Embedded memory subsystems for a CNN based processing unit and methods of making

Patent metadata

Field	Value
Publication number	US-10545693-B2
Application number	US-201916403676-A
Country	US
Kind code	B2
Filing date	May 6, 2019
Priority date	Apr 3, 2017
Publication date	Jan 28, 2020
Grant date	Jan 28, 2020

Abstract

Embedded memory subsystems in a digital integrated circuit for artificial intelligence are disclosed. A semi-conductor substrate contains CNN processing units. Each CNN processing unit includes CNN logic circuits and an embedded memory subsystem. The memory subsystem includes first embedded memory and second embedded memory. The first embedded memory contains an array of MTJ STT-RAM cells with each cell has a circular planar area with a diameter in a range of 40-120 nm. The second embedded memory contains an array of MTJ STT-RAM cells with each cell has a circular planar area having a diameter in a range of 30-75 nm.

First claim

What is claimed is:

1. A digital semi-conductor chip for artificial intelligence comprising:
a semi-conductor substrate containing integrated circuits of a plurality of cellular neural networks (CNN) processing units;
each CNN processing unit comprising CNN logic circuits, a first embedded memory, and a second embedded memory;
the first embedded memory containing an array of first spin transfer torque magnetic random access memory (STT-RAM) cells with each first STT-RAM cell having a first magnetic tunnel junction (MTJ) element located on a first transistor;
the second embedded memory containing an array of second STT-RAM cells with each second STT-RAM cell having a second MTJ element located on a second transistor; and
the CNN logic circuits further comprising at least a metal layer that includes a plurality of copper landing pads providing as the first and the second transistors.

2. The digital integrated circuit of claim 1 , further comprises at least one input/output data bus operatively coupling the plurality of the CNN processing units via the first embedded memory and the second embedded memory and a controller for controlling operations of the plurality of CNN processing units.

3. The digital integrated circuit of claim 2 , wherein the semi-conductor substrate comprises a silicon substrate.

4. The digital integrated circuit of claim 2 , wherein the first embedded memory is configured for requiring data being stored with higher retention rate than the second embedded memory.

5. The digital integrated circuit of claim 2 , wherein the second embedded memory is configured for higher endurance of balanced data read and write operations than the first embedded memory.

6. The digital integrated circuit of claim 1 , wherein each of the first MTJ element and the second MTJ element is made of two ferromagnetic layers located on either side of an oxide barrier layer orientated vertically.

7. The digital integrated circuit of claim 1 , wherein each of the first MTJ element and the second MTJ element operates in a Current Perpendicular to Plane configuration.

8. The digital integrated circuit of claim 1 , wherein the first MTJ element has a circular planar area with a first diameter in a range of 40-120 nm and the second MTJ element has a circular planar area with a second diameter in a range of 30-75 nm.

9. The digital integrated circuit of claim 1 , wherein the first embedded memory and the second embedded memory have a same pitch.

10. The digital integrated circuit of claim 1 , wherein the first memory and the second embedded memory have different pitches.

11. The digital semi-conductor chip for artificial intelligence of claim 1 , wherein the first transistor and the second transistor are independent with each other.

Assignees

• Gyrfalcon Tech Inc

Inventors

• Torng Chyu-Jiuh
• Yang Lin
• Dong Qi

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06F13/16

  for access to memory bus (G06F13/28 takes precedence) · CPC title

• G06F13/4068

  Electrical coupling · CPC title

• G06F3/0688

  Non-volatile semiconductor memory arrays · CPC title

• G11C11/161 Primary

  details concerning the memory cell structure, e.g. the layers of the ferromagnetic memory cell · CPC title