Re: [問題] watch真奧妙

作者: gun740201 (孤獨地擁抱了真實 )   2009-06-18 23:49:14
※ 引述《ric2k1 (Ric)》之銘言:
: 4. For a gate 'k' in the original "watching-v/known" lists of 'g' (i.e. 'g'
: is one of the watch pointers of 'k'), 'k' will try to update its watch
: pointer. Generally, this update process may result in:
: (i) The watch pointer is successfully updated. Let the new watch
: pointer be 'p' with watch value 'u' (u = 0, 1, or known), the 'k'
: will be added to the "watching-u" list of gate 'p'.
: (ii) The watch pointer cannot be updated due to implication conflict
: in 'k' (i.e. all the pins on the gate have watch values.
: e.g. all the fanins of an OR gate have value '0' and the output has
: value '1'). Then 'k' will be restored back to the "watching-v/known"
: lists of 'g' and the conflict analysis process will be triggered.
: (iii) The watch pointer cannot be updated and neither new implication
: nor conflict is generated on 'k' (i.e. the other watch pointer has
: non-watch value). Then no action will be taken except that 'k' will
: be restored back to the "watching-v/known" lists of 'g'.
: Hope this is clear.
我想我大概知道怎麼操作了
但是還是想知道哪種gate的哪種fan-in會在watching-known
而不是watching-0,1
另外,case(iii)是為了能detect未來的conflict嗎?
作者: timrau   2009-06-19 00:30:00
Inputs and outputs of XOR gates
作者: ric2k1 (Ric)   2009-06-19 00:38:00
case (iii) 像是一個 OR gate output has value '0',the watch pointer (input) was implied from x to '0',and other fanins have already had values '0'跟 detect 未來的 conflict 無關
作者: gun740201 (孤獨地擁抱了真實 )   2009-06-19 00:43:00
感謝回答,我想我應該懂了

Links booklink

Contact Us: admin [ a t ] ucptt.com