rgn_create
(which is a simple wrapper around
malloc) is also available to obtain sequential timings on Alewife, the
CM-5, or uniprocessor systems (e.g., desktop workstations).
To use the TCP/UNIX implementation of CRL, PVM [5] must also be installed on each machine; CRL leverages off of the group and process management facilities provided by PVM. The TCP/UNIX implementation included in CRL 1.0 is not intended to be a high-performance distributed shared memory implementation delivering speedups similar to those obtained with the Alewife and the CM-5 implementations. It is provided to allow development and experimentation by those without access to an Alewife or CM-5 platform and to display the ease with which CRL may be ported.
The rest of this document proceeds as follows. Section 2 provides an introduction to the CRL distributed shared memory system and its goals. Section 3 describes the CRL programming model and CRL functions. Section 4 mentions some of the caveats to keep in mind while using CRL 1.0. Sections 5, 6, and 7 provide instructions for using the CM-5, TCP/UNIX, and null versions of CRL 1.0, respectively. Section 8 gives instructions for obtaining the most recent CRL distribution package. Section 9 contains the copyright notice for CRL. Finally, Appendix A contains a detailed example of a simple application written using CRL 1.0.