Hi Steve.
Shift registers often are loaded in parallel (and I think
read out in parallel). Why should the SV language exclude
this possibility just because of operation of a register
file as a FIFO?
For example, on putting to sleep a block containing a FIFO,
one might want to offload it as a RAM; then, on wakeup for
a low-power mode, one might want to load it back up
without having to execute the FIFO write process. If
the FIFO happened to be involved in the sleep/wakeup,
one might want to wake it up by reloading it in some
different data-state than the original.
I think this idea is trying to make the compiler smarter than
a designer; actually, a compiler never should be smarter
than a designer -- only more attentive to detail.
--
John Michael Williams
jwill@BasicISP.net
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Fri Jul 23 09:17:42 2010
This archive was generated by hypermail 2.1.8 : Fri Jul 23 2010 - 09:18:14 PDT