Hi,
I may know the reason for the problem. Because there's no
fftwideband_real in the CASPER DSP in simulink even I have installed the
mlib_vevel library. Actually in the FFTs there are nothing can be
used,see the picture appendix.
Does anyone knows how to install the fftwideband_real and other items in
Casper DSP blockset? Without that item the tut3 can't run.
Best Regards.
Oliver Wang
-----原始邮件-----
*发件人:* "Wang Jinqing" <[email protected]>
*发送时间:* 2014年10月19日 星期日
*收件人:* [email protected]
*抄送:*
*主题:* about error report of tut3
Hi I havedownload the tut3.mdl and I want to do the simulation in
simulink sysgen(version=14.3.4477.). But the fft_wideband_real
component always report message as:
Error in 'tut3_back/fft_wideband_real <matlab:open_and_hilite_system
('tut3_back/fft_wideband_real')>': Initialization commands cannot be
evaluated.
fft_unscrambler block (mask) does not have a parameter named
'coeffs_bit_limit'
How can I fix this problem? Does anyone run into such a problem?
Best Regards.
Oliver Wang
> -----原始邮件-----
> 发件人: [email protected]
<mailto:[email protected]>
> 发送时间: 2014年10月17日 星期五
> 收件人: [email protected] <mailto:[email protected]>
> 抄送:
> 主题: casper Digest, Vol 83, Issue 11
>
> Send casper mailing list submissions to
> [email protected] <mailto:[email protected]>
>
> To subscribe or unsubscribe via the World Wide Web, visit
>
https://calmail.berkeley.edu/manage/list/listinfo/[email protected]
<mailto:[email protected]>
>
> or, via email, send a message with subject or body 'help' to
> [email protected]
<mailto:[email protected]>
>
> You can reach the person managing the list at
> [email protected]
<mailto:[email protected]>
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of casper digest..."
>
>
> Today's Topics:
>
> 1. Re: about simulation very slowly (David MacMahon)
> 2. Re: about simulation very slowly (Jason Manley)
> 3. Re: The roach2 always appear disconnected after a few
> miniutes! (Jason Manley)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Thu, 16 Oct 2014 21:18:06 -0700
> From: David MacMahon <[email protected]
<mailto:[email protected]>>
> Subject: Re: [casper] about simulation very slowly
> To: "Wang Jinqing" <[email protected] <mailto:[email protected]>>
> Cc: [email protected] <mailto:[email protected]>
> Message-ID: <[email protected]
<mailto:[email protected]>>
> Content-Type: text/plain; charset=us-ascii
>
> Hi, Oliver,
>
> On Oct 16, 2014, at 9:06 PM, Wang Jinqing wrote:
>
> > I have construct a very easy project under matlab simulink envirement
,thant is: only a katadc component and a scope to moniter the adc's output . I want
to do simulation so I set the start time 0.0,stop time 10. I found it'll spend about
20 minutes to give out the simulation result so that I can see the scope's output(
Although the result is righ). I think there's something about configuration may not
be correct. But I don't know where I should change the settings so the simulation
speed can be increased. Could some on give me some advice?
>
> I've encountered very slow simulations when using a network mounted home
directory. It turns out that I needed to set the DSP_CACHE_DIR variable to specify a
directory on a local disk. Without that, sysgen seemed to be repeatedly trying and failing
(for some reason) to start a "dspcache" server that used my home directory on the
network mounted filesystem (maybe CIFS?). I have no idea why that would be problematic
(other than slower), but at any rate it started running better once I set DSP_CACHE_DIR to
specify a directory on a local disk.
>
> Here is a Xilinx link that proved helpful:
>
>
http://forums.xilinx.com/t5/DSP-Tools/System-Generator-works-very-slow-Why/td-p/21140
>
> Here are some commands that provided clues:
>
> > $ /tools/commercial/xilinx/14.6/ISE_DS/ISE/sysgen/bin/lin64/dccli -s
> > Current number of entries: 299
> > Current usage in KB: 463
> > Hits/Misses: 3544/312 (91.9% hit ratio)
>
> Before fixing things, "dccli -s" showed errors about not being able to
start or attach to a cache server (and all the reported numbers were 0).
>
> > $ ps -f `pgrep dspcache`
> > UID PID PPID C STIME TTY STAT TIME CMD
> > davidm 18309 1 0 16:49 ? Ssl 0:00
/tools/commercial/xilinx/14.6/ISE_DS/ISE/sysgen/bin/lin64/dspcache.bin -r 120
>
> Before things were working, no dspcache.bin processes were running. The
dspcache process is supposed to get started automatically, if needed, but I think
when things were bad it would repeatedly try and fail to start this server.
>
> In Matlab, you can check the return values from this command:
>
> >>> [core, sg, usertemp] = xlCache('getpath')
> > (Should show three directories, preferable all local)
>
> Before setting DSP_CACHE_DIR to a local directory, the "sg" cache
directory was in the home directory on the network mounted filesystem.
>
> Hope this helps,
> Dave
>
>
>
>
> ------------------------------
>
> Message: 2
> Date: Fri, 17 Oct 2014 08:35:45 +0200
> From: Jason Manley <[email protected] <mailto:[email protected]>>
> Subject: Re: [casper] about simulation very slowly
> To: David MacMahon <[email protected]
<mailto:[email protected]>>
> Cc: Wang Jinqing <[email protected]
<mailto:[email protected]>>, Casper Lists
> <[email protected] <mailto:[email protected]>>
> Message-ID: <[email protected]
<mailto:[email protected]>>
> Content-Type: text/plain; charset=us-ascii
>
> Are you on a Linux environment? Matlab's GUI doesn't seem to like some X
window managers. If you have a scope open, depending on the block selected, it
redraws the screen after each clock tick and this can cause big delays. Does your
design simulate faster if the scope windows are closed and no blocks selected
(highlighted) in your design?
>
> Jason
>
>
> On 17 Oct 2014, at 6:18, David MacMahon <[email protected]
<mailto:[email protected]>> wrote:
>
> > Hi, Oliver,
> >
> > On Oct 16, 2014, at 9:06 PM, Wang Jinqing wrote:
> >
> >> I have construct a very easy project under matlab simulink envirement
,thant is: only a katadc component and a scope to moniter the adc's output . I want to
do simulation so I set the start time 0.0,stop time 10. I found it'll spend about 20
minutes to give out the simulation result so that I can see the scope's output( Although
the result is righ). I think there's something about configuration may not be correct.
But I don't know where I should change the settings so the simulation speed can be
increased. Could some on give me some advice?
> >
> > I've encountered very slow simulations when using a network mounted home
directory. It turns out that I needed to set the DSP_CACHE_DIR variable to specify a directory
on a local disk. Without that, sysgen seemed to be repeatedly trying and failing (for some
reason) to start a "dspcache" server that used my home directory on the network
mounted filesystem (maybe CIFS?). I have no idea why that would be problematic (other than
slower), but at any rate it started running better once I set DSP_CACHE_DIR to specify a
directory on a local disk.
> >
> > Here is a Xilinx link that proved helpful:
> >
> >
http://forums.xilinx.com/t5/DSP-Tools/System-Generator-works-very-slow-Why/td-p/21140
> >
> > Here are some commands that provided clues:
> >
> >> $ /tools/commercial/xilinx/14.6/ISE_DS/ISE/sysgen/bin/lin64/dccli -s
> >> Current number of entries: 299
> >> Current usage in KB: 463
> >> Hits/Misses: 3544/312 (91.9% hit ratio)
> >
> > Before fixing things, "dccli -s" showed errors about not being able to
start or attach to a cache server (and all the reported numbers were 0).
> >
> >> $ ps -f `pgrep dspcache`
> >> UID PID PPID C STIME TTY STAT TIME CMD
> >> davidm 18309 1 0 16:49 ? Ssl 0:00
/tools/commercial/xilinx/14.6/ISE_DS/ISE/sysgen/bin/lin64/dspcache.bin -r 120
> >
> > Before things were working, no dspcache.bin processes were running.
The dspcache process is supposed to get started automatically, if needed, but I think
when things were bad it would repeatedly try and fail to start this server.
> >
> > In Matlab, you can check the return values from this command:
> >
> >>>> [core, sg, usertemp] = xlCache('getpath')
> >> (Should show three directories, preferable all local)
> >
> > Before setting DSP_CACHE_DIR to a local directory, the "sg" cache
directory was in the home directory on the network mounted filesystem.
> >
> > Hope this helps,
> > Dave
> >
> >
>
>
>
>
> ------------------------------
>
> Message: 3
> Date: Fri, 17 Oct 2014 08:39:24 +0200
> From: Jason Manley <[email protected] <mailto:[email protected]>>
> Subject: Re: [casper] The roach2 always appear disconnected after a
> few miniutes!
> To: peter <[email protected] <mailto:[email protected]>>
> Cc: Casper Lists <[email protected]
<mailto:[email protected]>>
> Message-ID: <[email protected]
<mailto:[email protected]>>
> Content-Type: text/plain; charset="windows-1252"
>
> What version of the filesystem are you running? There was an old version
with a memory leak that I recall which behaved similarly to what you describe.
Unfortunately I don't have the latest tcpborphserver binary available for you to
update your netboot environment. We are currently using romfs soloboot on our
boards. Attached is the latest binary for this in case you want to try it.
>
> Jason Manley
> CBF Manager
> SKA-SA
>
> Cell: +27 82 662 7726
> Work: +27 21 506 7300
>
> On 17 Oct 2014, at 2:49, peter <[email protected]
<mailto:[email protected]>> wrote:
>
> > Hi all,
> > For our experiment,we have 3 roach2's for a test(Let's call them
roach1, roach2, roach3).I met a weird problem that roach3 board not stable while the
other two are fine.All roach2 boards have been setup by netboot,They all run the
PAPER correlator bof file.However,after a few minutes ,roach3 can not be connected
through katcp or ssh.The roach2 will work after restart ,but after about 10 minutes
it disconnect again.Sometime it need re-setup netboot to work again.I checked the
power supply and the electric wire,but nothing seems wrong.What's more strange is
that the LEDs of power and the NIC on roach2 are lighting,and the NIC?s LED is also
blinking!
> > Has anyone met the same problem like this?Hope for your help!
> > Best Regards!
> > peter niu
> >
> >
> >
> -------------- next part --------------
> A non-text attachment was scrubbed and removed.
> Name: roach2-root-fullmap-2014-08-12.romfs.zip
> Type: application/zip
> Size: 5013152 bytes
> Desc: not available
> Non-text attachments are only available in MIME digests.
>
> End of casper Digest, Vol 83, Issue 11
> **************************************
