. Our specialized hardware only executes the simplest part of the Boltzmann machine algorithm, namely multiplying matrices of numbers by vectors of bits. The other operations (which are complicated, but only require a modest amount of computation) are performed by the host processor.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Field Programmable Gate Array–based Implementation of an Improved Algorithm for Objects Distance Measurement (TECHNICAL NOTE)

In this work, the design of a low-cost, field programmable gate array (FPGA)-based digital hardware platform that implements image processing algorithms for real-time distance measurement is presented. Using embedded development kit (EDK) tools from Xilinx, the system is developed on a spartan3 / xc3s400, one of the common and low cost field programmable gate arrays from the Xilinx Spartan fami...

متن کامل

An Exact Hardware Implementation of the Boltzmann Machine

We present a fast implementation of the Boltzmann Machine based on specialized hardware. This realization faithfully implements the machine; in particular, it avoids the excess parallelism which makes other fast implementations of it only approximate. The current prototype performs 505 million additions and multiplications (or megasynapses) per second. It can emulate the fully connected Boltzma...

متن کامل

Implementation of D3Q19 Lattice Boltzmann Method with a Curved Wall Boundary Condition for Simulation of Practical Flow Problems

In this paper, implementation of an extended form of a no-slip wall boundary condition is presented for the three-dimensional (3-D) lattice Boltzmann method (LBM) for solving the incompressible fluid flows with complex geometries. The boundary condition is based on the off-lattice scheme with a polynomial interpolation which is used to reconstruct the curved or irregular wall boundary on the ne...

متن کامل

Implementation of a Restricted Boltzmann Machine in a Spiking Neural Network

Restricted Boltzmann Machines (RBMs) have been demonstrated to perform efficiently on a variety of applications, such as dimensionality reduction and classification. Implementing RBMs on neuromorphic hardware has certain advantages, particularly from a concurrency and lowpower perspective. This paper outlines some of the requirements involved for neuromorphic adaptation of an RBM and attempts t...

متن کامل

Practical Implementation of Scalar and Vector Control Methods on a Rotor Surface Type Permanent Magnent Synchronous Machine Drive/System Using a PC

In this paper, using a personal computer (PC), the practical implementation of scalar and vector control methods on a three–phase rotor surface- type permanent magnet synchronous machine drive is discussed. Based on the machine dynamic equations and the above control strategies, two block diagrams are presented first for closed-loop speed controlling of the machine drive/system. Then, the desig...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1992