An Architecture & Mechanism for Supporting Speculative Execution of a Context-full Reconfigurable Function Unit
نویسندگان
چکیده
Recently researchers have shown interest in integrating Reconfigurable logic into conventional processors as a Reconfigurable Function Unit (RFU). A context-full RFU supports holding intermediate results inside itself, which eliminates some data movement overheads and has some other benefits. Most contemporary processors support out-of-order execution and speculation. When a context-full RFU is integrated into a speculative processor, if a speculative RFU instruction has modified the RFU context but cannot be committed in the end, the RFU context must be recovered. Traditional mechanisms to handle speculative execution of instructions cannot effectively address this issue. Because of the design complexity, previous proposals did not support context-full RFUs in speculative processors. In this paper, we propose an architecture & mechanism for supporting speculative execution of a context-full RFU in in-order issue, out-of-order execution processors. It does not require too much extra space for the RFU context storage and the performance penalty shown to be low in practice.
منابع مشابه
An EPIC Processor with Pending Functional Units
The Itanium processor, an implementation of an Explicitly Parallel Instruction Computing (EPIC) architecture, is an in-order processor that fetches, executes, and forwards results to functional units inorder. The architecture relies heavily on the compiler to expose Instruction Level Parallelism (ILP) to avoid stalls created by in-order processing. The goal of this paper is to examine, in small...
متن کاملPerformance Enhancement of an Adaptive Dynamic Extensible Processor by Using a Heterogeneous Reconfigurable Functional Unit
In this paper, we develop a heterogeneous architecture for the reconfigurable functional unit of an extensible processor. To verify the efficiency of our architecture, we applied it to 8 applications of Mibench. The new architecture improves execution time of custom instructions by 20% to 30% on average while supporting more custom instructions. The area and the total wire length are reduced by...
متن کاملA Chip-Multiprocessor Architecture with Speculative Multithreading
ÐMuch emphasis is now placed on chip-multiprocessor (CMP) architectures for exploiting thread-level parallelism in an application. In such architectures, speculation may be employed to execute applications that cannot be parallelized statically. In this paper, we present an efficient CMP architecture for speculative execution of sequential binaries without source recompilation. We present the s...
متن کاملCircuit Morphing: Declarative Modeling of Reconfigurable Combinational Logic
Using a simple reconfigurable logic gate that combines an ITE gate and a 1-bit memory cell we devise a mechanism for synthesizing fine grained circuits that overlap multiple logic functions. A declarative model of the approach, including an exact synthesizer for small circuits, is provided as a literate Haskell program (code available at http://logic.csci.unt.edu/tarau/research/2009/fsyn.hs). P...
متن کاملThe Effect of Executing Mispredicted Load Instructions in a Speculative Multithreaded Architecture
Concurrent multithreaded architectures exploit both instructionlevel and thread-level parallelism in application programs. A single-threaded sequencing mechanism needs speculative execution beyond conditional branches in order to exploit more instruction-level parallelism. In addition, an aggressive multithreaded architecture should also use thread-level control speculation in order to exploit ...
متن کامل