2004•Unpublished venueRequires access

Low power instruction fetch using profiled variable length instructions

Mikael Collin, Mats Brorsson

Open publisher page 8 citations

Abstract

Computer system performance depends on high access rate and low miss rate in the instruction cache, which also affects energy consumed by fetching instructions. Simulation of a small computer typical for embedded systems shows that up to 20% of the overall processor energy is consumed in the instruction fetch path and as much as 23% of the execution time is spent on instruction fetch. Therefore it is of key importance to reduce the energy dissipated during instruction fetch. One way to increase the instruction memory bandwidth is to fetch more instructions in each access without increasing the bus width. We propose an extension to a RISC ISA, with variable length instructions, yielding higher information density without compromising programmability. Based on profiling of dynamic instruction usage and argument locality of a set of SPEC CPU2000 applications, we present a scheme using 816- and 24-bit instructions accompanied by lookup tables inside the processor Our scheme yields a 20-30% reduction in main memory usage, and experiments show that up to 60% of all executed instructions consist of short instructions. The overall energy savings are up to 15% for the entire data path and memory system, and up to 20% in the instruction fetch path.

About this research paper

What this paper is about

Computer system performance depends on high access rate and low miss rate in the instruction cache, which also affects energy consumed by fetching instructions. Simulation of a small computer typical for embedded systems shows that up to 20% of the overall processor energy is consumed in the instruction fetch path and as much as 23% of the execution time is spent on instruction fetch. Therefore it is of key importance to reduce the energy dissipated during instruction fetch. One way to increase the instruction memory bandwidth is to fetch more instructions in each access without increasing the bus width. We propose an extension to a RISC ISA, with variable length instructions, yielding higher information density without compromising programmability. Based on profiling of dynamic instruction usage and argument locality of a set of SPEC CPU2000 applications, we present a scheme using 816- and 24-bit instructions accompanied by lookup tables inside the processor Our scheme yields a 20-30% reduction in main memory usage, and experiments show that up to 60% of all executed instructions consist of short instructions. The overall energy savings are up to 15% for the entire data path and memory system, and up to 20% in the instruction fetch path.

Why it matters

OpenAlex reports 8 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Computer system performance depends on high access rate and low miss rate in the instruction cache, which also affects energy consumed by fetching instructions. Simulation of a small computer typical for embedded systems shows that up to 20% of the overall processor energy is consumed in the instruction fetch path and as much as 23% of the execution time is spent on instruction fetch. Therefore it is of key importance to reduce the energy dissipated during instruction fetch. One way to increase the instruction memory bandwidth is to fetch more instructions in each access without increasing the bus width. We propose an extension to a RISC ISA, with variable length instructions, yielding higher information density without compromising programmability. Based on profiling of dynamic instruction usage and argument locality of a set of SPEC CPU2000 applications, we present a scheme using 816- and 24-bit instructions accompanied by lookup tables inside the processor Our scheme yields a 20-30% reduction in main memory usage, and experiments show that up to 60% of all executed instructions consist of short instructions. The overall energy savings are up to 15% for the entire data path and memory system, and up to 20% in the instruction fetch path.

Key concepts: Fetch, Computer science, Instructions per cycle, Instruction set, Spec#, Parallel computing, Microarchitecture, Cache

Related papers

Back to paper searchBrowse research topicsOriginal source
Low power instruction fetch using profiled variable length instructions — Research Paper | ScholarLens