Hatcher, P. J. and Quinn, M. J., Data-Parallel Programming. The MIT Press, Cambridge, Mass., 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....program codes) and partial evaluation, both of which the current mode system can deal with naturally. So the rest of the paper will focus on the implementation level, leaving surface language design to future research. 4. 1 Matrix Multiplication We take matrix multiplication as the first example [3]. Consider the multiplication of an l Theta m matrix A and an m Theta n matrix B. The most natural data parallel modelling will be to create l Theta n processes each corresponding to an element of the result matrix C. Let P ij be the process for computing C ij . The creation of the P ij s can ....

....topic of future research. The result C can be built by concatenating one element subarrays computed by each process. Static analysis of the size of C will be easy even without array declarations and will simplify memory allocation. 4. 2 Prime Number Generation We take one more example from [3], the sieve of Eratosthenes. Process P k (k = 0; 1; holds a bit vector corresponding to the natural numbers in [nk; n(k 1) and strikes the multiples of primes in parallel. Since the bit vectors held by individual processes are independent, no difficulty arises. Instead of KL1 vectors ....

[Article contains additional citation context not shown here]

Hatcher, P. J. and Quinn, M. J., Data-Parallel Programming. The MIT Press, Cambridge, Mass., 1991.

....is supported, and there is no group concept. 9] compiles to a PRAM simulator while [16] offers back ends for several existing machines. ll [14] a similar approach, uses Pascal as base language. Further dataparallel languages in this tradition (see e.g. 17] are C , Dataparallel C [6], and dataparallel Fortran dialects such as HPF. The latter ones are mainly targeted towards distributed memory machines and require data layout directives to perform efficiently. Exact synchronicity is not supported as it is not available on the target architectures considered. Other ....

P.J. Hatcher and M.J. Quinn. Data-Parallel Programming. MIT Press, 1991.

....the target machine. The use of pointers in Fork95 is as flexible as in C, since the private address subspaces are embedded into the global shared memory of the PRAM. In particular, one does not have to distinguish between pointers to shared and pointers to private objects, in contrast to e.g. C [7]. Synchronous and asynchronous regions Fork95 provides synchronous and asynchronous programming modes, statically associated with program regions and functions. The farm statement farm stmt designates asynchronous mode and reinstalls synchronous mode at the end by bar rier synchronization. ....

....and direct shared memory access. Dataparallel variants of Modula, e.g. 15] support a subset of Fork95 s functionality. The main parallel constructs are synchronous and asynchronous parallel loops; there is no group concept. Other PRAM oriented dataparallel languages are Dataparallel C and C [7]. 3 PAD: A library of basic PRAM algorithms PAD is an attempt to provide support for the implementation of parallel algorithms as found in the current theoretical literature by providing access to some of the ubiquitous basic PRAM algorithms and computational paradigms like prefix sums, list ....

P. J. Hatcher and M. J. Quinn. Data-Parallel Programming. MIT Press, 1991.

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