Resolution, optimization, and encoding of pointer variables for the behavioral synthesis from C

Abstract

As designers may model mixed hardware-software systems using a subset of or ++, we present SpC, a solution to synthesize and optimize hardware models with pointers. In hardware, a pointer is not only the address of data in memory, but it may also reference data mapped to registers, ports, or wires. Pointer analysis is used to find the set of locations each pointer may reference in a program at compile time. In this paper, we address the problem of synthesizing and optimizing pointers to multiple variables or array elements. The value of the pointers are encoded and branching statements are used to dynamically access data referenced by pointers. A heuristic is used to efficiently encode the values of the pointers. Compiler techniques are also used to reduce storage before loads and stores. An implementation using the SUIF framework (Wilson et al., 1994; SUIF Compiler Framework) is presented, followed by some case studies and experimental results.

Automated bug detection for pointers and memory accesses in High-Level Synthesis compilers

“…In the following, we assume byte alignment, but any other alignment can be used without loss of generality. This concept of memory location is similar to the location sets introduced by Wilson and Lam [16] and also used by Séméria and De Micheli [8]. In compilers, aliasanalysis works on pointers, trying to recover the set of memory locations where they can point to: the points-to set, or location set.…”

“…Notice that the points-to sets, like our memory locations, are abstract concepts, independent of the target architecture. The authors of [8] describe how to use them to enhance the address decoding circuits. They use tags and indexes as internal encoding to optimize pointer operations in HW, getting rid of tags or indexes when possible.…”

“…The first point is usually inferred using alias-analysis (or pointsto analysis [15]) and/or decided with explicit directives. The process is described in [8], which also outlines some optimizations to save resources and superfluous accesses. The second option is tightly bound to the HLS implementation, to the specific memory architecture of the generated design, and to the set of performed memory optimizations.…”

“…Things get more complicated when the starting language for HLS is C. Pointers, access-by-address, and dynamic allocation (with malloc/free) complicates memory synthesis. Despite these obstacles, the large majority of academic and commercial HLS compilers are based on C. In fact, a number of improvements have been made and are currently ongoing to overcome the troubles with the mentioned features of the language [8]. For this reason, a debugging technique for HLS of memory subsystems must be able to handle the quirks of C to be both effective and useful.…”

Inter-procedural resource sharing in High Level Synthesis through function proxies

“…The work in [20], [21] suggests an alternative approach for the synthesis of function pointers. The paper propose to synthesize calls through function pointers by generating control blocks assign values to function pointers to resolve the call.…”