초록 ▼

A method and system for detecting race conditions computing systems. A parallel computing system includes multiple processor cores is coupled to memory. An application with a code sequence in which parallelism to be exploited is executed on this system. Different processor cores may operate on a giv

A method and system for detecting race conditions computing systems. A parallel computing system includes multiple processor cores is coupled to memory. An application with a code sequence in which parallelism to be exploited is executed on this system. Different processor cores may operate on a given memory line concurrently. Extra bits are associated with the memory data line and are used to indicate changes to corresponding subsections of data in the memory line. A memory controller may perform a comparison between check bits of a memory line to determine if more than one processor core modified the same section of data in a cache line and a race condition has occurred.

대표청구항 ▼

What is claimed is: 1. A method for use in a computing system, the method comprising: initiating execution of a code sequence; a first processing unit utilizing a first copy of a given memory line corresponding to the code sequence during said execution, and a second processing unit utilizing a sec

What is claimed is: 1. A method for use in a computing system, the method comprising: initiating execution of a code sequence; a first processing unit utilizing a first copy of a given memory line corresponding to the code sequence during said execution, and a second processing unit utilizing a second copy of the given memory line during said execution, said given memory line comprising a plurality of subsections and a check bit for each subsection; clearing first check bits prior to utilization of the first copy by the first processing unit, and clearing second check bits prior to utilization of the second copy by the second processing unit; the first processing unit setting a first check bit of the first check bits, in response to modifying a subsection of said first copy that corresponds to the first check bit, and the second processing unit setting a second check bit of the second check bits, in response to modifying a subsection of said second copy that corresponds to the second check bit; comparing first check bits of the first copy to second check bits of the second copy; detecting a race condition, in response to determining a same check bit is set in both of the first check bits and the second check bits; and generating an interrupt in response to detecting said race condition. 2. The method as recited in claim 1, further comprising storing address information corresponding to the memory line in response to detecting said race condition. 3. The method as recited in claim 1, wherein each of said subsections comprise a fixed number of bytes. 4. The method as recited in claim 3, further comprising: receiving said first copy from the first processing unit subsequent to completing said execution; receiving said second copy from the second processing unit subsequent to completing said execution; and storing data corresponding to both said first copy and said second copy in a main memory, if a race condition is not detected. 5. A computing system comprising: a first processing unit, including at least one processor, configured to utilize a first copy of a given memory line corresponding to a code sequence during execution; a second processing unit, including at least one processor, configured to utilize a second copy of the given memory line during said execution, said given memory line comprising a plurality of subsections and a check bit for each subsection; and the first processing unit setting a first check bit of first check bits, in response to modifying a subsection of said first copy that corresponds to the first check bit; the second processing unit setting a second check bit of second check bits, in response to modifying a subsection of said second copy that corresponds to the second check bit; a memory controller configured to: clear said first check bits prior to utilization of the first copy by the first processing unit, and clear said second check bits prior to utilization of the second copy by the second processing unit; compare first check bits of the first copy to second check bits of the second copy; detect a race condition, in response to determining a same check bit is set in both of the first check bits and the second check bits; and generate an interrupt in response to detecting said race condition. 6. The system as recited in claim 5, wherein the memory controller is further configured to store address information corresponding to the memory line in response to detecting said race condition. 7. The system as recited in claim 5, wherein each of said subsections comprise a fixed number of bytes. 8. The system as recited in claim 7, wherein the memory controller is further configured to: receive said first copy from the first processing unit subsequent to completing said execution; receive said second copy from the second processing unit subsequent to completing said execution; and store data corresponding to both said first copy and said second copy in a main memory, if a race condition is not detected. 9. A memory system comprising: a main memory comprising a plurality of memory lines; and a memory controller coupled to said main memory via a bus, wherein said memory controller is configured to: receive a first copy of a memory line; receive a second copy of the memory line; compare first check bits of the first copy to second check bits of the second copy; and detect a race condition, in response to determining a same check bit is set in both of the first check bits and the second check bits; wherein each of said first check bits corresponds to a particular subsection of the first copy, and each of said second check bits corresponds to a particular subsection of the second copy; and wherein the memory controller is further configured to generate an interrupt in response to detecting said race condition. 10. The memory system as recited in claim 9, wherein the memory controller is configured to perform said compare subsequent to completion of a code sequence execution. 11. The memory system as recited in claim 10, wherein the memory controller is configured to clear the first check bits and the second check bits prior to said execution. 12. The memory system as recited in claim 11, wherein a given bit of the first check bits is set responsive to modification of a corresponding subsection of the first copy, and a given bit of the second check bits is set responsive to modification of a corresponding subsection of the second copy.