Fine-grain parallelism with minimal hardware support: a compiler-controlled threaded abstract machine

Abstract

In this paper, we present a relatively primitive execution model for ne-grain parallelism, in which all synchronization, scheduling, and storage management is explicit and under compiler control. This is dened by a threaded abstract machine (TAM) with a multilevel scheduling hierarchy. Considerable temporal locality of logically related threads is demonstrated, providing an avenue for eective register use under quasi-dynamic scheduling.A prototype TAM instruction set, TL0, has been developed, along with a translator to a variety of existing sequential and parallel machines. Compilation of Id, an extended functional language requiring ne-grain synchronization, under this model yields performance approaching that of conventional languages on current uniprocessors.Measurements suggest that the net cost of synchronization on conventional multiprocessors can be reduced to within a small factor of that on machines with elaborate hardware support, such as proposed dataow architectures. This brings into question whether tolerance to latency and inexpensive synchronization require specic hardware support or merely an appropriate compilation strategy and program representation.

Data-triggered threads: Eliminating redundant computation

“…To provide a smoother transitions from von Neumann to dataflow architectures, hybrid architectures [9,14,23], StarT [2,22], EARTH [13], DDM [11,16], and Fuce [1] attempt to build dataflow machines on top of conventional architectures. However, these systems still require significant changes to a baseline architecture to better support message passing, thread management, synchronization, context switching, and memory accesses.…”

Resource Management in Dataflow-Based Multithreaded Execution

“…Most multithreaded models that use code-block-based frames [7,8,10,31] group threads within a code-block, and bind code-blocks to processors. Although the grouping technique presented in this paper uses the same code-block-based approach, our model does not preclude other methods, such as grouping according to the amount of communication between threads.…”

“…Examples are Monsoon [6], *T [7], EM-4 [8], and the EARTH project [9]. There also exist software abstractions of fine-grain multithreading as exemplified by TAM [10] that can be implemented on traditional multiprocessors such as the CM-5.…”

CILK: A Multi-Threaded Programming System for Meta-Computers

“…Cilk provides an inlet feature for this purpose. Inlets are inspired in part by the inlet feature of TAM [11]. An inlet is essentially a C function internal to a Cilk procedure.…”