IP Library Granted Patent US 8,933,942
Granted Patent B2
US 8,933,942 · App. 13/315,121 · Granted Jan 13, 2015

Partitioning resources of a processor

Inventors: Robert Scott Hartog (Windemere, FL); Ralph Clay Taylor (Deland, FL); Michael Mantor (Orlando, FL); Thomas Roy Woller (Austin, TX); Kevin McGrath (Los Gatos, CA); Rex McCrary (Oviedo, FL); Philip J. Rogers (Pepperell, MA); Mark Leather (Los Gatos, CA)
Assignee: Advanced Micro Devices, Inc.
G06F9/505
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 8,933,942
App. No.
13/315,121
Granted
Jan 13, 2015
Kind
B2
Abstract

Embodiments describe herein provide an apparatus, a computer readable medium and a method for simultaneously processing tasks within an APD. The method includes processing a first task within an APD. The method also includes reducing utilization of the APD by the first task to facilitate simultaneous processing of the second task, such that the utilization remains below a threshold.

Claims (28)

1. A method comprising:

processing a first task received by a processor within an accelerated processing device (APD); and

reducing utilization of the processor by the first task to facilitate simultaneous processing of a second task, until the reduced utilization equals an allotted utilization of the second task, wherein a combined utilization by the first task and the second task remains below a predetermined utilization threshold.

2. The method of claim 1 , wherein the first and second tasks are received from a central processing unit (CPU).

3. The method of claim 1 , wherein the first and second tasks include a plurality of wavefronts.

4. The method of claim 1 , wherein the threshold is determined in accordance with a number of local data shares available, or a predetermined allotment of APD resources.

5. The method of claim 1 , wherein the reducing utilization of the processor is performed by removing a wavefront of a plurality of wavefronts associated with the first task and temporarily context switching the wavefront to a memory.

6. The method of claim 5 , further comprising storing the wavefront of the plurality of wavefronts of the first task in the memory last one in last one out.

7. The method of claim 5 , further comprising maintaining, using a shader pipeline input, a record of the wavefront of the plurality of wavefronts that is temporarily context switched.

8. The method of claim 1 , wherein the threshold is determined in accordance with a number of single command memory devices in use.

9. The method of claim 1 , wherein the reducing utilization of the processor is performed by ceasing processing of portions of the first task.

10. The method of claim 9 , wherein the portions include one or more wavefronts from a plurality of wavefronts of the first task.

11. The method of claim 9 , wherein the ceasing processing includes removing one or more last sent wavefronts from the plurality of wavefronts of the first task.

12. A non-transitory computer readable media storing commands wherein the commands when executed by a processor are adapted to process work-items on an APD to perform operations comprising:

processing a first task by the processor within the APD; and

reducing utilization of the processor by the first task to facilitate simultaneous processing of a second task, until the reduced utilization equals an allotted utilization of the second task, wherein a combined utilization by the first task and the second task remains below a predetermined utilization threshold.

13. The non-transitory computer readable media of claim 12 , wherein the threshold is determined in accordance with a number of single command memory devices in use.

14. An apparatus, comprising:

a memory; and

a processor of an APD coupled to the memory, wherein the processor is configured to, based on a command stored in the memory:

process a first task; and

reduce utilization of the processor by the first task to facilitate simultaneous processing of a second task, until the reduced utilization equals an allotted utilization of the second task, wherein the combined utilization by the first task and the second task remains below a predetermined utilization threshold.

15. The apparatus of claim 14 , wherein the APD receives the first and second tasks from a CPU.

16. The apparatus of claim 14 , wherein the first and second tasks include a plurality of wavefronts.

17. The apparatus of claim 14 , wherein the threshold is a number of local data shares available, or a predetermined allotment of APD resources.

18. The apparatus of claim 14 , wherein the reduced utilization of the processor is performed by removing a wavefront of a plurality of wavefronts associated with the first task and temporarily context switching the wavefront to a memory.

19. The apparatus of claim 14 , wherein the reducing utilization of the processor is performed by ceasing processing of portions of the first task, wherein the portions include one or more wavefronts from a plurality of wavefronts of the first task.

20. The apparatus of claim 19 , wherein the ceasing processing includes removing one or more last sent wavefronts from the plurality of wavefronts of the first task.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded May 4, 2012
From: HARTOG, ROBERT SCOTT; TAYLOR, RALPH CLAY; MANTOR, MICHAEL; WOLLER, THOMAS; MCGRATH, KEVIN; MCCRARY, REX; ROGERS, PHILIP J.; LEATHER, MARK
To: ADVANCED MICRO DEVICES, INC.
Reel/Frame 028154/0384 →
Continuity (1)
Related Publication 20130147816A1 · Jun 13, 2013