Talk:Asynchronous circuit

From Wikipedia, the free encyclopedia

Contents 1 Merge 2 Merge with Asynchronous Systems 3 Corrections by 172.188.108.141 4 Removals

[edit] Merge

I agree with User:DavidCary that merging Asynchronous circuit and Asynchronous logic makes sense. --Daedalus-Prime 17:53, 7 November 2005 (UTC)

[edit] Merge with Asynchronous Systems

I think the articles should remain seperate - the systems page has a lot of general information that does not fit in the circuit article. Kcordina 13:31, 10 March 2006 (UTC)

Since no-one has offered further opinion, I have removed the merge tags in line with my opinion. Kcordina ^Talk 15:26, 22 March 2006 (UTC)

I think they should be merged. There is no information in the systems article that would be inappropriate here. The stuff on design styles, modularity, handshaking protocols, async pipelines are all essential knowledge if you're researching asynchronous circuits. Two of the three references in the systems article actually have "asynchronous circuit" in the title. Chris, 8 May 2008. —The preceding unsigned comment was added by 172.188.108.141 (talk) 16:44, 8 May 2007 (UTC).

Other articles to merge: Delay_insensitive_circuit Delay-insensitive

[edit] Corrections by 172.188.108.141

I reverted some additions by 172.188.108.141 because they are not strictly true. In the power section the "huge increase in circuit size" is not necessarily present. In fact bundled-data circuits can have a negative overhead due to the removal of clock trees. The Amulet processors was not "cracked" by NDS in Israel. The processor in question is the "SPA" and it wasn't cracked but there was some "information leak" despite its use of balanced dual-rail. Although "there is no significant difference between dual-rail synchronous designs and dual rail asynchronous designs in terms of EMI security" the asynchronous versions are still impermeable to clock glitch attacks. Brejc8 20:06, 8 May 2007 (UTC)

Well at the moment the article says that asynchronous circuits lead to lower power consumption, which isn't strictly true. And while it may be true that one class of attacks has been removed, the use of an asynchronous design does not strictly prevent information leakage, or any other attack. New attacks may be introduced; the clock is a control signal, and when every data line is also a control signal, it may be possible to manipulate them to produce a similar failure. —The preceding unsigned comment was added by 172.188.108.141 (talk) 13:33, 9 May 2007 (UTC).

[edit] Removals

I think the quotes should be removed. Who cares what two guys once said? Neither of them appears to be particularly famous. The Martin rumour should be removed. "Red Star seems to be.." - seems? It either is, or it isn't. The link to UARTs should be removed; asynchronous UARTS are not asynchronous circuits (at least, none that you can buy). Link "asynchronous logic" should be "a.l @ U. of Manchester". Comments? —The preceding unsigned comment was added by 172.188.108.141 (talk) 13:43, 9 May 2007 (UTC).

I would like to see redirects from asynchronous_logic and self_timed_logic, though I am not going to add them. I don't see the UART reference, but I believe that there should be such a reference. It should explain that asynchronous logic/circuit/system is not related to asynchronous_communication and link to that page.

76.22.75.98 18:12, 11 September 2007 (UTC)