Simultaneous Multithreading: Maximizing On-Chip Parallelism（1）

Time

2020.10.27

Summary

examines simultaneous multithreading(SM),a technique permitting several independent threads to issue instructions to a superscalar’s multiple functional units in a single cycle.

Research Objective

Problem Statement

Method(s)

We present several models of simultaneous multithreading and compare them with alternative organizations: a wide superscalar, a fine-grain multithreaded processor, and single-chip, multiple-issue multiprocessing architectures.

Evaluation

We evaluate several cache configurations made possible by this type of organization and evaluate tradeoffs between them.

We examine many of these complexities and evaluate alternative organizations in the design space.

Conclusion

both (single-threaded) superscalar and fine-grain multithreaded architectures are limited intheir ability to utilize the resources of a wide-issue processor. Simultaneous multithreading has the potential to achieve 4 times thethroughput of a superscalar, and double that of fine-grain multithreading.

Notes

Words

superscalar
超标量
tradeoffs
折中
outperform
做的比...好
high register file bandwidth,
高寄存器文件带宽
processor utilization
处理器利用率
long-latency
长时延
horizontal waste
横向浪费

Sentence

issue instructions
发出指令
the binding between thread and functional unit
线程和功能单元之间的绑定
issue up to four instructions per cycle from a single thread.
一个线程每个周期最多发出四个指令
employ multiple threads with fast context switch between threads.
使用多个线程，并在多个线程之间进行快速上下文切换。

TimtLine

Our results show the limits of superscalar execution and tradi-
tional multithreading to increase instruction throughput in future
processors.