IP Library › Granted Patent US 12,603,829
Granted Patent B2
US 12,603,829 · App. 18/540,782 · Granted Apr 14, 2026

Network collective offload message chunking management

Inventor: Josiah I. Clark (Austin, TX)
Assignee: Advanced Micro Devices, Inc.
H04L45/02H04L41/0823H04L41/12H04L45/34H04L49/30
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 12,603,829
App. No.
18/540,782
Granted
Apr 14, 2026
Kind
B2
Abstract

The disclosed device can perform a collective operation on received datasets, and split the result into chunks in accordance with a chunking scheme. The device can also forward the chunks in accordance with a routing scheme that can direct chunks to appropriate nodes of a collective network. Various other methods, systems, and computer-readable media are also disclosed.

Claims (38)

1 . A method comprising:

receiving a plurality of datasets;

performing a collective operation of a collective network on the plurality of datasets to produce a result dataset;

split the result dataset into a plurality of chunks using a chunking scheme that correlates index values of the plurality of datasets to a plurality of destinations in a round-robin style; and

send the plurality of chunks to the plurality of destinations based on the chunking scheme and a topology corresponding to the collective operation that defines paths for chunks traversing the collective network such that corresponding chunks reach a same destination.

2 . The method of claim 1 , wherein the topology defines connections between nodes of a collective network based on the collective operation such that the plurality of datasets are received from a prior level of the collective network and the plurality of chunks are sent to a subsequent level of the collective network.

3 . The method of claim 1 , wherein a number of the plurality of datasets corresponds to a number of input ports and a number of the plurality of chunks corresponds to a number of a plurality of output ports.

4 . The method of claim 1 , wherein the paths for chunks traversing the collective network ensures that chunks of matching vector indexes reach the same destination for the collective operation.

5 . A device comprising:

a plurality of output ports; and

a control circuit configured to:

perform a collective operation on a dataset to produce a result dataset;

split the result dataset into a plurality of chunks using a chunking scheme that correlates index values of the dataset to the plurality of output ports in a round-robin style; and

send the plurality of chunks through the plurality of output ports based on the chunking scheme.

6 . The device of claim 5 , wherein a number of the plurality of chunks matches a number of the plurality of output ports such that each of the plurality of chunks are sent through a respective one of the plurality of output ports.

7 . The device of claim 5 , wherein the device is part of collective network such that the plurality of output ports are connected to nodes of a subsequent level of the collective network.

8 . The device of claim 7 , wherein the plurality of output ports are connected to the nodes of the subsequent level based on a topology that defines paths for chunks traversing the collective network to send each of the plurality of chunks to destinations defined by the topology such that corresponding chunks reach a same destination.

9 . The device of claim 8 , wherein the topology is enforced using segment routing of the plurality of chunks.

10 . The device of claim 8 , wherein the topology defines connections between nodes based on the collective operation.

11 . The device of claim 5 , further comprising a plurality of input ports, wherein the control circuit is further configured to:

receive, through the plurality of input ports, a plurality of datasets; and

perform the collective operation on the plurality of datasets to produce the result dataset.

12 . The device of claim 11 , wherein a number of the plurality of datasets matches a number of the plurality of input ports such that each of the plurality of datasets are received from a respective one of the plurality of input ports.

13 . The device of claim 11 , wherein the device is part of collective network such that the plurality of input ports are connected to nodes of a prior level of the collective network.

14 . The device of claim 13 , wherein the plurality of datasets corresponds to chunks sent from the nodes of the prior level of the collective network.

15 . A device comprising:

a plurality of input ports;

a plurality of output ports; and

a control circuit configured to:

receive, through the plurality of input ports, a plurality of datasets;

perform a collective operation on the plurality of datasets to produce a result dataset;

split the result dataset into a plurality of chunks using a chunking scheme that correlates index values of the plurality of datasets to the plurality of output ports in a round-robin style; and

send the plurality of chunks through the plurality of output ports based on the chunking scheme.

16 . The device of claim 15 , wherein a number of the plurality of datasets matches a number of the plurality of input ports such that each of the plurality of datasets are received from a respective one of the plurality of input ports.

17 . The device of claim 15 , wherein a number of the plurality of chunks matches a number of the plurality of output ports such that each of the plurality of chunks are sent through a respective one of the plurality of output ports.

18 . The device of claim 15 , wherein the device is part of collective network such that the plurality of input ports are connected to nodes of a prior level of the collective network and the plurality of output ports are connected to nodes of a subsequent level of the collective network.

19 . The device of claim 18 , wherein the plurality of datasets corresponds to chunks sent from the nodes of the prior level of the collective network.

20 . The device of claim 18 , wherein the plurality of input ports are connected to the nodes of the prior level and the plurality of output ports are connected to the nodes of the subsequent level based on a topology that defines paths for chunks traversing the collective network and the plurality of chunks are sent through the plurality of output ports to destinations defined by the topology such that corresponding chunks reach a same destination.

Assignments (2)
CORRECTIVE ASSIGNMENT TO CORRECT THE APPLICATION NUMBER 18540779 PREVIOUSLY RECORDED AT REEL: 67116 FRAME: 127. ASSIGNOR(S) HEREBY CONFIRMS THE ASSIGNMENT . Recorded Jul 15, 2024
From: CLARK, JOSIAH I.
To: ADVANCED MICRO DEVICES, INC.
Reel/Frame 068376/0705 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Apr 16, 2024
From: CLARK, JOSIAH I.
To: ADVANCED MICRO DEVICES, INC.
Reel/Frame 067116/0127 →
Continuity (2)
Provisional Application 63432657 · Dec 14, 2022
Related Publication 20240205133A1 · Jun 20, 2024
References Cited (22)
US 20030137940A1 · Schwartz · 2003 [cited by examiner]
US 20040205242A1 · Xu · 2004 [cited by examiner]
US 20050169179A1 · Antal et al. · 2005 [cited by applicant]
US 20060182118A1 · Lam · 2006 [cited by examiner]
US 20070189283A1 · Agarwal · 2007 [cited by examiner]
US 20140059556A1 · Barsness et al. · 2014 [cited by applicant]
US 20140149715A1 · Inman · 2014 [cited by examiner]
US 20140279883A1 · Kostenko · 2014 [cited by applicant]
US 20180268296A1 · Zheng et al. · 2018 [cited by applicant]
US 20190042527A1 · Langer · 2019 [cited by examiner]
US 20190340263A1 · Porat-Stoler et al. · 2019 [cited by applicant]
US 20200322258A1 · Oprea · 2020 [cited by examiner]
US 20220100752A1 · Rastogi · 2022 [cited by applicant]
US 20220200906A1 · Ye et al. · 2022 [cited by applicant]
US 20220366220A1 · Roth et al. · 2022 [cited by applicant]
US 20220393698A1 · Sethi et al. · 2022 [cited by applicant]
US 20230177321A1 · Sanchez · 2023 [cited by examiner]
CN 111898730A · 2020 [cited by applicant]
JP 2018207182A · 2018 [cited by applicant]
International Search Report and Written Opinion dated Apr. 2, 2024, from corresponding International Application No. PCT/US2023/084145; 10 pages. [cited by applicant]
International Search Report and Written Opinion, dated Mar. 27, 2024, from International Application No. PCT/US2023/084144 (10 pages), by Yang, Jeong Rok. [cited by applicant]
International Search Report and Written Opinion, dated Apr. 8, 2024, from International Application No. PCT/US2023/084146; (10 pages), by Yang, Jeong Rok. [cited by applicant]
Cited By (5)
US 12,726,430 US 12,739,192 US 12,739,193 US 12,739,194 US 12,739,195