Advances in Computer Systems Architecture: 8th Asia-Pacific by Tetsuya Sato, Hitoshi Murai (auth.), Amos Omondi, Stanislav

By Tetsuya Sato, Hitoshi Murai (auth.), Amos Omondi, Stanislav Sedukhin (eds.)

This convention marked the ?rst time that the Asia-Paci?c computers structure convention was once held outdoors Australasia (i. e. Australia and New Zealand), and was once, we are hoping, the beginning of what is going to be a typical occasion. The convention began in 1992 as a workshop for laptop architects in Australia and for this reason constructed right into a full-?edged convention masking Austra- sia. extra significant alterations ended in the current convention. The ?rst was once a metamorphosis from “computer structure” to “computer platforms architecture”, a transformation that well-known the significance and shut dating to machine arc- tecture of yes degrees of software program (e. g. working platforms and compilers) and of alternative components (e. g. machine networks). the second one swap, which re?ected the expanding variety of papers being submitted from Asia, used to be the substitute of “Australasia” with “Asia-Paci?c”. This year’s occasion used to be hence rather signi?cant, in that it marked the start of a really “Asia-Paci?c” convention. it's meant that during the longer term the convention venue will exchange among Asia and Australia/New Zealand and, even if nonetheless small, we are hoping that during time the convention will become a tremendous person who represents Asia to an analogous - tent as present significant computer-architecture meetings in North the US and Europe characterize these regions.

Show description

Read Online or Download Advances in Computer Systems Architecture: 8th Asia-Pacific Conference, ACSAC 2003, Aizu-Wakamatsu, Japan, September 23-26, 2003. Proceedings PDF

Best nonfiction_11 books

Additional resources for Advances in Computer Systems Architecture: 8th Asia-Pacific Conference, ACSAC 2003, Aizu-Wakamatsu, Japan, September 23-26, 2003. Proceedings

Sample text

Multi-threading on the other hand makes any form of speculation unnecessary, although some multi-threaded approaches do rely on speculation [9]. Multi-threaded instruction execution need not suffer from the problems encountered using speculative execution, with one exception and that is fundamental, it is synchronising across many concurrently issued instructions and requires a large register file. In a threaded microprocessor, it is not necessary to issue instructions in a thread out-of-order and hence we need only deal with true data dependencies.

The justification for using registers as synchronisers for micro threads is to provide a very low-latency synchronisation mechanism within a single context and this model requires that all registers in the micro-architecture implement a modified i-structure [26]. g. the number of cycles to the register read stage in the pipeline + 1, assuming an I-cache hit on rescheduling). Thread suspension occurs at the register read stage when a read is attempted on an empty register. In this case the instruction reading the register is transformed into an instruction that writes the thread reference into the empty register.

The number of cycles to the register read stage in the pipeline + 1, assuming an I-cache hit on rescheduling). Thread suspension occurs at the register read stage when a read is attempted on an empty register. In this case the instruction reading the register is transformed into an instruction that writes the thread reference into the empty register. To do this, the thread’s reference travels down the pipeline with each instruction executed. A subsequent write to that register will extract and reschedule the thread whose reference is waiting there.

Download PDF sample

Rated 4.85 of 5 – based on 39 votes