https://norma.mbg.duth.gr/ 2024-12-31T10:57:05+02:00 https://norma.mbg.duth.gr/ https://norma.mbg.duth.gr/lib/images/favicon.ico text/html 2012-02-29T18:21:23+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:amd8coregtx570&rev=1330532483&do=diff 8-core AMD FX-8150 + GTX 570 [The new computing machine] New box with an 8-core AMD FX-8150, an NVIDIA GTX 570 and 4 Gbytes of memory at 1333 MHz arrived. After a few problems with perceus (rejected hardware address starting from 8C:...), all else appear to have worked smoothly. Started testing it with NAMD and comparing it with the i7-extreme-GTX295 machine (all measurements in nanoseconds per day) ... text/html 2009-05-12T11:21:15+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:cuda_flops&rev=1242116475&do=diff CUDA, Flops The hardware is: # ./deviceQuery There is 1 device supporting CUDA Device 0: "GeForce GTX 260" Major revision number: 1 Minor revision number: 3 Total amount of global memory: 939196416 bytes Number of multiprocessors: 24 Number of cores: 192 Total amount of constant memory: 65536 bytes Total amount of shared memory per block: 16384… text/html 2009-04-29T15:24:23+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:cuda_memory_bandwidth&rev=1241007863&do=diff Memory bandwidth for CUDA # ./deviceQuery There is 1 device supporting CUDA Device 0: "GeForce GTX 260" Major revision number: 1 Minor revision number: 3 Total amount of global memory: 939196416 bytes Number of multiprocessors: 24 Number of cores: 192 Total amount of constant memory: 65536 bytes Total amount of shared memory per block: 16384 byte… text/html 2009-02-04T19:22:51+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:hplinpack&rev=1233768171&do=diff HPL 2.0 Coming soon ... text/html 2013-09-11T20:46:44+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:ibmx3755m3&rev=1378921604&do=diff IBM System x3755 M3 No CUDA, NAMD multicore 2.9 1,632 atom system Cores Nanoseconds per day 1 22 4 69 8 99 … text/html 2009-02-04T19:14:36+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:linpack&rev=1233767676&do=diff Linpack benchmarks These benchmarks are based on Intel's Optimized LINPACK Benchmark as distributed with Intel's mkl libraries v.10.1.1.019. On a single node and using the defaults, the xlinpack_xeon64 executable gives: The log file from the run is ... text/html 2009-04-16T21:56:30+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:memory_bandwidth&rev=1239908190&do=diff Memory bandwidth Using the program cachebench the results are: # ./cachebench Memory Set Library Cache Test C Size Nanosec MB/sec % Chnge ------- ------- ------- ------- 256 0.05 20205.70 1.00 256 0.05 20210.63 1.00 384 0.05 20445.46 0.99 384 0.05 20447.49 1.00 51… text/html 2009-02-06T14:04:19+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namd1.6k&rev=1233921859&do=diff NAMD, pentapeptide benchmarks The solute is a peptapeptide. The system comprises 1632 atoms and the simulation involves full electrostatics on a 32x32x27 grid(see included script below). NAMD script used for these tests # # Input files # structure ionized.psf coordinates heat_out.coor velocities heat_out.vel extendedSystem heat_out.xsc parameters par_all27_prot_na.inp paraTypeCharmm on # # Output files & writing frequ… text/html 2009-02-06T13:56:00+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namd100k&rev=1233921360&do=diff NAMD, 100,000 atoms benchmarks This is a 99,744 atom system with a PME grid of 112x108x108 (script included below). For all the tests that follow we used the NAMD 2.6 amd64 executable as provided by the NAMD developers. NAMD script used for these tests text/html 2009-06-30T11:34:27+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namd27b1_further_tests&rev=1246350867&do=diff Some more tests with NAMD v.2.7b1 More measurements (in ns/day) concerning timesteps and friends for a couple of small systems: Number of atoms - Cores 13.3K - 4 timestep 2, nonBondedFreq 2, fullElectFrequency 4 7.26 timestep 2, nonBondedFreq 2, fullElectFrequency 2 7.28 timestep 2, nonBondedFreq 1, fullElectFrequency 2 4.40 timestep 1, nonBondedFreq 2, fullElectFrequency 2 3.86 text/html 2009-02-06T14:01:44+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namd60k&rev=1233921704&do=diff NAMD, 60,000 atoms benchmarks The system used for the test comprised 60660 atoms, with an orthogonal PBC box of dimensions ~124x77x63, an inner step of 2 fs, nonbonded every 4 fs, and electrostatics every 8 fs (script include below). All runs were allowed to run till stabilisation. The table below reports days per nanosecond of simulation for each combination indicated. text/html 2012-03-08T11:10:02+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namdlatestcuda&rev=1331197802&do=diff NAMD-CUDA, latest benchmarks on small (tiny) systems All simulations in explicit solvent (TIP3P), full PME electrostatics, 2-1-2 steps, and cutoffs ranging from 8 to 10 Angstrom. Number of atoms Hardware (single node) nanoseconds per day days per μs Adaptive tempering ? 1997 Q6600 - GTX460 75 14 … text/html 2009-09-09T18:06:37+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namdv27b1&rev=1252508797&do=diff NAMD v.2.7b1 For details of the systems used, see the corresponding pages for NAMD v.2.6, or view the scripts included at the end of this page. The ApoA1 benchmark is as distributed by the NAMD developers. If you'd rather prefer cutting a long story short, see the timings (in ns/day) shown on the very last table of this page. text/html 2010-05-01T18:10:10+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namdv27b1_i7&rev=1272726610&do=diff NAMD 2.7b1 on an i7 965 We compare eight threads vs. four threads vs. the performance obtained from a Q6600-based quad. All measurements are in nanoseconds per day and were made with an SMP-aware executable of NAMD 2.7b1 using the flags '+setcpuaffinity +LBSameCpus'. Details of the examples used can be found here. text/html 2012-02-13T19:56:16+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namdv27cuda&rev=1329155776&do=diff NAMD 2.8b1 with and without CUDA on an Intel i7 extreme with an Nvidia GTX 295 card For the non-CUDA runs eight threads were used. For the CUDA runs, four threads always gave better performance and were used throughout. The non-CUDA runs are based on an SMP-aware executable of NAMD 2.7b1 using the flags '+setcpuaffinity +LBSameCpus'. The CUDA runs are based on the 2.8b1 version of NAMD (obtained directly from the NAMD site). Details of the examples used can be found here (keeping in mind that… text/html 2012-02-13T19:54:32+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:namdv28cudagtx460&rev=1329155672&do=diff NAMD 2.8 with and without CUDA on Q6600 quads with nvidia's GTX460 cards ApoA1 The ApoA1 benchmark is as distributed by NAMD developers. The tests were performed on identical machines based on intel's Q6600 with nvidia's GTX460 on a gigabit interconnect. All measurements are in nanoseconds per day. text/html 2009-02-16T18:37:21+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:network&rev=1234802241&do=diff Network benchmarks The network benchmarks are based on NetPIPE. Each node comes with two gigabit network interfaces, which unfortunately, have very different performances, although they are both based on RealTek chipsets. The graphs below compare the on-board NIC (RTL8168B, green line) with the PCI NIC (RTL8169sb, orange line) with respect to throughput vs. message size. The PCI NIC has significantly lower throughput with messages longer than 64K, but appears to perform better with shorter mes… text/html 2009-06-10T18:02:52+02:00 https://norma.mbg.duth.gr/index.php?id=about:benchmarks:qsmpi&rev=1244646172&do=diff Qs v.1.3, MPI version Space group P6(1)22, 3166 reflections, 50K steps, average of three runs Number of cores Wall-clock time in seconds Parallel efficiency (%) 1 334 - 2 164 102 4 103 81 8 80 52 …