Re: JavaMemoryModel: Question: Why is it that Memory Models are Intuitively Hard?

From: Ben Wint (benedic@attglobal.net)
Date: Wed Feb 19 2003 - 14:52:25 EST


Jeremy asked

>>> I would be very curious to hear what people's
    thoughts ... are <<<

so I'll take the liberty. I'm a programmer, not academic.
I've been worrying *deeply* about memory model issues
for over 10 years, because I've seen so many people not
worrying about them, and, as David Holmes says,

>>> People have been writing unsafe non-portable
     MT C code for decades without realizing it. <<<

Well, with luck, and given time, I hope their code will
become safe, when the hardware guys cave in and make
machines sequentially consistent.

I think that programmers should revolt against these relaxed
memory models.

I would have thought the hardware guys could use some of the
googolplexes of gates they are getting, to figure out the rare
cases when they need to slow down to maintain sequential
consistency, and when they can go at full speed.

>>> Programmers have a genuinely incorrect notion of how
   their code is executed (Sequential consistency). <<<

The notion is correct. The hardware is incorrect.

Ben Wint

-------------------------------
JavaMemoryModel mailing list - http://www.cs.umd.edu/~pugh/java/memoryModel



This archive was generated by hypermail 2b29 : Thu Oct 13 2005 - 07:00:42 EDT