Re: [Wien] Intel(R) Xeon Phi™ coprocessor
This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM, Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at wrote: Before buying a couple of new computers, I was asking myself the same question and discussed this with some people of our computing departments. The conclusions: a) potentially very good, but in practice very questionable, because for most application you cannot get out the real speed (10 times faster than an Intel I7). This is true, even for many lapack/mkl subroutines where it should work better. They told me to wait, until the mkl becomes better (hopefully). I'm not too optimistic, when you see how badly the mkl-parallelization of multicore machines is working (2 cores is very good, but 4 or more is already very bad). b) The nature of our problem (big eigenvalue problem): A fast processor is useful only for large problems -- large memory. You can buy Phi coprocessors now with quite some large memory, but then they are terrible expensive (and 5 normal PCs are faster and cheaper) c) the hardware design has a VERY slow communication between main-memory and Phi-memory. This makes also parallelization over several PHI-nodes via mpi not really possible (if you need any significant data transfer, like for an eigenvalue problem). Thus I did not buy it. However, if anybody has access and time to try out WIEN2k on PHis, I'd would be very interested in getting feedback. (Maybe these computer-people were not good enough )
Re: [Wien] Intel(R) Xeon Phi™ coprocessor
Dear Prof. Blaha, Thank you very much for the explanations. They will be very useful !! All the best, Luis 2013/9/12 Peter Blaha pbl...@theochem.tuwien.ac.at This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pblaha@theochem.**tuwien.ac.at pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM, Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pblaha@theochem.**tuwien.ac.atpbl...@theochem.tuwien.ac.at wrote: Before buying a couple of new computers, I was asking myself the same question and discussed this with some people of our computing departments. The conclusions: a) potentially very good, but in practice very questionable, because for most application you cannot get out the real speed (10 times faster than an Intel I7). This is true, even for many lapack/mkl subroutines where it should work better. They told me to wait, until the mkl becomes better (hopefully). I'm not too optimistic, when you see how badly the mkl-parallelization of multicore machines is working (2 cores is very good, but 4 or more is already very bad). b) The nature of our problem (big eigenvalue problem): A fast processor is useful only for large problems -- large memory. You can buy Phi coprocessors now with quite some large memory, but then they are terrible expensive (and 5 normal PCs are faster and cheaper) c) the hardware design has a VERY slow communication between main-memory and Phi-memory. This makes also parallelization over several PHI-nodes via mpi not really possible (if you need any significant data
Re: [Wien] Intel(R) Xeon Phi™ coprocessor
While I'm not sure how easily normal desktop benchmarks transfer to parallel processing through wien2k I usually look at the following benchmarks when comparing CPUs (and prices): http://cpubenchmark.net/high_end_cpus.html Best regards, Michael Sluydts Op 12/09/2013 15:43, Luis Ogando schreef: Dear Prof. Blaha, Thank you very much for the explanations. They will be very useful !! All the best, Luis 2013/9/12 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM, Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at wrote: Before buying a couple of new computers, I was asking myself the same question and discussed this with some people of our computing departments. The conclusions: a) potentially very good, but in practice very questionable, because for most application you cannot get out the real speed (10 times faster than an Intel I7). This is true, even for many lapack/mkl subroutines where it should work better. They told me to wait, until the mkl becomes better
[Wien] L -4
Dear WIEN2k team, In case.clmsum and case.vcoul one finds designations like CLM(R) FOR L 4 M= 4 CLM(R) FOR L -4 M= 4 Does the second line mean CLM(R) FOR L 4 M= -4 ? John Rundgren / KTH ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Intel(R) Xeon Phi™ coprocessor
Thank you, Michael ! This will be useful too ! Nevertheless, we can not forget that, sometime ago, Prof. Blahacommented that the Cache Memory is more important than the clock for WIEN2K (discarding extreme cases, of course). All the best, Luis 2013/9/12 Michael Sluydts michael.sluy...@ugent.be While I'm not sure how easily normal desktop benchmarks transfer to parallel processing through wien2k I usually look at the following benchmarks when comparing CPUs (and prices): http://cpubenchmark.net/high_end_cpus.html Best regards, Michael Sluydts Op 12/09/2013 15:43, Luis Ogando schreef: Dear Prof. Blaha, Thank you very much for the explanations. They will be very useful !! All the best, Luis 2013/9/12 Peter Blaha pbl...@theochem.tuwien.ac.at This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM, Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at wrote: Before buying a couple of new computers, I was asking myself the same question and discussed this with some people of our computing departments. The conclusions: a) potentially very good, but in practice very questionable, because for most application you cannot get out the real speed (10 times faster than an Intel I7). This is true, even for many lapack/mkl subroutines where it should work better. They told me to wait, until the mkl becomes better (hopefully). I'm not too optimistic, when you see how badly the mkl-parallelization of multicore machines is working (2 cores is very good, but 4 or more is already very bad). b) The nature of our problem (big
Re: [Wien] L -4
yes On Thu, 12 Sep 2013, John Rundgren wrote: Dear WIEN2k team, In case.clmsum and case.vcoul one finds designations like CLM(R) FOR L 4 M= 4 CLM(R) FOR L -4 M= 4 Does the second line mean CLM(R) FOR L 4 M= -4 ? John Rundgren / KTH ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] L -4
No. We expand the density in real spherical harmonics, which are linear combinations ofY_44 +/- Y_4-4 and these two combinations are labeled that way. There are papers by Kurki-Suonio, or see the UG (lapw2). On 09/12/2013 04:05 PM, John Rundgren wrote: Dear WIEN2k team, In case.clmsum and case.vcoul one finds designations like CLM(R) FOR L 4 M= 4 CLM(R) FOR L -4 M= 4 Does the second line mean CLM(R) FOR L 4 M= -4 ? John Rundgren / KTH ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- P.Blaha -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWWW: http://info.tuwien.ac.at/theochem/ -- ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Intel(R) Xeon Phi™ coprocessor
I will add that I have been told that with the latest Xeon (not sure about i7), the memory speed can also matter. One additional option, the one I now use, is to have a vendor put together a small cluster for you. At least in the US this seems to be competitive as they get a discount on the price. My procedure if they have options is to benchmark lapw1 on a test machine (ideally machines and mpi) to compare options as well as different vendors. At least at the higher Xeon end I cannot invest the time to learn all the (constantly changing) hardware details and prefer to leave that to experts. I also get technical support included without extra cost, which can be very important. As one example, the list Michael just sent does not include the latest v2 models (too new). I will shortly be benchmarking some of these, they look cheaper but On Thu, Sep 12, 2013 at 9:06 AM, Luis Ogando lcoda...@gmail.com wrote: Thank you, Michael ! This will be useful too ! Nevertheless, we can not forget that, sometime ago, Prof. Blaha commented that the Cache Memory is more important than the clock for WIEN2K (discarding extreme cases, of course). All the best, Luis 2013/9/12 Michael Sluydts michael.sluy...@ugent.be While I'm not sure how easily normal desktop benchmarks transfer to parallel processing through wien2k I usually look at the following benchmarks when comparing CPUs (and prices): http://cpubenchmark.net/high_end_cpus.html Best regards, Michael Sluydts Op 12/09/2013 15:43, Luis Ogando schreef: Dear Prof. Blaha, Thank you very much for the explanations. They will be very useful !! All the best, Luis 2013/9/12 Peter Blaha pbl...@theochem.tuwien.ac.at This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM, Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at wrote:
Re: [Wien] Intel(R) Xeon Phi™ coprocessor
Nice page, although of course what for wien2k matters is the performance for lapack/blas calls for a diagonalization/matrix-matrix multiplication. Anyway, you can see that Intel Core i7-3930K @ 3.20GHz is pretty high ranked and costs(560$) just a quarter of the top Intel Xeon E5-2687W @ 3.10GHz (1930$), while its performance is only 15% less than the top value. And of course, this may change every couple of month On 09/12/2013 03:46 PM, Michael Sluydts wrote: While I'm not sure how easily normal desktop benchmarks transfer to parallel processing through wien2k I usually look at the following benchmarks when comparing CPUs (and prices): http://cpubenchmark.net/high_end_cpus.html Best regards, Michael Sluydts Op 12/09/2013 15:43, Luis Ogando schreef: Dear Prof. Blaha, Thank you very much for the explanations. They will be very useful !! All the best, Luis 2013/9/12 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM, Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at wrote: Before buying a couple of new computers, I was asking myself the same question and discussed this with some people of our computing departments. The
[Wien] Regarding Space group
Dear All I am facing problem when i am calculating the SCF for the compound of Monoclinic symmetry having space group # 12 (C2/c) and this # space group is different, which is B2/m in space group list. So please some one help that what should i do with this type of structure. I am waiting for reply, thanks in advance. with regards sikander ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Regarding Space group
I think you have a typo and mean #12 (C2/m), as C2/c is #15. You should be able to convert it with SETSTRU. The old post links below might be insightful: http://zeus.theochem.tuwien.ac.at/pipermail/wien/2013-January/018177.html http://zeus.theochem.tuwien.ac.at/pipermail/wien/2013-January/018156.html On 9/12/2013 1:00 PM, sikandar azam wrote: Dear All I am facing problem when i am calculating the SCF for the compound of Monoclinic symmetry having space group # 12 (C2/c) and this # space group is different, which is B2/m in space group list. So please some one help that what should i do with this type of structure. I am waiting for reply, thanks in advance. with regards sikander ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Intel(R) Xeon Phi™ coprocessor
Should probably also note with that page that the high end AMD CPUs are typically much cheaper, but they tend to have double the cores with nearly the same power usage per core (so double the power) compared to an i7 or xeon, at continuous operation this can lead to a painful power bill, possibly aggrevated by heat generation requiring more active cooling (though I don't remember data on the temperatures). Op 12/09/2013 17:33, Peter Blaha schreef: Nice page, although of course what for wien2k matters is the performance for lapack/blas calls for a diagonalization/matrix-matrix multiplication. Anyway, you can see that Intel Core i7-3930K @ 3.20GHz is pretty high ranked and costs(560$) just a quarter of the top Intel Xeon E5-2687W @ 3.10GHz (1930$), while its performance is only 15% less than the top value. And of course, this may change every couple of month On 09/12/2013 03:46 PM, Michael Sluydts wrote: While I'm not sure how easily normal desktop benchmarks transfer to parallel processing through wien2k I usually look at the following benchmarks when comparing CPUs (and prices): http://cpubenchmark.net/high_end_cpus.html Best regards, Michael Sluydts Op 12/09/2013 15:43, Luis Ogando schreef: Dear Prof. Blaha, Thank you very much for the explanations. They will be very useful !! All the best, Luis 2013/9/12 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at This depends a lot on what you want to do and how much money you have. The single-core speed of a fast I7 is at least as good (or faster) than most Xeons, and they are MUCH cheaper. So for all systems up to 64-100 atoms/cell, where you need several k-points, a small cluster of I7 cpus /4 cores, or more expensive 6 core) (with GB-network and a common NFS file system) is for sure the fastest platform and in particular has by FAR the best price/performance ratio (One powerful I7 computer may cost about 1000 Euro). For bigger clusters, a drawback can be the large space to put all PCs on a big shelf ), but if you have less than 1 Euros, this is probably the best choice. However, Xeons can be coupled (2-4 Xeons) to a single multicore computer (eg. 16 cores), which may work with mpi and can be used to handle systems up to 200-300 atoms. They also can be bought in small boxes and may fit in a single 19 inch cabinet. But of course such systems are much more expensive. From what I said above it should be clear, that it is completely useless to buy a single 4 core Xeon computer. The next step would be to buy an Infiniband switch+cards and couple your PCs with this fast network to a powerful multinode mpi-cluster. Since the switch/cards are fairly expensive, on usually takes here Xeons as platform. However, you need to know how to install/configure the software properly. I've seen such clusters even in computing centers, which were completely useless, because the network/mpi was instable and jobs would crash randomly every couple of hours . Our strategy: i) We have a GB-networked cluster with Intel I7 computers (which we maintain our-self and this cluster includes also all the user- workstations) and do all the calculations for systems up to 64 atoms/cell on these systems. 2) For bigger systems we go to our University computer-center and run there with a PBS queuing system. This has the advantage that we do not need to care about the installation of the infiniband network nor the mpi-infrastructure (but we use always intel-mpi together with ifort/mkl). On 09/11/2013 06:31 PM, Luis Ogando wrote: Dear Prof. Blaha, Just for curiosity, what processor did you buy ? Is the Xeon family better than the i7 one for WIEN2k calculations ? All the best, Luis 2013/9/11 Peter Blaha pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at mailto:pbl...@theochem.tuwien.ac.at I don't know what latest means. We use the latest one installed on our supercomputers (4.1.1.036) I have not seen any significant change with mpi in the last years. PS: I just got info that we have now a new ifort available for download ... On 09/11/2013 05:00 PM, Laurence Marks wrote: Thanks. One thing I will add/ask concerning the parallelization, the latest impi seems to be substantially better -- have you tried it? I have not just noticed this with Wien2k, but I am told that others have seen improvements in other codes. On Wed, Sep 11, 2013 at 9:42 AM,
Re: [Wien] L -4
Ah yes. I answered too fast. On Thu, 12 Sep 2013, Peter Blaha wrote: No. We expand the density in real spherical harmonics, which are linear combinations ofY_44 +/- Y_4-4 and these two combinations are labeled that way. There are papers by Kurki-Suonio, or see the UG (lapw2). On 09/12/2013 04:05 PM, John Rundgren wrote: Dear WIEN2k team, In case.clmsum and case.vcoul one finds designations like CLM(R) FOR L 4 M= 4 CLM(R) FOR L -4 M= 4 Does the second line mean CLM(R) FOR L 4 M= -4 ? John Rundgren / KTH ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- P.Blaha -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWWW: http://info.tuwien.ac.at/theochem/ -- ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html