Custom-Instruction Synthesis for Extensible-Processor Platforms

Abstract

Abstract-Efficiency and flexibility are critical, but often conflicting, design goals in embedded system design. The recent emergence of extensible processors promises a favorable tradeoff between efficiency and flexibility, while keeping design turnaround times short. Current extensible processor design flows automate several tedious tasks, but typically require designers to manually select the parts of the program that are to be implemented as custom instructions. In this work, we describe an automatic methodology to select custom instructions to augment an extensible processor, in order to maximize its efficiency for a given application program. We demonstrate that the number of custom instruction candidates grows rapidly with program size, leading to a large design space, and that the quality (speedup) of custom instructions varies significantly across this space, motivating the need for the proposed flow. Our methodology features cost functions to guide the custom instruction selection process, as well as static and dynamic pruning techniques to eliminate inferior parts of the design space from consideration. Furthermore, we employ a two-stage process, wherein a limited number of promising instruction candidates are first short-listed using efficient selection criteria, and then evaluated in more detail through cycle-accurate instruction set simulation and synthesis of the corresponding hardware, to identify the custom instruction combinations that result in the highest program speedup or maximize speedup under a given area constraint. We have evaluated the proposed techniques using a state-of-the-art extensible processor platform, in the context of a commercial design flow. Experiments with several benchmark programs indicate that custom processors synthesized using automatic custom instruction selection can result in large improvements in performance (up to 5.4 , an average of 3.4 ), energy (up to 4.5 , an average of 3.2 ), and energy-delay products (up to 24.2 , an average of 12.6 ), while speeding up the design process significantly.

Memory organization and data layout for instruction set extensions with architecturally visible storage

“…For both benchmarks the cheapest memory organization (8x8 clustered RF) had a feasible layout, which LICCA discovered; the runtime of our CAD flow was approximately one minute; this is significantly faster than the runtime required to find an 8-input 8-output ISE using most existing techniques [3,9,12]; this runtime is comparable to the runtime of the ISE identification method of Verma et al [13], which, to the best of our knowledge, is the fastest optimal algorithm published to date for this problem.…”

RISPP: A run-time adaptive reconfigurable embedded processor

“…Therefore, variations of graph partitioning algorithms may be investigated to transform monolithic SI graphs into modular SIs and to determine, which properties modular SIs demand. This can be used to modify the 'pruning' step in state-of-the-art automatic SI detection (see for instance [SRRJ04,VBI07]). To exploit the feature to share Atoms between different SIs, techniques like data-path merging [BKS04] may be adapted to identify reusable Atoms.…”

Architectural Exploration of Heterogeneous Multiprocessor Systems for JPEG

“…Such instructions are able to do the work of multiple instructions of a general-purpose processor. Extended instructions include fusion instructions, (21) SIMD/vector instructions and FLIX (22) instructions. Flexible Length Instruction Xtensions (FLIX) are VLIW-like instructions whereby multiple operations can be performed in a single instruction.…”