IP Library Granted Patent US 9,606,924
Granted Patent B2
US 9,606,924 · App. 13/530,802 · Granted Mar 28, 2017

System and method for supporting fast and deterministic execution and simulation in multi-core environments

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 9,606,924
App. No.
13/530,802
Granted
Mar 28, 2017
Kind
B2
Abstract

The exemplary embodiments described herein relate to supporting fast and deterministic execution and simulation in multi-core environments. Specifically, the exemplary embodiments relate to systems and methods for implementing determinism in a memory system of a multithreaded computer. A exemplary system comprises a plurality of processors within a multi-processor environment, a cache memory within the processor and including metadata, and a hardware check unit performing one of a load check and a store check on the metadata to detect a respective one of a load metadata mismatch and a store metadata mismatch, and invoking a runtime software routine to order memory references upon a detection of one of the load metadata mismatch and the store metadata mismatch.

Claims (43)

1. A non-transitory computer readable storage medium including a set of instructions executable by a processor, the set of instructions, when executed, resulting in a performance of the following:

inserting metadata into a cache memory of a first core in a system including a plurality of cores;

performing one of a load check and a store check for a thread to be executed by a first one of the cores on the metadata to detect a respective one of a load metadata mismatch and a store metadata mismatch, wherein the load metadata mismatch is a mismatch between metadata of a load to be executed by the first core and core metadata, and wherein the store metadata mismatch is a mismatch between metadata of a store to be executed by the first core and the core metadata, further wherein the core metadata identifies at least one of the cores with permission to execute the load and at least one of the cores with permission to execute the store; and

invoking a runtime software routine to order memory references upon a detection of one of the load metadata mismatch and the store metadata mismatch.

2. The non-transitory computer readable storage medium of claim 1 , wherein the invoking the runtime software routine includes transferring control to a runtime library.

3. The non-transitory computer readable storage medium of claim 1 , wherein the performing the load check includes:

comparing the metadata with a plurality of load values stored in a hardware thread register, and

signaling a trap upon a detected load metadata mismatch.

4. The non-transitory computer readable storage medium of claim 1 , wherein the performing the store check includes:

comparing the metadata with a store value stored in a hardware thread register,

and signaling a trap upon a detected store metadata mismatch.

5. The non-transitory computer readable storage medium of claim 1 , wherein the execution of the set of instructions further results in the performance of the following:

adding the load check and the store check to a cache controller of the first core; and

routing one of trap information and mismatch information to a local processing core upon detection of one of the load metadata mismatch and the store metadata mismatch.

6. The non-transitory computer readable storage medium of claim 1 , wherein the execution of the set of instructions further results in the performance of the following:

trapping a first processor that does not have a current permission to the runtime software routine as a result of one of the load metadata mismatch and the store metadata mismatch; and

distributing permission to each of the plurality of processors that require new permission.

7. The non-transitory computer readable storage medium of claim 1 , wherein the execution of the set of instructions further results in the performance of the following:

synchronizing a first processor with a remainder of the plurality of processors.

8. The non-transitory computer readable storage medium of claim 1 , wherein the metadata is inserted into a cache line of the cache and follows the cache line into a dynamic random-access memory (“DRAM”).

9. A system, comprising:

a plurality of cores within a multi-processor environment;

a cache memory within a first core of the cores and including metadata; and

a hardware check unit performing one of a load check and a store check for a thread to be executed by the first core on the metadata to detect a respective one of a load metadata mismatch and a store metadata mismatch, wherein the load metadata mismatch is a mismatch between metadata of a load to be executed by the first core and core metadata, and wherein the store metadata mismatch is a mismatch between metadata of a store to be executed by the first core and the core metadata, further wherein the core metadata identifies at least one of the cores with permission to execute the load and at least one of the cores with permission to execute the store, and invoking a runtime software routine to order memory references upon a detection of one of the load metadata mismatch and the store metadata mismatch.

10. The system of claim 9 , wherein the invoking the runtime software routine includes transferring control to a runtime library.

11. The system of claim 9 , wherein the performing the load check includes comparing the metadata with a plurality of load values stored in a hardware thread register and signaling a trap upon a detected load metadata mismatch.

12. The system of claim 9 , wherein the performing the store check includes comparing the metadata with a store value stored in a hardware thread register and signaling a trap upon a detected store metadata mismatch.

13. The system of claim 9 , wherein the hardware check unit is added to a cache controller of the first core and one of trap information and mismatch information is routed to a local processing core upon detection of one of the load metadata mismatch and the store metadata mismatch.

14. The system of claim 9 , wherein the runtime software routine traps a first processor that does not have a current permission to the runtime software routine as a result of one of the load metadata mismatch and the store metadata mismatch and distributes permission to each of the plurality of processors that require new permission.

15. The system of claim 9 , wherein the runtime software routine synchronizes a first processor with a remainder of the plurality of processors.

16. The system of claim 9 , wherein the metadata is inserted into a cache line of the cache and follows the cache line into a dynamic random-access memory (“DRAM”).

17. A method of updating an address table, comprising:

inserting metadata into a cache memory of a first core in a system including a plurality of cores;

performing one of a load check and a store check for a thread to be executed by the first core on the metadata to detect a respective one of a load metadata mismatch and a store metadata mismatch, wherein the load metadata mismatch is a mismatch between metadata of a load to be executed by the first core and core metadata, and wherein the store metadata mismatch is a mismatch between metadata of a store to be executed by the first core and the core metadata, further wherein the core metadata identifies at least one of the cores with permission to execute the load and at least one of the cores with permission to execute the store; and

invoking a runtime software routine to order at least one memory reference upon a detection of one of the load metadata mismatch and the store metadata mismatch.

18. The method of claim 17 , wherein the invoking the runtime software routine includes transferring control to a runtime library.

19. The method of claim 17 , further comprising:

adding the load check and the store check to a cache controller of the first core; and

routing one of trap information and mismatch information to a local processing core upon detection of one of the load metadata mismatch and the store metadata mismatch.

20. The method of claim 17 , further comprising:

trapping a first processor that does not have a current permission to the runtime software routine as a result of one of the load metadata mismatch and the store metadata mismatch;

synchronizing the first processor with a remainder of the plurality of processors; and

distributing permission to each of the plurality of processors that require new permission.

Assignments (3)
RELEASE OF SECURITY INTEREST Recorded Dec 28, 2022
From: GUGGENHEIM CORPORATE FUNDING, LLC
To: WIND RIVER SYSTEMS, INC.
Reel/Frame 062239/0590 →
PATENT SECURITY AGREEMENT Recorded Dec 24, 2018
From: WIND RIVER SYSTEMS, INC.
To: GUGGENHEIM CORPORATE FUNDING, LLC
Reel/Frame 049148/0001 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 8, 2017
From: ZEFFER, HAKAN
To: WIND RIVER SYSTEMS, INC.
Reel/Frame 041205/0883 →