R. Barve and J. S. Vitter. External memory algorithms with dynamically changing memory allocation. Technical Report CS199809, Duke University, June 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....algorithms are already known. We believe that the ability of our simulations to accommodate a variable number k, for 1 k v p , of virtual processors in the memory of a real processor allows our techniques to conform to a new memory adaptive EM model, recently proposed by Barve and Vitter [14]. In this model, the available memory changes periodically and unpredictable, perhaps due to the demands of unrelated applications. More investigation of this idea, as it applies to our techniques, seems warranted. Our new EM BSP algorithms have the advantage that they are essentially quite ....

R. Barve and J. S. Vitter. External memory algorithms with dynamically changing memory allocation. Technical Report CS199809, Duke University, June 1998.

....the initial results on halfspace range searching. 6 7 Miscellaneous There is some work on contour line extraction [AAM 98] and isosurface extraction [CS97] motivated by GIS applications. Vitter and Barve considered external memory algorithms with dynamically changing memory allocations [BV98] Arge and Miltersen [AM98] give a theoretical summary of lower bounds in external memory computational geometry. ....

R. Barve and J. Vitte. External memory algorithms with dynamically changing memory allocations. Technical Report CS-1998-09, Duke University, 1998.

No context found.

R. D. Barve and J. S. Vitter. External memory algorithms with dynamically changing memory allocations: Long version. Technical Report CS#1998#09, Duke University, 1998.

....of systems in which MA algorithms have been implemented [10, 11, 19] Allowing 2m I O operations in an allocation phase of size m is natural because the act of loading m blocks into internal memory and then (after internal processing) storing them back to disk requires 2m I O operations. In [3] we present a somewhat more detailed dynamic memory model where the duration of each allocation phase can be arbitrarily long; the model in Definition 2 is a constant factor approximation of the general model in [3] In our model, the allocation size can be as low as a constant number of ....

....internal processing) storing them back to disk requires 2m I O operations. In [3] we present a somewhat more detailed dynamic memory model where the duration of each allocation phase can be arbitrarily long; the model in Definition 2 is a constant factor approximation of the general model in [3]. In our model, the allocation size can be as low as a constant number of blocks; in particular, we do not make the assumption used in the work on hash join [11, 18] that the internal memory allocation always remains larger than the square root of the smaller input file s size. As a consequence ....

[Article contains additional citation context not shown here]

R. D. Barve and J. S. Vitter. External memory algorithms with dynamically changing memory allocations: Long version. Technical Report CS--1998--09, Duke University, 1998.

....adapt dynamically to whatever resources are available so as to preserve good performance [107] The algorithms in the previous sections assume a fixed memory allocation; they must resort to virtual memory if the memory allocation is reduced, often causing a severe performance hit. Barve and Vitter [28] discuss the design and analysis of EM algorithms that adapt gracefully to changing memory allocations. In their model, without loss of generality, a program P is allocated memory in phases: During the ith phase, P is allocated m i blocks of internal memory, and this memory remains allocated to P ....

....sorting if X i 2m i log m i = O(n log n) for all possible sequences m 1 , m 2 , of memory allocation. Intuitively, if P is dynamically optimal, no other program can perform more than a constant number of sorts in the worst case for the same sequence of memory allocations. Barve and Vitter [28] define the model in generality and give dynamically optimal strategies for sorting, matrix multiplication, and buffer trees operations. Their work represents the first theoretical model of dynamic allocation for EM algorithms. Pang et al. 107] and Zhang and Larson [147] give memory adaptive ....

R. D. Barve and J. S. Vitter. External memory algorithms with dynamically changing memory allocations: Long version. Technical Report CS--1998--09, Duke University, 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