Stephen W. Keckler. Fast thread communication and synchronization mechanisms for a scalable single chip multiprocessor. Ph. D. thesis, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, June 1998. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Processor Mechanisms for Software Shared Memory - Carter (Correct)
....communication allows the MAP to be programmed in two distinct manners. The first approach is to treat each of the clusters as an independent processor, using the inter cluster communication mechanisms to reduce the communication and synchronization overhead. This approach has been shown [18] [16] to be very effective, achieving noticably higher speedups then memory based communication on a number of fine grained applications. Figure 3.4: Multithreading Scheme H Thread 0 H Thread 1 H Thread 2 H Thread 3 H Thread 4 H Thread 0 H Thread 1 H Thread 2 H Thread 3 H Thread 4 H Thread 0 H Thread ....
....in singlechip multiprocessors. These changes simplified the implementation of the MAP, as each cluster now acts as a 3 wide VLIW, issuing all three operations within an instruction in the same cycle instead of allowing individual operations to issue in different cycles. Steve Keckler s thesis [16] compared the three methods of inter cluster synchronization available on the MAP: synchronization through memory, synchronization using the CBAR instruction, and implicit synchronization using the tightly coupled execution mode. As expected, synchronization using the CBAR instruction was more ....
Stephen W. Keckler. Fast thread communication and synchronization mechanisms for a scalable single chip multiprocessor. Ph. D. thesis, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, June 1998.
Mechanisms for Efficient, Protected Messaging - Lee (Correct)
....and chip size, and the emergence of affordable integrated processor components [2, 3, 4, 5, 6, 7] Such systems would not be economically viable if limited to exotic, huge size problems. Few existing architectures are able to take advantage of fine grain parallelism, although a recent study [8] shows that it is readily available even in common benchmark programs. Figure 1 1 (reproduced from pp.113 [8] courtesy of Stephen Keckler) for example, shows that grain sizes as small as 70 cycles are accessible by parallelizing the applications at the inner loop level. Nonetheless, most existing ....
....systems would not be economically viable if limited to exotic, huge size problems. Few existing architectures are able to take advantage of fine grain parallelism, although a recent study [8] shows that it is readily available even in common benchmark programs. Figure 1 1 (reproduced from pp.113 [8], courtesy of Stephen Keckler) for example, shows that grain sizes as small as 70 cycles are accessible by parallelizing the applications at the inner loop level. Nonetheless, most existing machines, burdened by communication overhead that ranges from many tens to many thousands of processor ....
[Article contains additional citation context not shown here]
Steve Keckler, "Fast Thread Communication and Synchronization Mechanisms for a Scalable Single Chip Multiprocessor", PhD. Thesis, Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, 1998.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC