Philip Heidelberger and Kishor S. Trivedi. Analytic queueing models for programs with internal concurrency. IEEE Trans. on Computers, C-32:73--82, January 1983.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and are not skewed by artifacts of the simulation, or the combination of TCP and parallel download. In order to do so we use a very simple queueing model which is depicted in Figure 3. This type of queueing network is known in the literature as a Fork Join system and has been well studied ( 13] [7], 9] 10] 4] In this model a client arrives according to a Poisson process with rate # and forks 7 Servers Processor Sharing Service Rate = mu Join (Synchronization) Request Arrival Poisson, Rate = lambda Fork Fig. 3. Queueing model multiple requests that are sent to the N servers. ....

Philip Heidelberger and Kishor S. Trivedi. Analytic queueing models for programs with internal concurrency. IEEE Transactions on Computers, 32(1):73--82, 1983.

.... of the above techniques along with possible heuristics and approximations) A general approach for analytic performance prediction for shared memory systems has been proposed by Siewiorek et al. in [4] while probabilistic models for parallel programs based on queueing theory have been presented in [5]. An analytic performance prediction technique based on the approximation of parallel flow graphs by sequential flow graphs has been proposed by Qin et al. in [6] The above approaches require users to explicitly model the application along with the entire system. A source based analytic ....

Philip Heildelberger and Kishore S. Trivedi, "Analytic Queueing Models for Programs with Internal Concurrency ", IEEE Transactions on Computers, C-32(1):73--82, Jan. 1983.

....who compete for resources. This is different from massively parallel computing in which the parallel paths all execute the same instructions and there may be frequent global synchronizations. Several authors have developed analytic models for parallel service systems. Heidelberger and Trivedi [7] considered a simple computational loop executed by each user, alternating between a sequential section and a parallel section. They did consider contention between the parallel subtasks, but only one layer of resources. They compared two modelling approaches, called decomposition and ....

....in section 3. The CCD analysis is applied in Section 4. to several examples, showing that: ffl the approach scales up to systems with many servers having parallel sections and to many independent users, ffl approximation errors are limited to a few percent, ffl from the 228 examples studied in [7], the CCD errors are about three times as large as errors from decomposition (CD errors are about ten times as large) It is concluded that the delay approximation combines usable accuracy with scalable complexity. 2. Example as a Layered Model Figure 2 shows a simplified version of the system ....

[Article contains additional citation context not shown here]

Philip Heidelberger and Kishor S. Trivedi, "Analytic queueing models for programs with internal concurrency," IEEE Transactions on Computers, vol. 32, no. 1, pp. 73--82, Jan. 1983.

.... The influence of such program structures has also been studied [43] and predictive models of concurrency have been developed for the last two classes above for certain sets of applications; these models use stochastic techniques to derive averages and distributions for the potential concurrency [62, 93, 94, 104]. T sc The approaches to this component exhibit the most variation. In most of its uses it is intended to encompass the overheads associated with both synchronisation and communication. However, it is useful to view these two aspects separately. Synchronisation In the literature, synchronisation ....

....units is examined. A variation on this in which multi stage interconnection networks were analysed has been investigated by Harrison [58, 59] The other area that has been modelled is the effect of the varying concurrency available in shared memory parallel machines in terms of fork and join [9, 62, 93, 94]. The analysis associated with these models is performed under the assumption that the environment is work conserving, ie no processor is ever idle if there is an available task to be run. This allows for the treatment of the available work as one queue to one (or more) servers, any processor ....

[Article contains additional citation context not shown here]

Philip Heidelberger and Kishor Trivedi. Analytic queueing models for programs with internal concurrency. IEEE Transactions on Computers, C32(1):73--82, January 1983.

....models is that the performance information that can be obtained and the levels at which it can be obtained is limited. Finally, heterogeneity cannot be easily incorporated into analytic models. A number of general analytic models have been developed to model both hardware [9, 16] and software [10, 17, 18, 19]. Analytic models for different interconnections and networks as well as transport protocols have also been proposed [20, 21] Models for heterogeneous systems are proposed in [22, 23] but are limited to fork join type of parallelism and apply to fixed system configurations. An analytic ....

Philip Heildelberger and Kishore S. Trivedi, "Analytic Queueing Models for Programs with Internal Concurrency ", IEEE Transactions on Computers, vol. C-32, pp. 73--82, Jan. 1983.

No context found.

Philip Heidelberger and Kishor S. Trivedi. Analytic queueing models for programs with internal concurrency. IEEE Trans. on Computers, C-32:73--82, January 1983.

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