View previous topic :: View next topic |
Author |
Message |
stfark1
Joined: 02 Sep 2008 Posts: 210
|
Posted: Wed May 26, 2010 7:50 pm Post subject: Windows 7 64 bit re-visit |
|
|
Hi Paul: Is there any reason to believe that the problems that I experienced in the past with Win 7 64bit have been solved? Please let me know. Thanks, Sid kraft |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 7924 Location: Salford, UK
|
Posted: Wed May 26, 2010 9:19 pm Post subject: |
|
|
As far as I know, no one has managed to reproduce your problem.
For myself, I am not yet geared up to testing under a 64bit OS. |
|
Back to top |
|
|
mecej4
Joined: 31 Oct 2006 Posts: 1886
|
Posted: Fri Jun 18, 2010 8:20 pm Post subject: FTN95 on Windows 7-X64 |
|
|
I have no trouble compiling and running Fortran console applications from the command line or from Plato.
Code: | s:\>ftn95 -nocolour -V
Silverfrost FTN95/.NET Copyright (C) 1993-2010 Silverfrost Ltd
---------------------------------------------------------------
Version: 5.50.0
Built: Sat Feb 27 20:28:53 2010
Operating System: Microsoft .NET on Windows 7 Home
CPU: Pentium(R) Dual-Core CPU T4300 @ 2.10GHz Model 7 Step 10
CPU Features: MMX, SSE, SSE2, x86-64
Status: Personal Edition |
|
|
Back to top |
|
|
stfark1
Joined: 02 Sep 2008 Posts: 210
|
Posted: Sun Aug 29, 2010 6:12 pm Post subject: Win 7 64 bit |
|
|
Andrew/Paul Laidler/ other:
I currently am using the SilverFrost Fortran/Plato3 with Win 7 32 bit. It runs fine under 32 bit but I had problems running it with the 64 bit version. I would like to use the 64 bit version to get support for more memory. Has there been any new updates to the SilverFrost Fortran/Plato3 system to enable it to run under Win 7, 64 bit? Please let me know. Thanks in advance, Sid Kraft |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 7924 Location: Salford, UK
|
Posted: Mon Aug 30, 2010 7:54 am Post subject: |
|
|
The current release of Plato is version 4.3.
If you can describe your problem and someone else can confirm it then the best I can do at the moment is to log it as something to be fixed.
Before too long I plan to have a testing environment for Windos 7 32bit and 64bit versions.
As you will appreciate, it is extremely difficult to fix bugs that I am unable to reproduce on my machine. |
|
Back to top |
|
|
aebolzan
Joined: 06 Jul 2007 Posts: 229 Location: La Plata, Argentina
|
Posted: Mon Aug 30, 2010 4:26 pm Post subject: |
|
|
No problems here with Win7 64 bit......Plato 4.3.0.
Agustin |
|
Back to top |
|
|
mecej4
Joined: 31 Oct 2006 Posts: 1886
|
Posted: Thu Sep 23, 2010 5:40 pm Post subject: Unstated problems while sunning FTN95 on Win7-64 |
|
|
Quote: | I would like to use the 64 bit version to get support for more memory | You are still using a 32-bit compiler, and the .EXE it produces is also a 32-bit program. Therefore, the .EXE is subject to the usual limitations of 32-bit programs. Your program cannot even tell that it is running on a 64-bit system, without using some rather special tricks (these would not be implemented in Fortran!). Any "support for more memory" would have to come from the 64-bit OS, not from the 32-bit compiler or the 32-bit EXE files that it produces.
The only advantage of running Salford FTN95 on a 64-bit Windows system is that your EXE does not have to compete for its share of 32-bit address space with those Widows processes that are 64-bit. Thus, a bigger fraction of the 4 GB address space may be available to your program than on 32-bit Windows, provided you have more than 4 GB of installed RAM. |
|
Back to top |
|
|
JohnCampbell
Joined: 16 Feb 2006 Posts: 2554 Location: Sydney
|
Posted: Fri Sep 24, 2010 1:20 am Post subject: |
|
|
I am now using a 64-bit OS. My experience to date is:
There is no way of getting more addressable memory with FTN95 than the past 1.7gb practical limit.
I've not found how to implement the /3gb switch either, although this was of limited practical value.
The big change is disk I/O is now much faster as the disk buffering for large files appears to be much better managed, possibly due to more physical memory and also possibly because the 64-bit OS is better suited to dealing with bigger files.
This means that my "2gb" limited programs which used disk storage for overflow ie "out of core solutions" work much faster when more than 2gb is required for the equation solution.
Also, make sure you understand the influence of:
-addressable memory, (limited to about 1.7gb with FTN95)
-available physical memory (limited to 4gb with 32 OS) and
-pageing memory limit (which can be much bigger, I don't know the limit).
They all interact to effect the run time. The 64 OS can increase all but the first to improve performance.
I've also implemented better reporting of processor and disk times (like task manager) and although at times it appears that little is happening apart from disk delays, the total disk time is not as bad these days.
John |
|
Back to top |
|
|
JohnCampbell
Joined: 16 Feb 2006 Posts: 2554 Location: Sydney
|
Posted: Fri Oct 01, 2010 10:02 am Post subject: |
|
|
How much a week can change things !!
Quote: | There is no way of getting more addressable memory with FTN95 than the past 1.7gb practical limit. |
See my posts this week in Support, as it appears that with Win-64 and WOW64, you can address nearly 4gb of memory, but only by using ALLOCATE. There also appears to be a limit of 2gb for a single array. Hopefully that may change too.
Addressable memory appears to have doubled !
My experience of XP-64 has been very promising.
John |
|
Back to top |
|
|
|