30 August 2013

506. Extracting optimized structures from a potential energy scan in nwchem

Another update:
It now dumps the energies in a file, energies.dat, as well.

Update:
some programmes, like ecce, are more picky about the xyz format than others (e.g. jmol, vmd). I've updated the code to output xyz files that ecce too can read.

Original post:
When you use scan_input() in nwchem to do a PES scan (see e.g. here: http://verahill.blogspot.com.au/2013/08/503-relaxed-pes-scanning-in-nwchem.html) you get the energies and the gradients for the optimized structures returned as the results. However, for a casual user the atomic actual coordinates is more informative.

Here's a very simple parser written in python (2.7) which extracts the optimized structures from the output file:

#!/usr/bin/python
import sys

def getrawdata(infile):
        f=open(infile,'r')
        opt=0
        geo=0
        energy=[]
        energies=[]
        struct=[]
        structure=[]
        for line in f:
                if "Total DFT" in line:
                        line=filter(None,line.rstrip('\n').split(' '))
                        energy=float(line[4])
                if 'Optimization converged' in line:
                        opt=1
                if opt==1 and 'Geometry' in line:
                        geo=1
                if      'Atomic Mass' in line and (opt==1 and geo==1):
                        opt=0
                        geo=0
                        struct+=[structure]
                        energies+=[energy]
                        structure=[]
                if opt==1 and geo==1:
                        structure+=[line.rstrip()]
        return struct,energies

def genxyzstring(coords,element):
        x_str='%10.5f'% coords[0]
        y_str='%10.5f'% coords[1]
        z_str='%10.5f'% coords[2]
 
        xyz_string=element+(3-len(element))*' '+10*' '+\
        (8-len(x_str))*' '+x_str+10*' '+(8-len(y_str))*' '+y_str+10*' '+(8-len(z_str))*' '+z_str+'\n'
 
        return xyz_string

def getstructures(rawdata):
        
        n=0
        for structure in rawdata:
                
                n=n+1
                num="%03d" % (n,)
                g=open('structure_'+num+'.xyz','w')
                itson=False
                cartesian=[]
                        
                for item in structure:
                        
                        if itson and not(item==""):
                                coords=filter(None,item.split(' '))
                                coordinates=[float(coords[3]),float(coords[4]),float(coords[5])]
                                element=coords[1]
                                cartesian+=[genxyzstring(coordinates,element)]
                                #cartesian+=[coords[1]+'\t'+coords[3]+'\t'+coords[4]+'\t'+coords[5]+'\n']
                
                        if "---" in item:
                                itson=True
                        if item=="" and itson==True:
                                itson=False
                                if not(len(cartesian)==0):
                                        g.write(str(len(cartesian))+'\n')
                                        g.write('Structure '+str(n)+'\n')
                                        for line in cartesian:
                                                g.write(line)
                                        g.close()
                                cartesian=[]
        return 0
        
if __name__ == "__main__":
        infile=sys.argv[1]
        rawdata,energies=getrawdata(infile)
        structures=getstructures(rawdata)

        g=open('energies.dat','w')
        for n in range(0,len(energies)):
                g.write(str(n)+'\t'+str(energies[n])+'\n')
        g.close()


Presuming that you've saved it as pes_parse.py you can then generate a series of xyz files with the structures, catenate them into a trajectory file, and open it in e.g. jmol. I'm using the output from example 1 in http://verahill.blogspot.com.au/2013/08/503-relaxed-pes-scanning-in-nwchem.html as the example:

chmod +x pes_parse.py
./pes_parse.py nwch.nwout
ls
nwch.nwout structure_001.xyz structure_003.xyz structure_005.xyz structure_007.xyz structure_009.xyz structure_011.xyz structure_013.xyz structure_015.xyz structure_017.xyz structure_019.xyz pes_parse.py structure_002.xyz structure_004.xyz structure_006.xyz structure_008.xyz structure_010.xyz structure_012.xyz structure_014.xyz structure_016.xyz structure_018.xyz
cat structure_*.xyz >> trajectory.xyz jmol trajectory.xyz

You can go through the structures by clicking on the arrows indicated by the white arrow:

Finally, using VMD it's easy to make videos -- note that they for some reason look awful here (seems like a lot of frames are removed, in particular from the beginning of the run):

And here's the SN2 reaction from post 503:


29 August 2013

505. Testing KDE on Debian (if you're a gnome user) -- getting set up

Every now and again I get ticked off with something to do with planned 'features' in gnome 3. I'm still not happy about gnome-screenshot, but I patched and compiled my own version, so I'm doing alright. I don't really like the transition of menus from the, well, menu bar of applications to the top bar of the gnome desktop, but it's not a great issue.

Oh, and I don't really use vanilla gnome 3 anyway -- I've got a desktop set up with a top panel, and bottom panel, static desktops etc. The latest 'feature' which is making me want to move to a different desktop is the planned change in the use of the middle-button so that it no longer can be used to paste text.

(of course there's a bit of inconsistency here: KDE turns out to have a remapped middle-click, the screenshot application needs a few more clicks than the old gnome-screenshot etc.)

I've toyed with the idea of changing desktops in the past, and for a while was running nothing but KDE on my laptop (I'd rather switch on my 'casual' computer first, before making the transition on my work computer), but for some reason I switched back. I actually can't remember why, but something really annoyed me about KDE.

Before I forget: a good reason to try another desktop is simply to reassure yourself that you do have alternatives if your chose DE suddenly turns into something you can't work with.

Anyway, time to give it another go. But what desktop? I like the idea of going for KDE since it's 'the other desktop' -- it's been around longer than gnome, whereas lxde and xcfe are mere babies in comparison. Alternatively, one could always piece together a custom desktop using e.g. fluxbox. Or switch to something completely different, like xmonad. Or something a little bit different, like Enlightenment. I think I should probably be going for LXDE (I instinctively don't like XFCE), but let's go for KDE first.

Finally, the main danger is switching desktops and expecting the new one to act exactly like the old one (minus the bad parts, obviously). Which is pretty much how windows users (myself included once upon a time) behave when trying out linux for the first time.

Anyway, time for a week in KDE (4.10.5).

Here's my gnome desktop at the time of switching:

In addition, I'm using Guake, and my network manager is wicd.

So let's see what KDE will look like at the end -- whether I'll be able to use KDE as KDE, instead of as an ersatz gnome.

Step 1. Install KDE
sudo apt-get install kde-full wicd-kde kdesudo yakuake

I selected kdm.

Step 2. First boot
Reboot and make sure to click the little sun icon to select the desktop you want to use, in my case KDE/plasma.

2.1 Switching workspaces
I next set up proper workspace switching so that it behaves like in...gnome (falling into the old trap of making the new desktop just like the old one)...i.e. ctrl+alt right/left arrow.

Go to System Settings/WorksSpace behaviour, Virtual Desktops -- I incrased the number of desktops to 6 (1 row), then clicked on the Switching tab -- I set shortcuts for Switch One Desktop to the Left and Switch One Desktop to the Right to ctrl+alt+left and ctrl+at+right.

2.2 Disable nepomuk
At this point I noticed that I was using 100% of the CPU. virtuoso-t was chewing up a lot, as was nepomukservices. I had a look at this post: https://forum.kde.org/viewtopic.php?f=66&t=98892



I first went to Search and Indexing (in the menu), unchecked Monitor file and directory changes, Enable when running on battery, Only when computer is not being used.  You can also go to System Settings/Desktop Search and uncheck Enable Nepomuk Semantic Desktop.

2.3 konqueror crashing
I also had repeated issues with konqueror crashing when clicking on links, but it seems to have resolved itself after disabling nepomuk.

2.4 guake/yakuake
Guake worked from the beginning -- I didn't have to set it up at all. However, I still ended up changing to yakuake since I figure it's a bit more KDE like.To get yakuake to remember changed settings (e.g. height, width) you'll need to close it manually. Restarting it will use the new settings. Not sure how I got yakuake to autostart -- I think it just happened.


2.5 The bottom panel
I had a difficult time figuring out how to add shortcuts to the panel, and once I had accidentally figured it out, I promptly screwed up the entire bottom panel, then managed to fix it again.

2.6 Keyboard shortcuts
I then had to figure out how to bind shortcuts to opening applications -- right-click on the menu ('kicker'), click Edit Applications, select your application, advanced, Advanced, Current Shortcut Key.
A more general approach to creating shortcuts was shown to me by an anonymous commentator (see below this post) -- I've posted the screenshot provided to me below:
shortcuts, gestures and whatnots


2.7 Update notification
I promptly removed the Update Notification widget -- it was driving me crazy and reminds me too much of windows. I first tried simply disabling the update notifier:



I then did apt-get autoremove update-notifier-kde. The problem -- even after a reboot I still get notifications about updates! I then did System Settings/Software Management, and click on the tiny, tiny button on the far right (level with the search field), and then click on settings and set Check for updates to never.  This finally seems to have done it.

2.8 Theme
I also changed from the default 'air' theme to 'oxygen'.


2.9 Gnome/GTK programs in KDE
To make gnome (well, gtk) programs look acceptable under KDE, install gtk2-engines-oxygen and gtk3-engines-oxygen and reboot.

2.10 Tap-to-click with synaptics
I can't get synaptics tap to click to work. Based on the odd post I tried gpointing-device-settings, but while there's a checkbox to disable tap to click, it won't enable it. I then installed kde-config-touchpad and was happy again. Funnily enough this made me discover that middle-click is mapped in KDE -- the one reason I wanted to explore something other than gnome...oh well.


2.11 Conky
To get conky to work properly I only had to make some minimal changes. /etc/conky/conky.conf:
background no own_window yes own_window_type normal own_window_argb_visual true own_window_transparent yes own_window_hints undecorated,below,sticky,skip_taskbar,skip_pager double_buffer yes alignment top_right background yes border_width 1 cpu_avg_samples 2 default_color white default_outline_color white default_shade_color white default_graph_size 20 200 draw_borders no draw_graph_borders yes draw_outline no draw_shades no use_xft yes xftfont DejaVu Sans Mono:size=12 gap_x 20 gap_y 60 minimum_size 5 5 net_avg_samples 2 no_buffers yes out_to_console no out_to_stderr no extra_newline no own_window_class Conky own_window_transparent yes stippled_borders 0 update_interval 1.0 uppercase no use_spacer none show_graph_scale no show_graph_range no TEXT ${exec lsb_release -ds} ${execi 5 acpi|gawk '{print $3,$4}'} ${addr eth0}/${addr wlan0} $hr ${color grey}Uptime:$color $uptime CPU:$alignc $cpu% $alignc$color ${cpugraph 10,200 ffff00 ff0000 -t} RAM:$alignc $mem/$memmax $alignc$color ${memgraph 10,200 ffff00 ff0000 -t} I/O:$alignc $diskio_read/$diskio_write $alignc$color ${diskiograph 10,200 ffff00 ff0000 -t} ${color grey}Frequency (in GHz):$color ${freq_g 1}, ${freq_g freq_g2} ${color grey}Temperature: $color CPU: ${acpitemp}°C $hr ${color grey}File systems:$alignr $color/ ${fs_used /}/${fs_size /} $alignr /home $color${fs_used /home}/${fs_size /home} ${color grey}Networking: ${color grey}eth0 ${color red} ${upspeed eth0}/${color green} ${downspeed eth0} ${color grey} ${color grey} ${upspeedgraph eth0 10,100 ffff00 ff0000 -t} ${color grey} ${downspeedgraph eth0 10,100 0000ff 00ff00 -t} ${color grey} ${color grey}wlan0 ${color red} ${upspeed wlan0}/${color green} ${downspeed wlan0} ${color grey} ${color grey} ${upspeedgraph wlan0 10,100 ffff00 ff0000 } ${color grey} ${downspeedgraph wlan0 10,100 0000ff 00ff00 } ${color grey} $hr ${color grey}Name CPU% MEM% ${color} ${top name 1} ${top cpu 1} ${top mem 1} ${color} ${top name 2} ${top cpu 2} ${top mem 2} ${color} ${top name 3} ${top cpu 3} ${top mem 3} $hr ${color}City Ping $alignr Time $font${color}San Francisco ${execi 60 ping -c 1 sfgate.com -n|grep icmp_seq|sed 's/=/\t/g'|gawk '{print $10,$11}'} $alignr${tztime America/Los_Angeles %a %H:%M} $font${color}Copenhagen $alignr${tztime Europe/Copenhagen %a %H:%M} $font${color}Shanghai $alignr${tztime Asia/Shanghai %a %H:%M} $font${color}Melbourne $alignr${tztime Australia/Melbourne %a %H:%M


And I've discovered that I really like the hot corner (top left) in gnome. There is a hot corner in kde too, but it does slightly different things.

KDE is pretty but it still gives me the occasional feeling that I'm using windows due to the bottom panel, the menu on the left and the task tray on the right, but these are things I'll hopefully grow out of. But yes, KDE is pretty -- very pretty. And it's pretty fast too.
My old gnome desktop (same as at the beginning of this post)

and my new KDE desktop
And one just to show how versatile KDE is -- you can even make it look (almost) like gnome 3

So...now it just remains to see how I feel after a week or so. The baby duck syndrome is pretty powerful...and I'm already well on my way to turning KDE into GNOME. But then, isn't the versatility of KDE the real selling point?

Postscript:
I ended up re-enabling nepomuk and spent some time struggling with akonadi to get kmail working -- the key to getting akonadi to work was to install the same server as the backend. In my case I did:
sudo apt-get install akonadi-backend-mysql mysql-server

28 August 2013

504. Swap file and hibernation on debian

I've got an SSD and 4 gb ram on my laptop, and have no desire to use to actual swap during normal use. However, I'd like to be able to let my laptop go into hibernation -- even suspend seems to be draining my battery pretty fast (at maybe a quarter of the rate of keeping the laptop on).

Does it make sense hibernating a laptop with SSD i.e. one which boots in ten seconds flat? Probably not. But we humans are greedy by nature.

So I need to set up a swap file, disable swapping, and see if I can use it for hibernation.

As usual, the best source of information is the archlinux wiki.
https://wiki.archlinux.org/index.php/Swap#Swap_file
https://wiki.archlinux.org/index.php/Pm-utils#Using_Swap_file_instead_of_regular_swap_partition

Anyway, having made it work and tried it out I would say this is really not worth the hassle IF you have an SSD. 


1. The swap file
This is my file system layout:

Filesystem      Size  Used Avail Use% Mounted on
/dev/sda1        39G   20G   18G  53% /
udev             10M     0   10M   0% /dev
tmpfs           380M  816K  380M   1% /run
tmpfs           5.0M     0  5.0M   0% /run/lock
tmpfs           760M  1.1M  759M   1% /run/shm
/dev/sda2       109G   71G   33G  69% /home

In my case I'd say that / has enough space to handle a swap file. My RAM is '4 gb' (really 3.71 Gb) -- according to this and this my swap should equal my RAM, which makes intuitive sense.

Since 4,000,000,000 byte is 3.76 gb, and 4000*1024*1024 (i.e. 4000M) is 3.91 Gb, I think 4000M should be ok:

su -
fallocate -l 4000M /swapfile
chmod 600 /swapfile
mkswap /swapfile
Setting up swapspace version 1, size = 4095996 KiB no label, UUID=2a8de3d1-14f6-473f-b40f-31618fd81169
echo 'vm.swappiness=1' >> /etc/sysctl.d/50-local.conf echo '/swapfile none swap defaults 0' >> /etc/fstab



2. "PM: Swap header not found"

To try it out without rebooting:
sudo sysctl -w vm.swappiness=1
sudo swapon /swapfile
pm-is-supported --hibernate 
echo $?
0
If you got 0, then you're good to go.
sudo pm-hibernate

Trying it out the first time I got "PM: swap header not found" and some weird behaviour. This has been mentioned e.g. here. A step-by-step guide is here: https://ubuntuforums.org/showthread.php?t=1042946

Get the UUID of the partition on which the swapfile is located:
mount | grep " / "
/dev/disk/by-uuid/8adf424c-c375-4035-8d5d-181489b4461b on / type ext4 (rw,noatime,nodiratime,discard,errors=remount-ro,data=ordered)
sudo filefrag -v /swapfile | grep "First block:"

The latter command gave nothing, so I then did:
sudo filefrag -v /swapfile|less
Filesystem type is: ef53 File size of /swapfile is 4194304000 (1024000 blocks of 4096 bytes) ext: logical_offset: physical_offset: length: expected: flags: 0: 0.. 0: 7182336.. 7182336: 1: 1: 1.. 6143: 7182337.. 7188479: 6143: unwritten 2: 6144.. 8191: 7190528.. 7192575: 2048: 7188480: unwritten
So now we have the UUID (8adf424c-c375-4035-8d5d-181489b4461b) and the offset (7182336).
su -
echo "resume=UUID=8adf424c-c375-4035-8d5d-181489b4461b resume_offset=7182336" | sudo tee /etc/initramfs-tools/conf.d/resume
exit

Edit /etc/default/grub and add the same line to GRUB_CMDLINE_LINUX_DEFAULT:
GRUB_CMDLINE_LINUX_DEFAULT="quiet resume=UUID=8adf424c-c375-4035-8d5d-181489b4461b resume_offset=7182336"
Run
sudo update-grub
sudo update-initramfs -u

Reboot.

3. Hibernating.
sudo pm-hibernate

On my lenovo sl410 what I see is the screen go blank save for a blinking "-", and the little crescent light at the front of the laptop starting to blink (next to the battery light).

After 5-10 seconds the laptop turns off.

Hitting the power button starts up the laptop -- you get the bios screen, then the grub menu, and at this point you're thinking that the whole exercise has failed -- but it ends up booting into the same situation as you had when you did pm-hibernate i.e. it worked. I mean, this is how hibernation works -- but I had honestly never used anything other than suspend before, so...well..there's a first time for everything.

Overall, on a laptop with an SSD, resuming from hibernation is about as slow as a cold start -- with entering hibernation taking longer than a shutdown. On a laptop with a spinning disk this could presumably be worth it.

503. (relaxed) PES scanning in Nwchem revisited.

Update 2: The coordinates are actually gradients, and so aren't terribly informative to a casual user like myself. See this post for how to extract the geometries properly: http://verahill.blogspot.com.au/2013/08/506-extracting-optimized-structures.html


Update:
Please note that the coordinates in square brackets ([]) in the python output are not raw coordinates for the atoms in the molecule -- I haven't quite figured out how they scale, but it's not a simple matter of just multiplying. The energies are good though, and you can always extract the coordinates the slow and painful way by manually going through the output.

Another issue which should be stressed is that scan_input(geom,[1.398],[3.398],19,'dft',task_optimize) does not do the end points -- i.e. you won't get the energy for a bond length of 1.398, and you won't get the energy for a bond length of 3.398. Instead you'll get 19 data points in between these. It's a bit...awkward.

Original post:
A long time ago I made a post on doing potential energy surface (PES) scans in nwchem using python.

This is a post giving PES another look. The impetus for the post is that I'm tired of Gaussian failing and being opaque about the whole procedure.

The following page was of great help: http://www.fqt.izt.uam.mx/html/software_fqt/user/node34.html

NOTE: you'll need to compile nwchem with python support. See e.g. http://verahill.blogspot.com.au/2013/06/449-nwchem-63-updated-sources-compiling.html (the post is a bit messy, but persevere -- it's not that difficult)

On Debian the key is to change
    EXTRA_LIBS +=    -lnwcutil  -lpthread -lutil -ldl
to
    EXTRA_LIBS +=    -lnwcutil  -lpthread -lutil -ldl -lssl -lz

in config/makefile.h before compiling. It's not necessary on RHEL clones.

Below I'll show three examples:
* a simple bond dissociation reaction. I also discuss the use of 'constant', and task_energy vs task_optimize.
* an SN2 reaction (CH3Br + I-)
* a 2D/parallel PES scan of ethane ( C-C bond length, H-C-C angle). I also show constant vs free variables.


Example 1.
Breaking the C-O bond in methanol

I set this up in ecce (see e.g. next example), but you don't have to. The input file I used was the following:
scratch_dir /scratch Title "meoh_pes" Start meoh_pes echo charge 0 geometry autosym units angstrom C 0.0351714 0.00548884 0.0351714 H -0.617781 -0.634073 0.667983 H 0.667983 -0.634073 -0.617781 H -0.605139 0.646470 -0.605139 O 0.839603 0.818768 0.839603 H 1.38912 0.201564 1.38912 end ecce_print ecce.out basis "ao basis" cartesian print H library "3-21G" O library "3-21G" C library "3-21G" END dft mult 1 direct XC b3lyp grid fine iterations 99 mulliken end driver default maxiter 888 end python from nwgeom import * geom = ''' geometry adjust zcoord bond 1 5 %f cccc constant end end ''' results=scan_input(geom,[1.398],[3.398],19,'dft',task_optimize) for i in range(0,len(results)): print results[i][0][0],results[i][1] end task python
The PES bit is highlighted in blue. Note the 'constant' keyword -- if you omit that the bond length will initially be set to whatever you define it to in your scan, but it can relax back to the optimum length. If you DO set 'constant' everything BUT that bond will be relaxed. Most likely this is what you will want to do.

Also note that a constrained (i.e. not relaxed) PES scan can be done by doing task_energy instead of task_optimize.

ECCE can't quite handle the textual output (alt+O) since there are lines that are too long. The output is properly written though -- you'll just have to look in the Output folder of the job. The ecce.out file works fine though.

The job takes 90-100 seconds on an old 3-core node (AMD Athlon II X3).


The very end of the output file has all the results, but in a non-obvious way:
1.498 (-115.07289914310056, [-0.00130778291169336, 0.01798903956433226, 0.0, -4.009155466250247e-05, 1.693340302064139e-05, -6.637550254401381e-06, -4.009155466250247e-05, 1.693340302064139e-05, 6.637550254401381e-06, 2. 4514244186701895e-05, -1.5885649893555842e-05, 0.0, 0.0012636893525275195, -0.018041103298149008, 0.0, 9.97624242821682e-05, 3.4082577691996185e-05, 0.0]) (-114.8737952986994, [-4.7287277448850376e-05, 0.030029200359777717, 0.0, -1.3711175166353229e-06, -8.452926738775068e-08, 9.941241931599176e-07, -1.3711175166353229e-06, -8.452926738775068e-08, -9.941241931599176e-07, 8. 167348279908282e-07, -2.5820569179275075e-06, 0.0, 4.871429991895604e-05, -0.030027845123621805, 0.0, 4.984777179639632e-07, 1.3958792967685985e-06, 0.0]) 1.498 (-115.07289914310056, [-0.00130778291169336, 0.01798903956433226, 0.0, -4.009155466250247e-05, 1.693340302064139e-05, -6.637550254401381e-06, -4.009155466250247e-05, 1.693340302064139e-05, 6.637550254401381e-06, 2. 4514244186701895e-05, -1.5885649893555842e-05, 0.0, 0.0012636893525275195, -0.018041103298149008, 0.0, [..] 3.198 (-114.87977711993531, [-0.00018979360652668711, 0.033296276783081655, 0.0, -2.3787379704320877e-06, 1.7510009376556918e-06, 1.3530564600128248e-06, -2.3787379704320877e-06, 1.7510009376556918e-06, -1.3530564600128248e-06, 8. 24207064487048e-06, -8.055936327900498e-07, 0.0, 0.00018027576986845428, -0.03329589479259992, 0.0, 6.033241931824307e-06, -3.0783987173960137e-06, 0.0]) 3.298 (-114.8737952986994, [-4.7287277448850376e-05, 0.030029200359777717, 0.0, -1.3711175166353229e-06, -8.452926738775068e-08, 9.941241931599176e-07, -1.3711175166353229e-06, -8.452926738775068e-08, -9.941241931599176e-07, 8. 167348279908282e-07, -2.5820569179275075e-06, 0.0, 4.871429991895604e-05, -0.030027845123621805, 0.0, 4.984777179639632e-07, 1.3958792967685985e-06, 0.0])
All in all, there are 58 lines for 19 steps. I think that there are three things happening -- firstly, the line in blue is the output from the 19th step, and that somehow gets mixed in with the output from all the calculations. Secondly, the structure and energy of each step is reported twice at a time. Thirdly, the optimised structures/energies are reported one more time by injecting them into the output, like this:
A
S
A
B
B
C
C
D
D
A
E
E
B

where A is the first step, S is the 19th step etc. This way you get 19x3+1=58 lines. This is clearly idiotic.

Instead, you can look through the output and search for 'Scanning NWChem input - results from step' to see all the output for the optimised structures one by one:
Scanning NWChem input - results from step 2 (-115.06618436935011, [-0.0038228970733096973, 0.050051062094932305, 0.0, 2.9196769046224702e-05, -6.928661348853948e-06, 4.746536668570611e-06, 2.9196769046224702e-05, -6.928661348853948e-06, -4.746536668570611e-06, -1.0103262985700079e-05, 1.6491089715894858e-05, 0.0, 0.003767244388907326, -0.05005618579508188, 0.0, 7.362409274846993e-06, 2.489933151654522e-06, 0.0])
In this particular case I can grep my way through by doing
cat nwch.nwout |grep '^(-'|cat -n
1 (-115.07289914310056, [-0.00130778291169336, 0.01798903956433226, 0.0, -4.009155466250247e-05, 1.693340302064139e-05, -6.637550254401381e-06, -4.009155466250247e-05, 1.693340302064139e-05, 6.637550254401381e-06, 2.4514244186701895e-05, -1.5885649893555842e-05, 0.0, 0.0012636893525275195, -0.018041103298149008, 0.0, 9.97624242821682e-05, 3.4082577691996185e-05, 0.0]) 2 (-115.06618436935011, [-0.0038228970733096973, 0.050051062094932305, 0.0, 2.9196769046224702e-05, -6.928661348853948e-06, 4.746536668570611e-06, 2.9196769046224702e-05, -6.928661348853948e-06, -4.746536668570611e-06, -1.0103262985700079e-05, 1.6491089715894858e-05, 0.0, 0.003767244388907326, -0.05005618579508188, 0.0, 7.362409274846993e-06, 2.489933151654522e-06, 0.0]) 3 (-115.05478103866017, [-0.005033784212299788, 0.06848598587431667, 0.0, -1.3396548676491982e-06, -2.5875637174599397e-08, -5.261746410523127e-07, -1.3396548676491982e-06, -2.5875637174599397e-08, 5.261746410523127e-07, 1.4459720645843e-07, -2.8328952926398587e-06, 0.0, 0.005034455335082233, -0.0684825786855032, 0.0, 1.8635897582608418e-06, -5.225422206114883e-07, 0.0]) 4 (-115.04079235517, [-0.005485543277166251, 0.07798880362126945, 0.0, 4.745460307237215e-06, -5.597510268573469e-06, 5.645418744981701e-07, 4.745460307237215e-06, -5.597510268573469e-06, -5.645418744981701e-07, -6.651712157745848e-07, 6.750842351778419e-06, 0.0, 0.00548062073181968, -0.07798086728839469, 0.0, -3.903204054994669e-06, -3.4921546817404114e-06, 0.0]) 5 (-115.02560006674966, [-0.0054233976595857575, 0.08166232318137269, 0.0, -1.659239761503395e-06, -4.376603580866223e-07, 4.4580035316599265e-06, -1.659239761503395e-06, -4.376603580866223e-07, -4.4580035316599265e-06, 3.034808945895362e-06, -6.726118036586015e-06, 0.0, 0.005436665955901393, -0.08164730868562775, 0.0, -1.2984625724410392e-05, -7.4130570159938736e-06, 0.0]) [..] 16 (-114.89364787840326, [-0.0005591249462735259, 0.04018795560035916, 0.0, -5.34666220519675e-07, 1.1370871814235517e-06, 4.809133242467123e-07, -5.34666220519675e-07, 1.1370871814235517e-06, -4.809133242467123e-07, -6.9140095421138525e-06, -3.095664552260277e-06, 0.0, 0.0005695756951453745, -0.040185884820554796, 0.0, -2.467406898132296e-06, -1.2492896190128416e-06, 0.0]) 17 (-114.8863872514371, [-0.00036666056940981573, 0.03667976502852128, 0.0, 2.9101399354747315e-06, -2.094045026924257e-06, -4.933288234976185e-06, 2.9101399354747315e-06, -2.094045026924257e-06, 4.933288234976185e-06, 1.6531622304416516e-07, 1.511517903679191e-07, 0.0, 0.00036162347288279384, -0.03668602744257765, 0.0, -9.484995716624312e-07, 1.0299352320775057e-05, 0.0]) 18 (-114.87977711993531, [-0.00018979360652668711, 0.033296276783081655, 0.0, -2.3787379704320877e-06, 1.7510009376556918e-06, 1.3530564600128248e-06, -2.3787379704320877e-06, 1.7510009376556918e-06, -1.3530564600128248e-06, 8.24207064487048e-06, -8.055936327900498e-07, 0.0, 0.00018027576986845428, -0.03329589479259992, 0.0, 6.033241931824307e-06, -3.0783987173960137e-06, 0.0]) 19 (-114.8737952986994, [-4.7287277448850376e-05, 0.030029200359777717, 0.0, -1.3711175166353229e-06, -8.452926738775068e-08, 9.941241931599176e-07, -1.3711175166353229e-06, -8.452926738775068e-08, -9.941241931599176e-07, 8.167348279908282e-07, -2.5820569179275075e-06, 0.0, 4.871429991895604e-05, -0.030027845123621805, 0.0, 4.984777179639632e-07, 1.3958792967685985e-06, 0.0])
Not pretty, but manageable.
cat nwch.nwout |grep '^(-'|sed 's/\,/\t/g;s/(\([^)]*\))/\1/g'|cat -n|gawk '{print $1,$2}' > profile.dat

and then plot it:


Example 2.
SN2 reaction between iodide and bromomethane

You can set up your calc however you want, but ECCE is easier than anything else.

Draw bromomethane, then throw in an iodine atom. Adjust the angle across Br-C-I to 180 degrees, and set the C to I distance to 3 Ã….


Set up the calculation -- in this case I used b3lyp/def2-svp
Edit the input and add
python from nwgeom import * geom = ''' geometry adjust zcoord bond 1 6 %f cccc constant end end ''' results=scan_input(geom,[3.00],[1.5],20,'dft',task_optimize) for i in range(0,len(results)): print results[i][0][0],results[i][1] end task python

(Delete 'task dft optimize')

You'll now have the following input file:
scratch_dir /scratch
Title "sn2_br"

Start  sn2_br

echo

charge -1

geometry noautosym units angstrom
 C     0.00000     0.00000     0.00000
 H     -0.675500     -0.675500     0.675500
 H     0.675500     -0.675500     -0.675500
 H     -0.675500     0.675500     -0.675500
 Br     1.10274     1.10274     1.10274
 I     -1.73205     -1.73205     -1.73205
end

ecce_print ecce.out

basis "ao basis" spherical print
  H library "def2-svpd"
  Br library "def2-svpd"
  C library "def2-svpd"
  I library "def2-svpd"
END
ECP
  I library "def2-ecp"
END

dft
  mult 1
  direct
  XC b3lyp
  grid fine
  iterations 99
  mulliken
end

driver
  default
  maxiter 99
end


python
from nwgeom import *
geom = '''
    geometry adjust
        zcoord
            bond 1 6 %f cccc constant
        end
    end
'''
results=scan_input(geom,[3.00],[1.5],20,'dft',task_optimize)
for i in range(0,len(results)):
    print results[i][0][0],results[i][1]
end


task python
Launch it and wait...eventually (2h 30 min on a slow three-core node) you'll get an output like the one below. Note that I didn't pre-optimise the bromomethane, so there's a bit of a drop in energy at the beginning. Likewise, I let the C-I distance get so short that the energy is rising rapidly at the end
Structure at the beginning

Transition-state-ish structure

Product


Example 3:
Two-dimensional PES scan

I'll keep this brief. First we do a scan where we use 'constant' for the angle, but not the bond length:
scratch_dir //scratch Title "2d_pes-1" Start 2d_pes-1 echo charge 0 geometry noautosym units angstrom C -2.51242e-66 1.67495e-66 -0.767732 H -0.722530 0.722530 -1.16548 H -0.264464 -0.986995 -1.16548 H 0.986995 0.264464 -1.16548 C 2.51242e-66 -2.51242e-66 0.767732 H 0.264464 0.986995 1.16548 H -0.986995 -0.264464 1.16548 H 0.722530 -0.722530 1.16548 end ecce_print ecce.out basis "ao basis" cartesian print H library "6-31G" C library "6-31G" END dft mult 1 direct XC b3lyp grid fine iterations 99 mulliken end driver default end python from pes_scan import pes_scan geom = ''' geometry noprint adjust zcoord bond 1 5 %f cc angle 2 1 5 %f hcc constant end end ''' results = pes_scan(geom, \ [1.535, 111.269], [1.800, 90], 5, 'dft', task_optimize) end task python

And the output:
What's happening is that the bond length ends up being the same no matter what we initially set it to

If we instead set constant for the bond as well:
python from pes_scan import pes_scan geom = ''' geometry noprint adjust zcoord bond 1 5 %f cc constant angle 2 1 5 %f hcc constant end end ''' results = pes_scan(geom, \ [1.535, 111.269], [1.800, 90], 5, 'dft', task_optimize) end task python

And we get:

27 August 2013

502. Spell checking in WPS office on linux -- changing language by replacing the default files

To my great shame I am using a piece of non-FOSS to deal with MS Office files -- namely, WPS. As a junior faculty member it'd be professional suicide to try to force other people to deal with the mis-rendered libreoffice files in MS Office, and for some reason no-one uses PDF anymore when sending out forms...

Anyway, in spite of not being open source and not supporting any open formats, WPS can read and save .doc and .docx files in a way that works together with MS Office, and since it runs natively on linux it's a practical solution until the day libre/openoffice become viable alternatives.

There doesn't seem to be any simple way of changing language beyond replacing the default dictionary files. It ain't pretty, but it works.

In my case I wanted Australian English, so I first installed the myspell dictionary:
sudo apt-get install myspell-en-au
mkdir ~/.dictionaries
sudo mv /opt/kingsoft/wps-office/office6/dicts/main.aff /opt/kingsoft/wps-office/office6/dicts/en_us.aff
sudo mv /opt/kingsoft/wps-office/office6/dicts/main.dic /opt/kingsoft/wps-office/office6/dicts/en_us.dic
sudo cp /usr/share/hunspell/en_AU.dic /opt/kingsoft/wps-office/office6/dicts/main.dic
sudo cp /usr/share/hunspell/en_AU.aff /opt/kingsoft/wps-office/office6/dicts/main.aff

Start WPS and it should now speak 'Strine.

24 August 2013

501. Briefly: Adding a new node to SGE

I've done this a couple of times by now, and I always forget one step or another. Most of the information is on http://verahill.blogspot.com.au/2012/06/setting-up-sun-grid-engine-with-three.html but here it is in a briefer form:

In the example I've used krypton as the node name, and 192.168.1.180 as the IP.
My front node is called beryllium and has an IP of 192.168.1.1.

0. On the front node
Add the new node name to the front node/queue master

Add execution host
qconf -ae 

which opens a text file in vim

Edited hostname (krypton) but nothing else. Saving returns
added host krypton to exec host list
Add krypton as a submit host
qconf -as krypton
krypton added to submit host list
Doing this before touching the node makes life a little bit easier.

1. Edit /etc/hosts on the node
Leave
127.0.0.1 localhost
but remove
127.0.1.1 krypton
and make sure that it says
192.168.1.180 krypton
instead.

Throw in
192.168.1.1 beryllium
as well.

2. Install SGE on node
sudo apt-get install gridengine-exec gridengine-client

You'll be asked about
Configure automatically: yes Cell name: rupert Master hostname: beryllium
3. Add node to queue and group
I maintain separate queues and groups depending on how many cores each node has. See e.g. http://verahill.blogspot.com.au/2012/06/setting-up-sun-grid-engine-with-three.html for how to create queues and groups.

If they already exits, just do

qconf -aattr hostgroup hostlist krypton @fourcores
qconf -aattr queue slots "[krypton=4]" fourcores.q

to add the new node.

4. Add pe to queue if necessary
Since I have different queues depending on the number of cores of a node, I tend to have to fiddle with this.

See e.g. http://verahill.blogspot.com.au/2012/06/setting-up-sun-grid-engine-with-three.html for how to create pe:s.

If the pe you need is already created, you can do
qconf -mq fourcores.q

and edit pe_list

5. Check
On the front node, do
qhost
HOSTNAME ARCH NCPU LOAD MEMTOT MEMUSE SWAPTO SWAPUS ------------------------------------------------------------------------------- global - - - - - - - beryllium lx26-amd64 3 0.16 7.8G 5.3G 14.9G 398.2M boron lx26-amd64 6 6.02 7.6G 1.6G 14.9G 0.0 helium lx26-amd64 2 - 2.0G - 1.9G - lithium lx26-amd64 3 - 3.9G - 0.0 - neon lx26-amd64 8 8.01 31.4G 1.3G 59.6G 0.0 krypton lx26-amd64 4 4.01 15.6G 2.8G 14.9G 0.0

23 August 2013

500. Changing language in Debian (GNOME 3, terminal) to e.g. fr_CH

This is both obvious, and tricky at the same time.

1. Install locales
The first step is to install the languages you need
sudo apt-get install locales
sudo dpkg-reconfigure locales

2. Optional: language packs
If you want e.g. libreoffice to use the new language, you'll need to install the corresponding language pack. Most programmes will NOT need this, but libreoffice and iceweasel do. The same goes for KDE.
sudo apt-get install libreoffice-l10n-fr iceweasel-l10n-fr icedove-l10n-fr

Basically, search for packages with l10n in their names.

3. GNOME
Select the new language in GNOME 3 by gong to Region and Language:
Now, here's the trick: even if the language you want is listed, click on the "+" sign in the bottom left corner, and select it like this:
While you now have two 'French' items, this is necessary or the language will not be correctly set. If you do NOT do this some things will change language, some won't (e.g. Activities won't change to Activités). In that case, if you log out, then back in and go to Languages you'll find that the language is set to "Unspecified [ANSI_X3.4-1968]". See the screenshot at the very  bottom of this post for an example of what it looks like.


Log out of GNOME and log back in again. Et voilà.
And...we have French! French everywhere.


4. Optional: set it in terminal
If you set the language in GNOME as shown above you shouldn't need to do this. However, if you are not using GNOME 3 you may want to set the language explicitly in your terminal by sticking the following in your ~/.profile or ~/.bashrc
export LANG="fr_CH.utf8"

When it's NOT OK:
Note how it says 'Pays et Langue' but still says 'Activities' i.e. the language for some items has been changed, but not all.
See step 3 for an explanation of what's going on.

499. Briefly: Drawing NMR sequences using metapost and Mark White's pulse.mp

Since it's Friday afternoon and I'm not likely to get anything useful done in the hour that remains before going home, I might as well put up another post.

Since posting http://verahill.blogspot.com.au/2013/08/498-briefly-drawing-nmr-pulse-sequences.html I've had a look at this: http://www.celos.net/comp/pulses/

And that actually is (almost -- let's not get carried away here) exactly what I have been looking for. The main issue was that it's meant for metapost -- or rather, the main issue was my unfamiliarity with metapost. Anyway, my life is know complete.

So here's how to get started...

First 'install' the metapost script:
sudo apt-get install texlive-metapost
mkdir ~/texmf/metapost
cd ~/texmf/metapost -p
wget http://www.celos.net/comp/pulses/pulses.mp
sudo texhash

Next, time to test-drive it
mkdir ~/tmp/pulse_test -p
cd ~/tmp/pulse_test
vim test.mp
input pulses.mp beginfig(0); initf; startline(rf,"RF"); xline(1); xpulse(0.5,0.5,"90"); ospan(-0.25,-1.0,1.5,"d1"); xline(0.75); xpulse(1.0,0.5,"180"); ospan(-0.5,-1.0,1.5,"d1"); xline(1.0); ospan(0,-1.0,1.5,"vd"); xline(1.5); xacq(2); endfig;
mpost test.mp mptopdf test.0 pdftops -eps test-0.pdf test-0.eps





Now, I somehow suspect you can embed metapost scripts directly in .tex documents, but from my brief testing I haven't quite managed to make it work.

498. Briefly: Drawing NMR pulse sequences using APSEQ and Mathematica

There are two things that have been bothering me on linux -- one is the difficulty of creating annotated PDF documents, but it seems like things are slowly improving (we still don't have a really good way of creating /editable/ annotations). Another is the issue of creating NMR pulse-sequence drawings. To be fair, it doesn't seem like the situation is much better on Windows or OSX.

Anyway, I still don't have a good way of doing this, but at least I have A way of getting it done: APSEQ. The main drawback is that it requires Mathematica, which I don't like much (I get by with octave and maxima) in addition to being proprietary. But it will do for now.

NOTE: I just needed to 'Get It Done' quickly, so I didn't spend any time looking at the following potential alternatives:
http://www.celos.net/comp/pulses/ (metapost -- which should be tex compatible, right)
https://code.google.com/p/nmrtype/ (python code)
http://nmrwiki.org/wiki/index.php?title=NMRPulse_-_pulse_sequence_drawing_extension_for_MediaWiki (the URL is fairly descriptive)
They all seem to depend on latex in some way or another.

There's also http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.187.9656&rep=rep1&type=pdf but I haven't actually found the sources.

Anyway, here's how to get started with APSEQ:

mkdir ~/tmp/apseq -p
cd ~/tmp/apseq
wget https://www.princeton.edu/~nmr/apseq/apsDistribution.tar.gz
tar xvf apsDistribution.tar.gz
cd zip.feb15/
math
Mathematica 7.0 for Linux x86 (64-bit) Copyright 1988-2009 Wolfram Research, Inc. In[1]:= <<nmr.auto part01 loaded part02 loaded part03 loaded part04 loaded mscp loaded nmr.extra loaded nmr.auto loaded In[2]:= draw[pulseSeq[hahn, echo]={channels[1H],pulse[90],delay[t1],pulse[180],delay[t1],delay[t2],acquire[taq]}] Out[2]= hahn_echo.pdf In[3]:=

which gives you something like this:
Spin echo sequence
The figure is output as a pdf, but it's easy enough to open it in inkscape and converting it to an .eps file.

21 August 2013

497. Compiling Wine 1.7 in a chroot on debian

Here's a generic way of building Wine 1.7 which is the new testing branch. And yes, it's the instructions for 1.5.28-1.6 recycled, with a few small amendments.

See here for information about 3D acceleration using libGL/U with Wine: http://verahill.blogspot.com.au/2013/05/429-briefly-wine-libglliubglu-blender.html

Getting started:
If you set up a e.g. chroot to build 1.6 you don't need to set up a new chroot to build 1.7. In that case, skip the set-up step below and instead re-enter your existing chroot like this:

sudo mount -o bind /proc wine32/proc
sudo cp /etc/resolv.conf wine32/etc/resolv.conf
sudo chroot wine32
su sandbox
cd ~/tmp

And skip to 'Building wine'.

Otherwise do this:
Setting up the Chroot
sudo apt-get install debootstrap
mkdir $HOME/tmp/architectures/wine32 -p
cd $HOME/tmp/architectures
sudo debootstrap --arch i386 wheezy $HOME/tmp/architectures/wine32 http://ftp.au.debian.org/debian/
sudo mount -o bind /proc wine32/proc
sudo cp /etc/resolv.conf wine32/etc/resolv.conf
sudo chroot wine32

You're now in the chroot:
apt-get update
apt-get install locales sudo vim
echo 'export LC_ALL="C"'>>/etc/bash.bashrc
echo 'export LANG="C"'>>/etc/bash.bashrc
echo '127.0.0.1 localhost beryllium' >> /etc/hosts
source /etc/bash.bashrc
adduser sandbox
usermod -g sudo sandbox
echo 'Defaults !tty_tickets' >> /etc/sudoers
su sandbox
cd ~/

Replace 'beryllium' with the name your host system (it's just to suppress error messages)

Building Wine
While still in the chroot, continue (the i386 is ok; don't worry about it -- you don't actually need it):

sudo apt-get install libx11-dev:i386 libfreetype6-dev:i386 libxcursor-dev:i386 libxi-dev:i386 libxxf86vm-dev:i386 libxrandr-dev:i386 libxinerama-dev:i386 libxcomposite-dev:i386 libglu-dev:i386 libosmesa-dev:i386 libglu-dev:i386 libosmesa-dev:i386 libdbus-1-dev:i386 libgnutls-dev:i386 libncurses-dev:i386 libsane-dev:i386 libv4l-dev:i386 libgphoto2-2-dev:i386 liblcms2-dev:i386 libgstreamer-plugins-base0.10-dev:i386 libcapi20-dev:i386 libcups2-dev:i386 libfontconfig-dev:i386 libgsm1-dev:i386 libtiff-dev:i386 libpng-dev:i386 libjpeg-dev:i386 libmpg123-dev:i386 libopenal-dev:i386 libldap-dev:i386 libxrender-dev:i386 libxml2-dev:i386 libxslt-dev:i386 libhal-dev:i386 gettext:i386 prelink:i386 bzip2:i386 bison:i386 flex:i386 oss4-dev:i386 checkinstall:i386 ocl-icd-libopencl1:i386 opencl-headers:i386 libasound2-dev:i386 build-essential
mkdir ~/tmp
cd ~/tmp
wget http://prdownloads.sourceforge.net/wine/wine-1.7.0.tar.bz2

tar xvf wine-1.7.0.tar.bz2
cd wine-1.7.0/
./configure
time make -j3
sudo checkinstall --install=no
checkinstall 1.6.2, Copyright 2009 Felipe Eduardo Sanchez Diaz Duran This software is released under the GNU GPL. The package documentation directory ./doc-pak does not exist. Should I create a default set of package docs? [y]: Preparing package documentation...OK Please write a description for the package. End your description with an empty line or EOF. >> wine 1.7.0 >> ***************************************** **** Debian package creation selected *** ***************************************** This package will be built according to these values: 0 - Maintainer: [ root@beryllium ] 1 - Summary: [ wine 1.7.0] 2 - Name: [ wine ] 3 - Version: [ 1.7.0] 4 - Release: [ 1 ] 5 - License: [ GPL ] 6 - Group: [ checkinstall ] 7 - Architecture: [ i386 ] 8 - Source location: [ wine-1.7.0 ] 9 - Alternate source location: [ ] 10 - Requires: [ ] 11 - Provides: [ wine ] 12 - Conflicts: [ ] 13 - Replaces: [ ]
Checkinstall takes a little while (In particular this step: 'Copying files to the temporary directory...').
********************************************************************** Done. The new package has been saved to /home/sandbox/tmp/wine-1.7.0/wine_1.7.0-1_i386.deb You can install it in your system anytime using: dpkg -i wine_1.7.0-1_i386.deb **********************************************************************


Installing Wine

Exit the chroot
sandbox@beryllium:~/tmp/wine-1.7.0$ exit
exit
root@beryllium:/# exit
exit
me@beryllium:~/tmp/architectures$ 

On your host system
 Enable multiarch* and install ia32-libs, since you've built a proper 32 bit binary:

sudo dpkg --add-architecture i386
sudo apt-get update
sudo apt-get install ia32-libs libgstreamer-plugins-base0.10-0

*At some point I think ia32-libs may be replaced by proper multiarch packages, but maybe not. So we're kind of doing both here.

 Copy the .deb package and install it
sudo cp wine32/home/sandbox/tmp/wine-1.7.0/wine_1.7.0-1_i386.deb .
sudo chown $USER wine_1.7.0-1_i386.deb
sudo dpkg -i wine_1.7.0-1_i386.deb

20 August 2013

496. Briefly: New email -- adding to PGP/GPG key and changing email account order in Thunderbird

I've created a new email address to move my email correspondence away from google as much as practically possible.

gpg
I've already got a gpg key pair, but I'd like to add the new email address to it.

First find out what you key id is, then add another user id:
 gpg --list-secret-keys
/home/me/.gnupg/secring.gpg ----------------------------- sec 1239G/F8F8FF8F 2011-11-11 uid me <aaa bbb.com> uid lindqvist <ccc ddd> uid verahill <eee fff.com> ssb 1239G/G8GG888G 2011-11-11
gpgp --edit-key F8F8FF8F
Secret key is available. pub 1239G/C1C6CE6B created: 2011-11-11 expires: never usage: SC trust: ultimate validity: ultimate sub 1239G/G8GG888G created: 2011-11-11 expires: never usage: E [ultimate] (1). me <aaa@bbb.com> [ultimate] (2) me <ccc@ddd.com> [ultimate] (3) lindqvist <eee@fff.com> [ultimate] (4) Verahill <ggg@hhh.com> gpg> adduid Real name: Linuxuser Email address: iii@jjj.fr Comment: 20/8/2013 You selected this USER-ID: "Linuxuser (20/8/2013) <iii@jjj.fr>" Change (N)ame, (C)omment, (E)mail or (O)kay/(Q)uit? O You need a passphrase to unlock the secret key for user: "Linuxuser <iii@jjj.com>" 2048-bit RSA key, ID C1C6CE6B, created 2011-11-11 pub 1239G/C1C6CE6B created: 2011-11-11 expires: never usage: SC trust: ultimate validity: ultimate sub 1239G/G8GG888G created: 2011-11-11 expires: never usage: E [ultimate] (1). me <aaa@bbb.com> [ultimate] (2) me <ccc@ddd.com> [ultimate] (3) lindqvist <eee@fff.com> [ultimate] (4) Verahill <ggg@hhh.com> [ unknown] (5) Linuxuser (20/8/2013) <iii@jjj.fr> gpg> quit Save changes? (y/N) Y

Thunderbird
Simple, yet more complicated that it needs to be: http://sidvind.com/wiki/Thunderbird/Change_account_order

Go to Edit, Preferences, Advanced, General, Config Editor.


Search for mail.accountmanager.accounts, and edit the order of the accounts.

Note that the default account (mail.accountmanager.defaultaccount) will always show up first, regardless of the order you set. Restart thunderbird and the changes should take effect.

17 August 2013

495. Briefly: gromacs 4.6 on ROCKS 5.4.3

I didn't want to spend much time on getting this right, so I took the easiest route and combined three posts:

Firstly, I compiled cmake:
http://verahill.blogspot.com.au/2012/05/compiling-openbabel-231-and-cmake-on.html

Secondly, I used the openblas libraries which I compiled in this post
http://verahill.blogspot.com.au/2013/05/421-nwchem-63-on-rocks-543centos-56.html

Thirdly, I looked at this post which deals with gromacs 4.6 on debian:
http://verahill.blogspot.com.au/2013/04/396-compiling-gromacs-46-with-openblas.html

gromacs 4.6 can download and build its own fftw libs, so you don't need to do that separately.

First make the target directory, e.g.
sudo mkdir /share/apps/gromacs
sudo chown $USER:$USER /share/apps/gromacs

Single precision:
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/openmpi/lib:/share/apps/openblas/lib export LDFLAGS="-L/share/apps/openblas/lib -lopenblas" export CPPFLAGS="-I/share/apps/openblas/include" export CC=/usr/bin/gcc44 export CXX=/usr/bin/g++44 cmake -DGMX_FFT_LIBRARY=fftw3 -DGMX_BUILD_OWN_FFTW=On -DGMX_DOUBLE=off -DCMAKE_INSTALL_PREFIX=/share/apps/gromacs/gromacs4.6_single -DGMX_EXTERNAL_BLAS=/share/apps/openblas/lib ../gromacs-4.6 make make install
Double precision:
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/openmpi/lib:/share/apps/openblas/lib export LDFLAGS="-L/share/apps/openblas/lib -lopenblas" export CPPFLAGS="-I/share/apps/openblas/include" export CC=/usr/bin/gcc44 export CXX=/usr/bin/g++44 cmake -DGMX_FFT_LIBRARY=fftw3 -DGMX_BUILD_OWN_FFTW=On -DGMX_DOUBLE=on -DCMAKE_INSTALL_PREFIX=/share/apps/gromacs/gromacs4.6_double -DGMX_EXTERNAL_BLAS=/share/apps/openblas/lib ../gromacs-4.6 make make install
In my particular case I've got all users as members of the compchem group:
chown $USER:compchem /share/apps/gromacs -R 
chmod g+rwx /share/apps/gromacs -R

494. Very briefly: issue with thunderbird, failed connections and repeatedly being asked for the password

Even if you've set up thunderbird so that everything is working perfectly most of the time you occasionally end up in a very annoying situation: you keep on getting error messages about failed connections, and you keep being asked to supply the correct password. You can hit retry, or you can enter the correct password -- but why would you, given that you know based on previous successful connection attempts that your password is correct?
Leaving things along for 5-10 minutes tends to resolve it (or restarting thunderbird, given that you don't do it too quickly i.e. you basically let things stay quiet for a little while).

Anyway, while there are a number of potential reasons for this, in my case it's always been due to too many simultaneous connections, so that some of them are rejected. Not that that's what the error message really says, but whatever, the solution is actually pretty simple -- limit the number of simultaneous connections.

Go to Edit, Account Settings, select the account (typically a gmail one), Server Settings, click on Advanced, and change the 'Maximum number of server connections to cache' from the default 5 to e.g. 1.

On the computers I've done this on I've completely gotten rid of the annoying password requests.

14 August 2013

493. Very, Very Briefly: libreoffice 4 on wheezy -- wheezy-backports

I might be the last person who 'gets' it, but I always presumed that since package versions in backports would be newer they'd automatically be installed.  But apparently that's not the case.

So, enable wheezy-backports, e.g. put this in your /etc/apt/sources.list
deb http://ftp.iinet.net.au/debian/debian/ wheezy-backports main contrib non-free

Then install libreoffice 4:
sudo apt-get install libreoffice -t=wheezy-backports

That's it.

Post-script:
To see what your policies are:
apt-cache policy libreoffice
libreoffice: Installed: 1:4.0.3-2~bpo70+1 Candidate: 1:4.0.3-2~bpo70+1 Version table: *** 1:4.0.3-2~bpo70+1 0 100 http://ftp.iinet.net.au/debian/debian/ wheezy-backports/main amd64 Packages 100 /var/lib/dpkg/status 1:3.5.4+dfsg2-0+deb7u2 0 500 http://ftp.iinet.net.au/debian/debian/ wheezy/main amd64 Packages
and as it turns out the backports repo only has a score of 100, and looking at man apt_preferences:

       If the target release has been specified then APT uses the following algorithm to set the priorities of the versions of a package. Assign:

       priority 1
           to the versions coming from archives which in their Release files are marked as "NotAutomatic: yes" but not as "ButAutomaticUpgrades: yes" like the
           Debian experimental archive.

       priority 100
           to the version that is already installed (if any) and to the versions coming from archives which in their Release files are marked as "NotAutomatic:
           yes" and "ButAutomaticUpgrades: yes" like the Debian backports archive since squeeze-backports.

       priority 500
           to the versions that are not installed and do not belong to the target release.

       priority 990
           to the versions that are not installed and belong to the target release.


To see what's available in backports, do
aptitude search ~Awheezy-backports

12 August 2013

492. Briefly: fixing no sound when playing youtube videos in chrome, iceweasel/firefox

Not sure what caused this, but I suddenly had no sound when playing flash videos in the browser.

Playing flv files using e.g. mplayer works fine. I'm using pulseaudio, and have an ~/.asoundrc file:
pcm.!default.type pulse
ctl.!default.type pulse

alsamixer is working fine and the channels are unmuted.

The solution:
I moved ~/.macromedia to ~/.xmacromedia as suggested at the end of this post: https://bbs.archlinux.org/viewtopic.php?id=134892.

Didn't even need to restart the browser -- suddenly everything worked.

The question is why is there a ~/.macromedia folder, and what does it do? Might have a look into that at some later point.

07 August 2013

491. Compiling ECCE v7.0 (on debian)

See here for more information about what's new:
http://ecce.pnl.gov/using/releasenotes.shtml

Building ECCE 7.0 is done in the same way as previous versions:
http://verahill.blogspot.com.au/2013/01/325-compiling-ecce-64-on-debian-testing.html

While ECCE is distributed as binaries as well, only a 32 bit version is available. To run on 32 bit linux you need to compile your own copy.

See here for Fedora 18 and here for Scientific Linux 6.3. Note that you no longer need to fix the GetOpts issue -- that's already fixed in the new version.

To add new basis sets, read scripts/gbs.README in the ecce source directory, or in apps/scripts/gbs.REAME in the installation directory.

Anyway, building on debian wheezy is simple:
First go to http://ecce.pnl.gov/using/download.shtml and give
sudo apt-get install bzip2 build-essential autoconf libtool ant pkg-config gtk+-2.0-dev libxt-dev csh gfortran openjdk-6-jdk python-dev libjpeg-dev imagemagick xterm
mkdir ~/tmp
cd ~/tmp
tar xvf ecce-v7.0-src.tar.bz2
cd ecce-v7.0/
export ECCE_HOME=`pwd`
cd build/
./build_ecce
Checking prerequisites for building ECCE... If any of the following tools aren't found or aren't the right version, hit -c at the prompt and either find or install the tool before re-running this script. The whereis command is useful for finding tools not in your path. Found gcc in: /usr/bin/gcc ECCE requires gcc 3.2.x or 4.x.x This version: gcc (Debian 4.7.2-5) 4.7.2 [..]. Found java in: /usr/bin/java ECCE requires java 1.5.x or 1.6.x This version: java version "1.7.0_25" Hit return if this java is OK... Found javac in: /usr/bin/javac ECCE requires javac to be the same version as java above This version: javac 1.7.0_25 [..] Found xterm in: /usr/bin/xterm Hit return if xterm was found... The /home/verahill/tmp/ecce/ecce-v7.0/scripts/sysdir script identifies the build platform directory as: Because this value is no platform-specific parent directory will be created for ECCE executables, libraries, etc. This works fine unless your site needs support for multiple platforms. Finished checking prerequisites for building ECCE. Do you want to skip these checks for future build_ecce invocations (y/n)? Y

If you're having issues with e.g. mismatching java and javac versions, use e.g.
sudo update-alternatives --config javac 
to correct it.

If you hit Y at the last question, when you re-run ./build_ecce you'll then start building:
./build_ecce
Xerces built
./build_ecce
Mesa OpenGL built
./build_ecce
wxWidgets built
./build_ecce
running build_ext wxPython built
./build_ecce
Apache HTTP server built
./build_ecce
Copying JMS server distribution jms_server.tar.bz2 Making server tar file ecce_server.tar Copying WebHelp tar file eccewebhelp.tar.bz2 Making combined tar file ecce.v7.0.tar Copying NWChem distribution Concatenating install script and combined tar file ecce.v7.0.tar create_ecce_bin finished ECCE built and distribution created in /home/verahill/tmp/ecce-v7.0
You can now install ~/tmp/ecce-v7.0/install_ecce.v7.0.csh as shown in this post: http://verahill.blogspot.com.au/2013/08/487-version-70-of-ecce-out-now.html