Hi,
I recently went through the toolflow setup with
RHEL5/Matlab_2008b/Xilinx 11.5/mlib10_1 from svn as advised by this list
in late May.
( http://www.mail-archive.com/[email protected]/msg01530.html )
The setup went smoothly and I was able to simulate/run tut1 and iADC
tutorial so far. I ran into couple of glitches:
1. It took me a while to realize that I need Matlab "Signal Processing
Blockset" (different than toolbox) license. This was not part of our
presumably fully licensed Matlab. So it may worth adding the list of
Matlab toolboxes to the "toolflow setup" wiki page.
2. I found setting up Python2.6 under RHEL5 quite frustrating. (RHEL5
only supports Python 2.4). I had to rebuild numpy and scipy from source
which required installing many packages (python-devel, gtk++-devel,
pygobject, pygtk, etc.) and I ran into incompatibility issues when
getting those packages from RHEL5. In this regard, I would have much
preferred to deal with Ubuntu than RHEL5.
Regards,
Mandana Amiri
Jonathan Landon wrote:
In March there was some discussion about setting up the CASPER tools
on Windows vs. Linux, and at the time the answer seemed to be that
Linux is the way of the future but it hadn't been well-tested, so a
new user should set up the 10.1 tools on Windows. Since then, the
11.x tools have had plenty of testing, so the official CASPER wisdom
is for users starting from scratch to use the Xilinx 11.4 tools on
RHEL5 with Matlab 2009b and mlib_devel_10_1 from the git repository --
is that right?
The related CASPER wiki pages were last updated 12 February 2010
(http://casper.berkeley.edu/wiki/Linux_xps),
(http://casper.berkeley.edu/wiki/Xilinx_ISE_11.4_Setup) and there was
additional discussion March 2010 on this list (included below), so I
just want to make sure I have the latest info in case anything changed
since then.
Thanks,
Jonathan Landon
[casper] Simulink/Xilinx integration
Andrew Martens
Tue, 16 Mar 2010 06:43:51 -0700
---------- Forwarded message ----------
From: Andrew Martens <[email protected]>
Date: 16 March 2010 14:51
Subject: Re: [casper] Simulink/Xilinx integration
To: Steve Maher <[email protected]>
Hi Steve
As Jason has mentioned, going with Linux is choosing the path less
traveled within the CASPER group. We are moving that way for various
reasons but are not there yet. If you are looking for an easy route
into the tools, Windows-XP combined with Matlab2007b and the Xilinx
10.x tools are the way to go.
(http://casper.berkeley.edu/wiki/MSSGE_Toolflow_Setup)
If you need to go the Linux route, tread lightly. Firstly ensure that
you can get as close to the official Xilinx supported configuration as
possible (http://www.xilinx.com/ise/ossupport/index.htm). Red Hat and
SUSE enterprise Linux are officially supported. Centos is a freely
available Linux version
based on Red Hat source. Note that an email was sent out a few weeks
ago with a configuration that CASPER people at UC Berkeley have tried
and recommend
(http://www.mail-archive.com/[email protected]/msg01228.html).
http://casper.berkeley.edu/wiki/Xilinx_ISE_11.4_Setup gives a setup
using Centos 5. I have set up (independently and using slightly
different steps) using Centos 5 and have successfully compiled a small
design.
Please note that at the moment Xilinx does not officially support
System Generator on any 64-bit version of Windows. For large devices
(e.g to be on ROACH2) a 64 bit operating system will be crucial
(http://www.xilinx.com/ise/products/memory.htm).
Regards
Andrew
On 16 March 2010 14:01, Steve Maher <[email protected]> wrote:
On Mon, Mar 15, 2010 at 12:38 PM, Jason Manley
<[email protected]>wrote:
Actually, the most stable flow right now (at least I've found) is
Windows XP 32-bit with 10.1.3.1386 and Matlab R2007b. This is what I
would recommend.
Well, since I have to wait to get the license for R2007b from Matlab
(since it requires different types of keys than current releases) AND
I'm losing our only Windows XP 32-bit machine at the end of this week
(it's moving to a telescope in Spain), I'm going to try things on
Linux (Ubuntu). I think
I'll start with 11.x while I'm waiting for the license from Matlab.
If that doesn't work I'll back down to 10.1.
Simultaneously I guess I'll try to track down Windows XP 32 bit. (We
have a WIndows XP 64-bit but that doesn't seem to install Xilinx
correctly).
Any warnings about the paths I'm heading down are appreciated.
Steve
I'm still investigating the 11.x flow on Linux. It's not ready for
prime-time yet: I sometimes have Matlab disappearing on me, compiles
that sometimes take significantly longer (22hrs), ridiculous memory
usage (over 16GB) etc etc.
Jason
# Re: [casper] Simulink/Xilinx integration Jason Manley
# Re: [casper] Simulink/Xilinx integration John Ford
# Re: [casper] Simulink/Xilinx integration Jason Manley
# Re: [casper] Simulink/Xilinx integration Steve Maher
# Re: [casper] Simulink/Xilinx integration Jason Manley
# Re: [casper] Simulink/Xilinx integration Billy Mallard
# Re: [casper] Simulink/Xilinx integration Billy Mallard
# Re: [casper] Simulink/Xilinx integration John Ford
# Re: [casper] Simulink/Xilinx integration Jason Manley
# Re: [casper] Simulink/Xilinx integration Steve Maher
# [casper] Simulink/Xilinx integration Andrew Martens
