C.-K. Luk and T. C. Mowry, "Automatic compiler-inserted prefetching for pointer-based applications." IEEE Transactions on Computers, vol. 48, no. 2, pp. 134--141, 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and Adve examine the effectiveness of software prefetching for scientific programs on a shared memory multiprocessor built with modern ILP processors [87] 2.6. 4 Prefetching Linked Structures: Luk and Mowry Luk and Mowry develop three prefetching schemes for recursive data structures (RDS) [68, 70, 71]. These include greedy prefetching, history pointer prefetching, and datalinearization. In the initial work, Luk and Mowry have a compiler implementation for greedy prefetching only. But, in his dissertation, Luk implements and evaluates all three techniques [68] They use the Olden benchmarks to ....

Chi-Keung Luk and Todd C. Mowry. Automatic compiler-inserted prefetching for pointer-based applications. IEEE Transactions on Computers, 48(2):134--141, February 1999.

....in advance. Once available in a local cache, those objects can be retrieved with minimal 23 delay, enhancing the user experience. Prefetching is a well known approach to decrease access times in the memory hierarchy of modern computer architectures (e.g. Smi82, Mow94, KCK 96, VL97, KW98a, LM99, JG99, SSS99] and has been proposed by many as a mechanism for the same in the World Wide Web (e.g. PM96, KLM97] While the idea is promising, prefetching is not straightforward to evaluate (as we point out later in this thesis and then propose an alternative approach in Chapter 10) and has ....

Chi-Keung Luk and Todd C. Mowry. Automatic compiler-inserted prefetching for pointer-based applications. IEEE Transactions on Computers, 48(2), 1999. Special Issue on Cache Memory and Related Problems.

No context found.

C.-K. Luk and T. C. Mowry. Automatic Compiler-Inserted Prefetching for Pointer-Based Applications. IEEE Transactions on Computers, 48(2):134--141, Feb. 1999.

....approach as done in some other studies [12, 31] we propose extending the underlying instruction set architecture (ISA) by adding a few instructions which allow software to control preexecution directly. It is interesting to note that this approach is analogous to software controlled prefetching [21, 22]. The advantages of this approach are its programmability and flexibility. In addition, we expect the software overhead to be low given today s wide issue superscalar processors, which is generally quite true for software controlled prefetching. Figure 4 shows our proposed ISA extensions, which ....

....involving indirect array references (equake) and arrays of pointers (raytrace) Mowry s extended algorithm [22] for prefetching these references was used. For the remaining five applications, we experimented with the greedy prefetching and jump pointer prefetching proposed by Luk and Mowry [20, 21] as well as the extensions of jump pointer prefetching (full jumping, chain jumping, and root jumping) proposed by Roth and Sohi [27] In all cases, a wide range of prefetching distances (whenever applicable) were tested, and the best performing one was chosen for the comparison. Figure 6(a) ....

C.-K. Luk and T. C. Mowry. Automatic compiler-inserted prefetching for pointer-based applications. IEEE Transactions on Computers (Special Issue on Cache Memory), 48(2):134--141, February 1999.

....miss latencies to the extent permitted by the memory system, as illustrated in Figure 3(c) In order to do that, we must first overcome the pointer chasing problem. 3. 1 Solving the Pointer Chasing Problem Figure 4(a) illustrates the pointer chasing problem, which was observed by Luk and Mowry [13, 14] in the context of prefetching pointer linked data structures (i.e. linked lists, trees, etc. in general purpose applications. Assuming that three nodes worth of computation are needed to hide the miss latency, then when node n i in Figure 4(a) is visited, we would like to be launching a prefetch ....

....distance Jump Pointer Array (c) Linked list with a jump pointer array Figure 4: Addressing the pointer chasing problem. thereby defeating our goal of hiding the miss latency of n i 3 . Luk and Mowry proposed two solutions to the pointerchasing problem that are applicable to linked lists [13, 14]. The first scheme (data linearization prefetching) involves arranging the nodes in memory such that their addresses can be trivially calculated without dereferencing any pointers. For example, if the leaf nodes of the B Tree are arranged sequentially in contiguous memory, they would be ....

[Article contains additional citation context not shown here]

C.-K. Luk and T. C. Mowry. Automatic Compiler-Inserted Prefetching for Pointer-Based Applications. IEEE Transactions on Computers, 48(2):134--141, Feb. 1999.

No context found.

C.-K. Luk and T. C. Mowry, "Automatic compiler-inserted prefetching for pointer-based applications." IEEE Transactions on Computers, vol. 48, no. 2, pp. 134--141, 1999.

No context found.

C.-K. Luk and T. C. Mowry. Automatic compiler-inserted prefetching for pointer applications. IEEE Transactions on Computers, 48(2):134--141, February 1999. 2.1.3, 5.5

No context found.

Chi-Keung Luk and Todd C. Mowry. Automatic compiler-inserted prefetching for pointer-based applications. IEEE Transactions on Computers, 48:2:134-- 141, 1999.

No context found.

C.-K. Luk and T. C. Mowry. Automatic compiler-inserted prefetching for pointer applications. IEEE Transactions on Computers, 48(2):134--141, February 1999. 2.1.3, 5.5

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