Abstract:With the availability of chip multiprocessor (CMP) and simultaneous multithreading (SMT) machines, extracting thread level parallelism from a sequential program has become crucial for improving performance. However, many sequential programs cannot be easily parallelized due to the presence of dependences. To solve this problem, different solutions have been proposed. Some of them make the optimistic assumption that such dependences rarely manifest themselves at runtime. However, when this assumption is violated, the recovery causes very large overhead. Other approaches incur large synchronization or computation overhead when resolving the dependences. Consequently, for a loop with frequently arising cross-iteration dependences, previous techniques are not able to speed up the execution. In this paper we propose a compiler technique which uses state separation and multiple value prediction to speculatively parallelize loops in sequential programs that contain frequently arising cross-iteration dependences. The key idea is to generate multiple versions of a loop iteration based on multiple predictions of values of variables involved in cross-iteration dependences (i.e., live-in variables). These speculative versions and the preceding loop iteration are executed in separate memory states simultaneously. After the execution, if one of these versions is correct (i.e., its predicted values are found to be correct), then we merge its state and the state of the preceding iteration because the dependence between the two iterations is correctly resolved. The memory states of other incorrect versions are completely discarded. Based on this idea, we further propose a runtime adaptive scheme that not only gives a good performance but also achieves better CPU utilization. We conducted experiments on 10 benchmark programs on a real machine. The results show that our technique can achieve 1.7x speedup on average across all used benchmarks.

Simultaneous multithreading trace processors: Improving trace processors performance

Simultaneous Multithreading Trace Processors

Path-based next N trace prefetch in trace processors

Boosting SMT trace processors performance with data cache misssensitive thread scheduling mechanism

Trace Cache and Trace Processor Technology

Simultaneous Multithreading Fault Tolerance Processor

TraceDo: an on-chip trace system for real-time debug and optimization in multiprocessor soc

A Flexible Chip Multiprocessor Simulator Dedicated for Thread Level Speculation

Position-aware Thread-Level Speculative Parallelization for Large-Scale Chip-Multiprocessor.

MKtrace: an Innovative Debugging Tool for Multi-Threaded Programs on Multiprocessor Systems

Reverse Compilation for Speculative Parallel Threading

Parallel simulation of chip-multiprocessor by using multi-threading

A high performance fault-tolerant approach based on simultaneous multithreading

Acceleration Techniques for Chip-Multiprocessor Simulator Debug

Software and Hardware Co-designed Multi-level TLBs for Chip Multiprocessors

Speculative Parallelization Using State Separation and Multiple Value Prediction.

Balancing Thread Partition for Efficiently Exploiting Speculative Thread-Level Parallelism

Design and research of two-level trace cache

Software-Based Lightweight Multithreading to Overlap Memory-Access Latencies of Commodity Processors

Performance Evaluation and Optimization of Cache Architecture for Simultaneous Multithreading Processor

Thread Level Parallel Processing Research for Network Processor