From owner-chemistry@ccl.net Mon Feb 25 00:10:01 2008 From: "Darryl Reid darryl|,|simbiosys.ca" To: CCL Subject: CCL: Computing on graphical processors Message-Id: <-36370-080224213108-2477-1dsOW4BgeyDb/slBg8DKHA===server.ccl.net> X-Original-From: Darryl Reid Content-Disposition: inline Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="iso-8859-15" Date: Sun, 24 Feb 2008 21:31:16 -0500 MIME-Version: 1.0 Sent to CCL by: Darryl Reid [darryl(a)simbiosys.ca] Thanks for your comments Jerome, I have duplicated my response (and your question) on our blog (www.simbiosys,ca/blog) as I am sure others are interested in this issue. We, personally, have not done any tests with QM codes on the Cell, however, currently the IBM QS21 already comes with 2GB, specs: http://www-03.ibm.com/systems/bladecenter/hardware/servers/qs21/ While the QS22 will have up to 16GB . Another option is the mercury Cell Accelerator Board which has 1GB XDR + 4GB DDR2 RAM: http://www.mc.com/microsites/cell/productdetails.aspx?id=2590 As I said, we haven't done any tests of QM, but I feel the newer implementations will have the memory to handle these codes. Now learning to program on them in an optimal fashion is a different question. Darryl On Friday 22 February 2008 6:45 am, Jerome Kieffer jerome.Kieffer * terre-adelie.org wrote: > Sent to CCL by: Jerome Kieffer [jerome.Kieffer-x-terre-adelie.org] > > Nice paper ... > > Have you made some tests on QM codes ? because 1GB of memory is really > tinny for QM. > > Best regards. -- Darryl Reid Application Scientist SimBioSys Inc. "Your Software Partner for Rational Drug Design" http://www.simbiosys.ca Tel: 1-416-741-4263 From owner-chemistry@ccl.net Mon Feb 25 10:12:01 2008 From: "Rebecca Claire Wade rebecca.wade||eml-r.villa-bosch.de" To: CCL Subject: CCL: EMBO Practical Course on Biomolecular Simulation - July 1-8,2008 Message-Id: <-36371-080225101013-10461-vUdw6sPzBMXnB/hh0rJP5Q__server.ccl.net> X-Original-From: "Rebecca Claire Wade" Date: Mon, 25 Feb 2008 10:10:08 -0500 Sent to CCL by: "Rebecca Claire Wade" [rebecca.wade---eml-r.villa-bosch.de] The 4th EMBO Practical Course on Biomolecular Simulation will take place from July 1-8 2008 at the Pasteur Institute in Paris, France. The course topics will include: Molecular and Brownian dynamics simulation, Monte Carlo techniques, electrostatics, free energy calculations, QM/MM, homology modeling, drug design Course organizers: Michael Nilges (Pasteur Institute, Paris) Rebecca Wade (EML Research, Heidelberg) Course lecturers: Arnaud Blondel Valerie Daggett Konrad Hinsen Leslie Kuhn Richard Lavery Phineus Markwick Alan Mark Adrian Mulholland Tom Simonson Anna Tramontano Registration forms and course details are at: http://cwp.embo.org/pc08-12/ Registration deadline: 10 April 2008 -- Rebecca Wade http://www.eml-r.org/english/homes/wade/ From owner-chemistry@ccl.net Mon Feb 25 11:10:01 2008 From: "Jens Thomas j.m.h.thomas a dl.ac.uk" To: CCL Subject: CCL: CCP1 Meeting - Molecular Modelling: Tools, GUIs and Visualisation Message-Id: <-36372-080225104348-4413-ETc9kVj+VqHwexQEb/3djg---server.ccl.net> X-Original-From: Jens Thomas Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 25 Feb 2008 14:46:37 +0000 MIME-Version: 1.0 Sent to CCL by: Jens Thomas [j.m.h.thomas],[dl.ac.uk] Molecular Modelling: Tools, GUIs and Visualisation 11th - 13th March 2008 This is to announce a CCP1 meeting looking at graphical interfaces, visualisation and tools for molecular modelling in general, but with the focus towards electronic structure. There are a large number of graphical interfaces and associated tools for working with electronic structure and molecular mechanics/dynamics codes in use in the computational chemistry community today. Many of these tools do similar jobs, but all have their own specialities, strengths and weaknesses. The aim of this meeting is to bring together the developers and users of these tools to: - raise awareness of the existing software tools and capabilities. - allow users to give the developers feedback on what is needed and which areas are not being sufficiently covered. - explore novel ways of visualising molecular data and other tasks. - provide an opportunity to discuss collaborations. We already have speakers confirmed from projects including OpenBabel, Avogadro, Molekel, Molpro, cclib, GaussSum, The CCP1GUI, GAMESS, and GAMESS-UK. The meeting will be held at the Holiday Inn in Runcorn, Cheshire, from 11th-13th March (starting after lunch on the 11th and running through to lunch on the 13th). The afternoon of the 13th will be reserved for developers to run tutorials/practicals at Daresbury Laboratory. We still have some slots available for speakers, so if you are a code developer or user, and would be interested in speaking, then please contact: j.m.h.thomas~~dl.ac.uk For registration information, please visit the meeting webpage at: http://www.ccp1.ac.uk/chemtoolsmeet We look forward to seeing you in March! -- =================================================================== Jens Thomas, email: j.m.h.thomas~~dl.ac.uk STFC Daresbury Lab, tel: +44-1925-603849 Warrington, fax: +44-1925-603634 WA4 4AD, UK. http: http://www.cse.scitech.ac.uk =================================================================== From owner-chemistry@ccl.net Mon Feb 25 12:42:01 2008 From: "Ted Cabeen cabeen=-=chem.ucsb.edu" To: CCL Subject: CCL: running SYBYL via a Mac Message-Id: <-36373-080225123458-21388-pn+ZGqAgxEFDqIVJSDK/rA|-|server.ccl.net> X-Original-From: Ted Cabeen Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 25 Feb 2008 09:35:23 -0800 MIME-Version: 1.0 Sent to CCL by: Ted Cabeen [cabeen[]chem.ucsb.edu] Daniil Bratashov dn2010[a]gmail.com wrote: > It depends on unix version installed. You have to set variable DISPLAY > to your_mac_ip:0 (or number of X11 server started). In some shell's > (csh) it is: > setenv DISPLAY macintosh:0 > > in other (bash): > export DISPLAY=macintosh:0 > > You also have to enable X forwarding over ssh, if it is not done > before, and enable connections to X11 from other hosts with xhost > command on your macintosh computer. This is incorrect. There are two ways to export your SYBYL X display to a macintosh X server, one using ssh X11 forwarding, and the other using xhost and manually setting DISPLAY. Here they are in detail: -- SSH X11 Forwarding (preferred) 1) Start the Macintosh X11 server. 2) From the terminal window it creates, ssh to the machine that has SYBYL installed with the -X or -Y command line arguments specified. 3) Once you've connected to the SYBYL machine, run the following to confirm that X11 forwarding is configured: echo $DISPLAY You should get something like "localhost:10.0" in response. If you get a blank line, then X11 forwarding is not working properly. 4) If X11 forwarding is working, then run SYBYL normally. You may need to tell SYBYL to use the XDEV graphics mode, as a previous poster mentioned. *) Note in this method, you do not have to run xhost at all, nor do you set your environment variables. If you do so, it may break the X11 forwarding that the ssh client and server have configured. -- xhost and DISPLAY method (discouraged) This method is the older and more complicated method and should only be used if you cannot use the ssh X11 forwarding method. This method is the only way if you are connecting over telnet, or if you ssh to the machine without the -X or -Y arguments. 1) Start the Macintosh X11 server. 2) In the terminal window it creates run the following: xhost +hostname.of.the.sybyl.machine.example.com (Do not run "xhost +", as this opens a large security hole on your machine. If you run "xhost +", you are allowing anyone on the internet to place X11 windows on your desktop, including fake username and password prompts which could steal your login credentials) 3) Connect to the machine running sybyl using ssh or telnet. If you use ssh, do not include the -X or -Y arguments. 4) On the SYBYL machine, configure the DISPLAY environment variable to point at your desktop's IP address, like this: Bourne-shell derivatives: export DISPLAY=my.ip.address.here:0 csh derivatives: setenv DISPLAY my.ip.address.here:0 5) Once this is all configured, run sybyl normally. Again, You may need to tell SYBYL to use the XDEV graphics mode. If you get to the last step of the process, but sybyl won't start, you can test your X11 configuration by running xterm from the remote machine. If the remote X11 display is working properly, it will open a new terminal window on your local machine. If not, then you know that something is misconfigured. I hope this has been helpful to people, and that you will find it useful in the future. This technique will work for all X11-based programs on a remote machine. --Ted From owner-chemistry@ccl.net Mon Feb 25 13:28:01 2008 From: "Jerome Kieffer jerome.Kieffer^^terre-adelie.org" To: CCL Subject: CCL: Computing on graphical processors Message-Id: <-36374-080225075731-30160-ui+blySKpQ3Q1biC0HaDcQ[A]server.ccl.net> X-Original-From: Jerome Kieffer Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-15 Date: Mon, 25 Feb 2008 13:07:07 +0100 Mime-Version: 1.0 Sent to CCL by: Jerome Kieffer [jerome.Kieffer-,-terre-adelie.org] On Sun, 24 Feb 2008 21:31:16 -0500 "Darryl Reid darryl|,|simbiosys.ca" wrote: >=20 > Sent to CCL by: Darryl Reid [darryl(a)simbiosys.ca] > Thanks for your comments Jerome, I have duplicated my response (and your= =20 > question) on our blog (www.simbiosys.ca/blog) as I am sure others are=20 > interested in this issue. Just another comment about the use of the Cell precessor, it should be especially efficent for crystal structure resolution from powder data as the are very little data and (many) monte-carlo simulations in parallel.=20 Well, as Accelrys does not seems to known the =AB-O=BB option of the compiler, I don't excpect them to port PowderSolve to the cell processor. Maybe somebody at CCDC Cambridge ?? The solution provided by Mercury seems very promising as GPU are not (yet) able to do double precision calculation. Regards. --=20 J=E9r=F4me KIEFFER=20