IP Library Granted Patent US 11,782,879
Granted Patent B2
US 11,782,879 · App. 17/234,007 · Granted Oct 10, 2023

System and method for secure, fast communications between processors on complex chips

Inventors: Joshua Cooper (Columbia, SC); Charles Yeomans (Orinda, CA); Aliasghar Riahi (Orinda, CA); Gregory Caltabiano (Palo Alto, CA); Mojgan Haddad (Orinda, CA)
Assignee: ATOMBEAM TECHNOLOGIES INC.
G06F16/1752G06F3/067G06F3/0608G06F3/0641
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,782,879
App. No.
17/234,007
Filed
Apr 19, 2021
Granted
Oct 10, 2023
Kind
B2
Art Unit
2136
USPC
707/692
Abstract

A system and method for fast communication between processors on complex chips using encoding, wherein a training data set is used to find patterns and associated smaller indices, or codewords, which are stored in a reference codebook library, and where reconstruction and deconstruction algorithms are used to encode and decode data as it is received. The codebook and algorithms may be stored in the firmware of a semiconductor which enable reduced resources and cost when transmitting data between or among devices that utilize such semiconductors.

Claims (29)

1. A system for secure and fast communication between processors on complex chips, comprising:

a complex chip comprising at least two processors and a memory for each processor;

wherein each of a first processor and a second processor on the complex chip stores a pre-trained reference codebook embedded as firmware in its respective memory, wherein:

the pre-trained reference codebook is generated by using training data to identify chunklets and their associated codewords, and storing the associated codewords in the pre-trained reference codebook; and

codewords within the pre-trained reference codebook are assigned such that more frequently-occurring chunklets are assigned shorter codewords based on a combination of:

an observed frequency of occurrence of chunklets in the training data; and

an estimated frequency of occurrence of chunklets not in the training;

wherein each of a first processor and a second processor on the complex chip stores a deconstruction algorithm embedded as firmware in in its respective memory, wherein the deconstruction algorithm, when operating on the first processor, causes the first processor to:

receive data;

deconstruct the data into a plurality of chunklets; and

encode the data by retrieving the codeword for each chunklet from the pre-trained reference codebook; and

wherein each of a first processor and a second processor on the complex chip stores a reconstruction algorithm embedded as firmware in in its respective memory, wherein the reconstruction algorithm, when operating on the second processor, causes the second processor to:

receive a codeword;

retrieve a chunklet for each received codeword from the pre-trained reference codebook; and

assemble the chunklets to reconstruct the data.

2. A method for secure and fast communication between processors on complex chips, comprising the steps of:

storing a pre-trained reference codebook embedded as firmware in the memory of a first processor and in the memory of a second processor, wherein:

the pre-trained reference codebook is generated by using training data that identifies chunklets and their associated codewords, and storing the associated codewords in the pre-trained reference codebook; and

codewords within the pre-trained reference codebook are assigned such that more frequently-occurring chunklets are assigned shorter codewords based on a combination of:

an observed frequency of occurrence of chunklets in the training data; and

an estimated frequency of occurrence of chunklets not in the training;

storing a deconstruction algorithm embedded as firmware in the memory of the first processor-and in the memory of the second processor, wherein the deconstruction algorithm, when operating on the first processor, causes the first processor to:

receive data;

deconstruct the data into a plurality of chunklets; and

encode the data by retrieving the codeword for each chunklet from the pre-trained reference codebook; and

storing a reconstruction algorithm embedded as firmware in the memory of the first processor and in the memory of the second processor, wherein the reconstruction algorithm, when operating on the second processor, causes the second processor to:

receiving a codeword;

retrieving the chunklet for each received codeword from the pre-trained reference codebook; and

assembling the chunklets to reconstruct the data.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 30, 2021
From: COOPER, JOSHUA; YEOMANS, CHARLES; RIAHI, ALIASGHAR; CALTABIANO, GREGORY; HADDAD, MOJGAN
To: ATOMBEAM TECHNOLOGIES INC.
Reel/Frame 057332/0958 →
Continuity (11)
Continuation In Part 17180439 · Feb 19, 2021
Continuation In Part 16923039 · Jul 7, 2020
Continuation In Part 16716098 · Dec 16, 2019
Continuation 16455655 · Jun 27, 2019
Continuation In Part 16200466 · Nov 26, 2018
Continuation In Part 15975741 · May 9, 2018
Provisional Application 63140111 · Jan 21, 2021
Provisional Application 63027166 · May 19, 2020
Provisional Application 62926723 · Oct 28, 2019
Provisional Application 62578824 · Oct 30, 2017
Related Publication 20210232544A1 · Jul 29, 2021