Subject: RE: [vhdl-200x] Re: Fork/Join
From: Peter Ashenden (peter@ashenden.com.au)
Date: Mon Jun 09 2003 - 09:57:12 PDT
I second Kevin's suggestion. A further reason against fork/join is that it
is hard to provide a modular abstraction of a process. By that, I mean that
it is hard to declare a parameterized process definition that can be
instantiated as an identifiable object with a separate thread of control.
VHDL already has processes that represent concurrently executing objects. A
cleaner approach would be to provide for parameterized process declarations
and a new form of concurrent statement that instantiates a declared process.
Dynamic creation of processes could be achieved with a sequential statement
that instantiates a declared process.
We looked at this issue in SUAVE some time ago. The ideas there might be
relevant.
Cheers,
PA
-- Dr. Peter J. Ashenden peter@ashenden.com.au Ashenden Designs Pty. Ltd. www.ashenden.com.au PO Box 640 Ph: +61 8 8339 7532 Stirling, SA 5152 Fax: +61 8 8339 2616 Australia Mobile: +61 414 70 9106> -----Original Message----- > From: owner-vhdl-200x@eda.org > [mailto:owner-vhdl-200x@eda.org] On Behalf Of vhdl-200x@grfx.com > Sent: Sunday, 8 June 2003 12:45 > To: vhdl-200x@eda.org > Subject: [vhdl-200x] Re: Fork/Join > > > > I noticed that adding fork/join (a la SystemVerilog) was on > someone's list of enhancements. I would highly recommend not > adding fork/join and instead just add a means of creating > single threads dynamically and add proper thread control > in the style of the Posix threads library - the fork/join > syntax doesn't lend itself to capturing thread identifiers > and makes fine-grained control difficult. > > Kev. >
This archive was generated by hypermail 2b28 : Mon Jun 09 2003 - 09:57:18 PDT