EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index 2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: RE: R3.14.8 testing
From: "Jeff Hill" <[email protected]>
To: "'Jeff Hill'" <[email protected]>, "'Eric Norum'" <[email protected]>, "'Ernest L. Williams Jr.'" <[email protected]>
Cc: "'Marty Kraimer'" <[email protected]>, "'Andrew Johnson'" <[email protected]>, "'Janet Anderson'" <[email protected]>, <[email protected]>
Date: Wed, 2 Nov 2005 15:00:42 -0700

Ø      IMHO the important issue for EPICS developers is instead this: should we define an EPICS base default which requires

Ø     EPICS applications to compile with a non-standard sizeof(long) default. The downside is that if the users inadvertently

Ø     take the compiler default an incomprehensible crash will occur.

 

I should have said “if layered products inadvertently take the compiler default”

 

Jeff

 

-----Original Message-----
From: Jeff Hill [mailto:[email protected]]
Sent:
Wednesday, November 02, 2005 2:58 PM
To: 'Eric Norum'; 'Ernest L. Williams Jr.'
Cc: 'Marty Kraimer'; 'Andrew Johnson'; 'Janet Anderson'; [email protected]
Subject: RE: R3.14.8 testing

 

Ø     I don't see this as an argument for making 'long' a 64-bit type. It's an argument for 64-bit pointers, yes,

Ø     but I think that making 'sizeof(long)==8' is a separate issue.

 

There may be compiler switches that provide sizeof(long)==4, but I expect that the defacto standard compiler default will eventually be sizeof(long)==8.

 

I can think of some reasons why that default might become ubiquitous, but my opinion doesn’t matter. You can take that issue up with the compiler vendors.

 

IMHO the important issue for EPICS developers is instead this: should we define an EPICS base default which requires EPICS applications to compile with a non-standard sizeof(long) default. The downside is that if the users inadvertently take the compiler default an incomprehensible crash will occur.

 

Jeff

 

-----Original Message-----
From: Eric Norum [mailto:[email protected]]
Sent:
Wednesday, November 02, 2005 2:24 PM
To: Ernest L. Williams Jr.
Cc: Marty Kraimer; Andrew Johnson; Janet Anderson; Jeff Hill
Subject: Re: R3.14.8 testing

 

On Nov 2, 2005, at 2:59 PM, Ernest L. Williams Jr. wrote:

 

 

We are moving to the use of linux-x86_64 and want to run native 64-bit

apps.  

We want to take advantage of not only the AMD64 hardware but also the

larger address space which will be very useful for database applications

and systems which have a lot of RAM e.g. our ChannelArchiver. 

 

There is a very small performance penalty within the OS for translating

32-bit app to use the AMD64. Probably not much.

 

I don't see this as an argument for making 'long' a 64-bit type. It's an argument for 64-bit pointers, yes, but I think that making 'sizeof(long)==8' is a separate issue.

 

I seem to recall some earlier discussions where we had decided that EPICS would use 'long long' where 64-bit integers were needed.

 

-- 

Eric Norum <[email protected]>

Advanced Photon Source

Argonne National Laboratory

(630) 252-4793

 

 


References:
RE: R3.14.8 testing Jeff Hill

Navigate by Date:
Prev: RE: R3.14.8 testing Jeff Hill
Next: Re: R3.14.8 testing Marty Kraimer
Index: 2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: RE: R3.14.8 testing Jeff Hill
Next: RE: R3.14.8 testing Jeff Hill
Index: 2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 02 Feb 2012 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·