Re: [gmx-users] Workstation choice
BTW, I'd recommend caution when using the dated d.dppc benchmark for drawing performance conclusions both because it may not be too representative of other workloads (small size, peculiar settings) and because it uses all-bonds constrained with 2fs time step which is not recommended these days (unless using larger time steps). Constraining only h-bonds is generally considered best practice and it is faster, especially with GPUs. I'd also like to warn about drawing conclusions about hypothetical production setups from performance measurements with under-utilized CPUs. Leaving cores empty creates conditions which are not representative to real-world scenarios (unless of course you plan to leave those core empty during production runs too). The least I'd do is to put some CPU load on those cores (e.g. on an 8-core CPU run mdrun on 4 cores + 1 GPU and keep the other 4 cores busy with some dummy load like a stress tester tool or a CPU-only mdrun run). -- Szilárd On Tue, Sep 11, 2018 at 11:13 PM Wahab Mirco < mirco.wa...@chemie.tu-freiberg.de> wrote: > On 11.09.2018 22:37, Olga Selyutina wrote: > >> How would the 8600K perform without GPU (using the "-nb cpu" > >> option on the mdrun binary)? > > 8600K with frequency 4600Mhz was used > > Command line: > > gmx mdrun -nb cpu > > > > Core t (s) Wall t (s)(%) > >Time: 1529.942 254.990 600.0 > > (ns/day)(hour/ns) > > Performance: 169.4190.142 > > Interesting - That means, for your record: > >i5-8600K, 6C/6T, AVX2_512: 169.419 ns/day >R2-2700X, 8C/16T, AVX2_128: 152.728 ns/day >R1-1700X, 8C/16T, AVX2_128: 141.456 ns/day > > your overclocked coffee-lake i5 would already > be slightly faster than a Ryzen2 2700X. > > But if you want to simulate larger membrane systems, > you should possibly test also on larger membrane > systems. There's a DPPC/Water membrane benchmark included > in the Gromacs benchmark set. Unfortunately, it (d.dppc) > wouldn't work out of the box with newer versions. > I patched it up and converted it into a NPT membrane-benchmark: > http://gromacs.deppenfraktur.de/gmxbench-d.dppc-npt.tgz > and already got some results for some of our antique boxes: > > Results for 2018.3: > >*** i5-7600K, 4C/4T, AVX2_512: 5.512 ns/day >i7-6700K, 4C/8T, AVX2_256: 6.858 ns/day >R1-1700X, 8C/16T, AVX2_128: 7.490 ns/day >* i7-5820K, 6C/12T, AVX2_256: 9.162 ns/day >* R-2 2700X, 8C/16T, AVX2_128: 9.401 ns/day > >** GTX-980, i5-7600K, 4C/4T, AVX2_512: 28.071 >** GTX-1080, R1-1700X, 8C/16T, AVX2_128: 42.908 > > If you get some people with contemporary hardware to test for > such workloads related to your projects, you can possibly buy > boxes accordingly. > > > M. > -- > Gromacs Users mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before > posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. > -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
On Tue, Sep 11, 2018 at 5:16 PM Wahab Mirco < mirco.wa...@chemie.tu-freiberg.de> wrote: > On 07.09.2018 20:41, Olga Selyutina wrote: > > ... > > These sets of CPU and GPU are suitable for price (in our region): > > *GPU* > > GTX 1070 ~1700MHz, cuda 1920 - $514 > > GTX 1080 ~1700MHz, cuda 2560 - $615 > > GTX 1070Ti ~1700MHz, cuda 2432 - $615 > > GTX 1080Ti ~1600MHz, cuda 3584 - $930 > > > > *CPU* > > Ryzen 7 2700X - $357 > > 4200MHz, 8/16 cores/threads, cache L1/L2/L3 768KB/4MB/16MB, 105W, max.T > 85C > > > > Threadripper 1950X - $930 > > 4000MHz, 16/32 cores/threads, cache L1/L2/L3 1.5/8/32MB, 180W, max.T 68C > > > > i7 8086K - $515 > > 4800MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C > > > > i7 8700K - $442 > > 4600MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C > > > > The most suitable combinations CPU+GPU are as follows: > > 1) Ryzen 7 2700X + two GTX 1080 - $1587 > > 1.1) Ryzen 7 2700X + one GTX 1080 + one GTX 1080*Ti* - $1900 (maybe?) > > 2) Threadripper 1950X + one GTX 1080Ti - $1860 > > 3) i7 8700K + two GTX 1080 - $1672 > > 4) Ryzen 7 2700X + three GTX 1070 - $1900 > > My suggestions: > > Variant 1 seems to be the most suitable. > > Variant 2 seems to be suitable only if the single simulation is running > on > > workstation > > It’s a bit confusing that in synthetic tests/games performance of i7 8700 > > is higher than Ryzen 7 2700. > > ... > > Sorry for jumping into the thread at this point, but depending on > the problem size and type, it might happen that: > >- a single R2-2700X possibly cannot always saturate TWO > GTX-1080/1080Ti (maybe Szilárd Páll can add some rule of > thumb whether and when it can), > Depends on the definition of "saturate"; for a typical ~100k membrane protein simulation with h-bonds constraints, already four low-end Xeon cores per GTX 1080 are sufficient to get ~90% of the ~ about peak perf (say what you'd get 2-3x more cores/GPU). With a 1080 Ti and ~4 wimpy Xeon cores, that'll be more around 75-80% of the peak perf. (Side-note: that's with a single run/GPU -- for small inputs like Olga's I'd definitely try to have 2 runs per GPU for efficiency.) Ref: https://goo.gl/8DVqPV Now, considering that the 2700X has 8 _fast_ cores, I'd expect it to work fine with two 1080s (and likely with 1080 Ti's too), though it does depend on the amount of CPU work. >- a larger Threadripper may suffer from memory bandwidth issues on > many parallel threads in Gromacs, > The number of cores will outweigh most bandwidth issues. >- due to AVX2_512, 6C/12T "Coffee Lake" i7 may be somehow faster than > even 8C/16T Ryzen2 (at some scenarios significantly?) > Without a GPU it may, with a GPU it will not. As I said earlier, the benefits of AVX512 are mostly gone with the nonbondeds offloaded to a GPU. > But it most probably depends on the circumstances. If you want to > spend your money effectively, you have to perform at least some tests. > > I can provide some performance numbers I got after updating > Gromacs to v.2018.3 lately on our aging lab hardware. > > Regards, > > Mirco > > ==> (I can provide the input file if required) ==> > -- > Gromacs Users mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before > posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
On 11.09.2018 22:37, Olga Selyutina wrote: >> How would the 8600K perform without GPU (using the "-nb cpu" >> option on the mdrun binary)? > 8600K with frequency 4600Mhz was used > Command line: > gmx mdrun -nb cpu > > Core t (s) Wall t (s)(%) >Time: 1529.942 254.990 600.0 > (ns/day)(hour/ns) > Performance: 169.4190.142 Interesting - That means, for your record: i5-8600K, 6C/6T, AVX2_512: 169.419 ns/day R2-2700X, 8C/16T, AVX2_128: 152.728 ns/day R1-1700X, 8C/16T, AVX2_128: 141.456 ns/day your overclocked coffee-lake i5 would already be slightly faster than a Ryzen2 2700X. But if you want to simulate larger membrane systems, you should possibly test also on larger membrane systems. There's a DPPC/Water membrane benchmark included in the Gromacs benchmark set. Unfortunately, it (d.dppc) wouldn't work out of the box with newer versions. I patched it up and converted it into a NPT membrane-benchmark: http://gromacs.deppenfraktur.de/gmxbench-d.dppc-npt.tgz and already got some results for some of our antique boxes: Results for 2018.3: *** i5-7600K, 4C/4T, AVX2_512: 5.512 ns/day i7-6700K, 4C/8T, AVX2_256: 6.858 ns/day R1-1700X, 8C/16T, AVX2_128: 7.490 ns/day * i7-5820K, 6C/12T, AVX2_256: 9.162 ns/day * R-2 2700X, 8C/16T, AVX2_128: 9.401 ns/day ** GTX-980, i5-7600K, 4C/4T, AVX2_512: 28.071 ** GTX-1080, R1-1700X, 8C/16T, AVX2_128: 42.908 If you get some people with contemporary hardware to test for such workloads related to your projects, you can possibly buy boxes accordingly. M. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Hi, Mirco. 2018-09-11 3:29 GMT+07:00 Wahab Mirco : > > How would the 8600K perform without GPU (using the "-nb cpu" > option on the mdrun binary)? > > 8600K with frequency 4600Mhz was used Command line: gmx mdrun -nb cpu Core t (s) Wall t (s)(%) Time: 1529.942 254.990 600.0 (ns/day)(hour/ns) Performance: 169.4190.142 Finished mdrun on rank 0 Wed Sep 12 02:29:41 2018 It may be interesting, the run with only 3 threads: Command line: gmx mdrun -nb cpu -nt 3 Core t (s) Wall t (s)(%) Time: 1435.090 478.363 300.0 (ns/day)(hour/ns) Performance: 90.3080.266 Finished mdrun on rank 0 Wed Sep 12 02:44:50 2018 - a single R2-2700X possibly cannot always saturate TWO > GTX-1080/1080Ti (maybe Szilárd Páll can add some rule of > thumb whether and when it can), > >- due to AVX2_512, 6C/12T "Coffee Lake" i7 may be somehow faster than > even 8C/16T Ryzen2 (at some scenarios significantly?) > Thank you, it's very helpful information. > I can provide some performance numbers I got after updating > Gromacs to v.2018.3 lately on our aging lab hardware. > > ==> (I can provide the input file if required) ==> Yes, it will be great if you can provide some test results with input file (maybe it can help) -- Best regards, Olga Selyutina -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Hi, Páll 2018-09-11 21:50 GMT+07:00 Szilárd Páll : > I've presonally not tried the 2nd-gen Ryzen, but based on the performance > of the 1800X and the compute benchmarks I've seen published, these will > likely be the best value for money (possibly the 2600X if that allows > getting better GPUs?). > It makes sence. Benchmarks are very similar but price is about 30% lower! I'll think carefully about it. > Special edition CPU, not worth the money unless you like the name ;) > > It's for lovers of rare names :) > Don't forget that you need a few other components too in particular note > that that the motherboard will be more expensive for Threadripper, cheaper > for Ryzen 7 (and unless Intel boards are also a bit more pricey than the > Ryzens). > > Yes, I remember. The cost of other components was taken into account when choosing CPU and GPU. 2018-09-11 21:55 GMT+07:00 Szilárd Páll : > > Also, I can't (yet) recommend AMD GPUs as a buying option for > consumer-grade stuff as we don't yet have PME offload support in OpenCL, > but this will soon change. > > Thanks for advice, I haven't take AMD GPUs into account yet. > > Glad to hear. Feel free to follow up if you have further questions (but do > allow some time for replies ;). > > Thank you, I will :) -- Best regards, Olga Selyutina -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
On 07.09.2018 20:41, Olga Selyutina wrote: > ... > These sets of CPU and GPU are suitable for price (in our region): > *GPU* > GTX 1070 ~1700MHz, cuda 1920 - $514 > GTX 1080 ~1700MHz, cuda 2560 - $615 > GTX 1070Ti ~1700MHz, cuda 2432 - $615 > GTX 1080Ti ~1600MHz, cuda 3584 - $930 > > *CPU* > Ryzen 7 2700X - $357 > 4200MHz, 8/16 cores/threads, cache L1/L2/L3 768KB/4MB/16MB, 105W, max.T 85C > > Threadripper 1950X - $930 > 4000MHz, 16/32 cores/threads, cache L1/L2/L3 1.5/8/32MB, 180W, max.T 68C > > i7 8086K - $515 > 4800MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C > > i7 8700K - $442 > 4600MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C > > The most suitable combinations CPU+GPU are as follows: > 1) Ryzen 7 2700X + two GTX 1080 - $1587 > 1.1) Ryzen 7 2700X + one GTX 1080 + one GTX 1080*Ti* - $1900 (maybe?) > 2) Threadripper 1950X + one GTX 1080Ti - $1860 > 3) i7 8700K + two GTX 1080 - $1672 > 4) Ryzen 7 2700X + three GTX 1070 - $1900 > My suggestions: > Variant 1 seems to be the most suitable. > Variant 2 seems to be suitable only if the single simulation is running on > workstation > It’s a bit confusing that in synthetic tests/games performance of i7 8700 > is higher than Ryzen 7 2700. > ... Sorry for jumping into the thread at this point, but depending on the problem size and type, it might happen that: - a single R2-2700X possibly cannot always saturate TWO GTX-1080/1080Ti (maybe Szilárd Páll can add some rule of thumb whether and when it can), - a larger Threadripper may suffer from memory bandwidth issues on many parallel threads in Gromacs, - due to AVX2_512, 6C/12T "Coffee Lake" i7 may be somehow faster than even 8C/16T Ryzen2 (at some scenarios significantly?) But it most probably depends on the circumstances. If you want to spend your money effectively, you have to perform at least some tests. I can provide some performance numbers I got after updating Gromacs to v.2018.3 lately on our aging lab hardware. Regards, Mirco ==> (I can provide the input file if required) ==> -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Sadly, I can't recommend packaged versions of GROMACS for anything other than pre- or post-processing or non-performance critical work; these are compiled with proper SIMD support which is generally wasteful. Also, I can't (yet) recommend AMD GPUs as a buying option for consumer-grade stuff as we don't yet have PME offload support in OpenCL, but this will soon change. Additionally and importantly, I can't recommend the MESA stack, it's just not competitive in performance. Use ROCm (or AMDGPU-PRO). -- Szilárd On Mon, Sep 10, 2018 at 8:21 PM Benson Muite wrote: > Some results (probably suboptimal) for d.poly-ch2 on a desktop running > Fedora 28 and using Gromacs-Opencl from Fedora repositories: > > Log file opened on Mon Sep 10 21:00:25 2018 > Host: mikihir pid: 32669 rank ID: 0 number of ranks: 1 >:-) GROMACS - gmx mdrun, 2018.2 (-: > > GROMACS is written by: > Emile Apol Rossen Apostolov Paul Bauer Herman J.C. > Berendsen > Par BjelkmarAldert van Buuren Rudi van Drunen Anton Feenstra >Gerrit GroenhofAleksei Iupinov Christoph Junghans Anca Hamuraru > Vincent Hindriksen Dimitrios KarkoulisPeter KassonJiri Kraus >Carsten Kutzner Per Larsson Justin A. LemkulViveca Lindahl >Magnus Lundborg Pieter MeulenhoffErik Marklund Teemu Murtola > Szilard Pall Sander Pronk Roland Schulz Alexey Shvetsov > Michael Shirts Alfons Sijbers Peter TielemanTeemu > Virolainen > Christian WennbergMaarten Wolf > and the project leaders: > Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel > > Copyright (c) 1991-2000, University of Groningen, The Netherlands. > Copyright (c) 2001-2017, The GROMACS development team at > Uppsala University, Stockholm University and > the Royal Institute of Technology, Sweden. > check out http://www.gromacs.org for more information. > > GROMACS is free software; you can redistribute it and/or modify it > under the terms of the GNU Lesser General Public License > as published by the Free Software Foundation; either version 2.1 > of the License, or (at your option) any later version. > > GROMACS: gmx mdrun, version 2018.2 > Executable: /usr/bin/gmx > Data prefix: /usr > Working dir: /home/benson/Projects/GromacsBench/d.poly-ch2 > Command line: >gmx mdrun > > GROMACS version:2018.2 > Precision: single > Memory model: 64 bit > MPI library:thread_mpi > OpenMP support: enabled (GMX_OPENMP_MAX_THREADS = 64) > GPU support:OpenCL > SIMD instructions: SSE2 > FFT library:fftw-3.3.5-sse2-avx > RDTSCP usage: disabled > TNG support:enabled > Hwloc support: hwloc-1.11.6 > Tracing support:disabled > Built on: 2018-07-19 19:45:21 > Built by: mockbuild@ [CMAKE] > Build OS/arch: Linux 4.17.3-200.fc28.x86_64 x86_64 > Build CPU vendor: Intel > Build CPU brand:Intel Core Processor (Haswell, no TSX) > Build CPU family: 6 Model: 60 Stepping: 1 > Build CPU features: aes apic avx avx2 clfsh cmov cx8 cx16 f16c fma intel > lahf mmx msr pcid pclmuldq popcnt pse rdrnd rdtscp sse2 sse3 sse4.1 > sse4.2 ssse3 tdt x2apic > C compiler: /usr/bin/cc GNU 8.1.1 > C compiler flags:-msse2 -O2 -g -pipe -Wall -Werror=format-security > -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions > -fstack-protector-strong -grecord-gcc-switches > -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 > -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic > -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection > -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 > -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong > -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 > -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic > -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection > -DNDEBUG -funroll-all-loops -fexcess-precision=fast > C++ compiler: /usr/bin/c++ GNU 8.1.1 > C++ compiler flags: -msse2 -O2 -g -pipe -Wall -Werror=format-security > -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions > -fstack-protector-strong -grecord-gcc-switches > -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 > -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic > -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection > -std=c++11 -DNDEBUG -funroll-all-loops -fexcess-precision=fast > OpenCL include dir: /usr/include > OpenCL library: /usr/lib64/libOpenCL.so > OpenCL version: 2.0 > > > Running on 1 node with total 8 cores, 8 logical cores, 1 compatible GPU > Hardware detected: >CPU info: > Vendor: AMD > Brand: AMD FX(tm)-8350 Eight-Core Processor > Family: 21 Model: 2 Stepping: 0 > Features: aes amd apic avx clfsh
Re: [gmx-users] Workstation choice
Hi, On Fri, Sep 7, 2018 at 8:40 PM Olga Selyutina wrote: > Hi, > A lot of thanks for valuable information. > If it isn’t difficult for you, could you answer how the growth of > performance under using the second GPU on the single simulation was changed > in GROMACS 2018 vs older versions (2016, 5.1, it was 20-30% higher)? > Short answer: Greatly depends on the simulation setup and system, under circumstances ideal for scaling (large input, little constraint work, etc.) ~1.7x without PME offload and max 1.5x with PME offload. Note that the PME offload of the current release was not optimized for scaling but rather for single GPU performance. It's worth to note that a peculiar behavior with scaling is that from 1 to 2 GPUs you get less efficient scaling than from 2 to 4 (assuming that you have enough work per GPU to scale). This is because domain decomposition (DD) incurs a "one-time performance hit" due to the additional work involved in decomposing the system. 2018-09-07 23:25 GMT+07:00 Szilárd Páll : > > > > > Are you intending to use it mostly/only for running simulations or also > as > > a desktop computer? > > > > Yes, it will be mostly used for simulations. > If so, I'd recommend getting decent cooling and consider a chassis with soundproofing if you'll have the machine in an office, especially if you'll have 2x GPUs. > > I'm not on the top of pricing details so you should probably look at some > > configs and get back with concrete CPU + GPU (+price) combinations and we > > might be able to guesstimate what's best. > > > > > These sets of CPU and GPU are suitable for price (in our region): > *GPU* > GTX 1070 ~1700MHz, cuda 1920 - $514 > GTX 1080 ~1700MHz, cuda 2560 - $615 > GTX 1070Ti ~1700MHz, cuda 2432 - $615 > GTX 1080Ti ~1600MHz, cuda 3584 - $930 > 1080 and 1070Ti should be about the same in performance, but the latter should (in theory) be a little cheaper. > *CPU* > Ryzen 7 2700X - $357 > 4200MHz, 8/16 cores/threads, cache L1/L2/L3 768KB/4MB/16MB, 105W, max.T 85C > I've presonally not tried the 2nd-gen Ryzen, but based on the performance of the 1800X and the compute benchmarks I've seen published, these will likely be the best value for money (possibly the 2600X if that allows getting better GPUs?). > Threadripper 1950X - $930 > 4000MHz, 16/32 cores/threads, cache L1/L2/L3 1.5/8/32MB, 180W, max.T 68C > The 1920X still has 12 cores and should cost significantly less. Side-note: one more thing you're getting with Thredripper vs Ryzen or Intel Coffe Lake is more PCIe lanes that will matter somewhat if you use two GPUs. i7 8086K - $515 > 4800MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C > Special edition CPU, not worth the money unless you like the name ;) > i7 8700K - $442 > 4600MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C > Not sure about the perf/price difference, but the 8700 (non-K) might be reasonable option too. > The most suitable combinations CPU+GPU are as follows: > 1) Ryzen 7 2700X + two GTX 1080 - $1587 > 1.1) Ryzen 7 2700X + one GTX 1080 + one GTX 1080*Ti* - $1900 (maybe?) > 2) Threadripper 1950X + one GTX 1080Ti - $1860 > 3) i7 8700K + two GTX 1080 - $1672 > 4) Ryzen 7 2700X + three GTX 1070 - $1900 > My suggestions: > Variant 1 seems to be the most suitable. > Variant 2 seems to be suitable only if the single simulation is running on > workstation > Don't forget that you need a few other components too in particular note that that the motherboard will be more expensive for Threadripper, cheaper for Ryzen 7 (and unless Intel boards are also a bit more pricey than the Ryzens). > It’s a bit confusing that in synthetic tests/games performance of i7 8700 > is higher than Ryzen 7 2700. > Consumer and especially gaming benchmarks rarely reflect the performance of an HPC workload. Also note that CPU-only GROMACS benchmarks are _not_ a good indicative of performance with GPU offload as the most compute-intensive parts of the code that generally run very efficiently on CPUs too (and would favor Intel Skylake with AVX512 units) runs in these cases on the GPU. The code that remains is generally less arithmetically intensive. > Thanks a lot again for your advice, it has already clarified a lot! > Glad to hear. Feel free to follow up if you have further questions (but do allow some time for replies ;). > -- > Best regards, Olga Selyutina > -- > Gromacs Users mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before > posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe
Re: [gmx-users] Workstation choice
d.poly-ch2 from gmxbench-3.0.tar.gz available at: http://www.gromacs.org/About_Gromacs/Benchmarks 6000 atoms timestep 0.001 On 09/10/2018 11:37 PM, Albert wrote: May I ask how many atoms in the system? Which forcefield did you use? And what's the time step? regards On 09/10/2018 09:14 PM, Olga Selyutina wrote: (50 ts instead of 5000 ts, otherwise it's too fast) That would be a factor of about 227! On an available machine (not used for simulations) self-compiled 2018.3, CUDA 9.1 the similar result, factor is about 136: Working dir: /home/user/dev/gromacs/gmxbench-3.0/d.poly-ch2 Command line: gmx mdrun -ntmpi 1 -nt 6 Running on 1 node with total 6 cores, 6 logical cores, 1 compatible GPU Hardware detected: CPU info: Vendor: Intel Brand: Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz GPU info: Number of GPUs detected: 1 #0: NVIDIA GeForce GTX 1050 Ti, compute cap.: 6.1, ECC: no, stat: compatible On 1 MPI rank, each using 6 OpenMP threads Core t (s) Wall t (s) (%) Time: 1102.853 183.809 600.0 (ns/day) (hour/ns) Performance: 235.027 0.102 Finished mdrun on rank 0 Tue Sep 11 01:53:36 2018 -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
May I ask how many atoms in the system? Which forcefield did you use? And what's the time step? regards On 09/10/2018 09:14 PM, Olga Selyutina wrote: (50 ts instead of 5000 ts, otherwise it's too fast) That would be a factor of about 227! On an available machine (not used for simulations) self-compiled 2018.3, CUDA 9.1 the similar result, factor is about 136: Working dir: /home/user/dev/gromacs/gmxbench-3.0/d.poly-ch2 Command line: gmx mdrun -ntmpi 1 -nt 6 Running on 1 node with total 6 cores, 6 logical cores, 1 compatible GPU Hardware detected: CPU info: Vendor: Intel Brand: Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz GPU info: Number of GPUs detected: 1 #0: NVIDIA GeForce GTX 1050 Ti, compute cap.: 6.1, ECC: no, stat: compatible On 1 MPI rank, each using 6 OpenMP threads Core t (s) Wall t (s)(%) Time: 1102.853183.809 600.0 (ns/day)(hour/ns) Performance: 235.0270.102 Finished mdrun on rank 0 Tue Sep 11 01:53:36 2018 -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
On 10.09.2018 21:15, Olga Selyutina wrote: >> >> >> (50 ts instead of 5000 ts, otherwise it's too fast) >> >> >> That would be a factor of about 227! > > > On an available machine (not used for simulations) > self-compiled 2018.3, CUDA 9.1 > the similar result, factor is about 136: > > Working dir: /home/user/dev/gromacs/gmxbench-3.0/d.poly-ch2 > Command line: >gmx mdrun -ntmpi 1 -nt 6 > > Running on 1 node with total 6 cores, 6 logical cores, 1 compatible GPU > Hardware detected: >CPU info: > Vendor: Intel > Brand: Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz > >GPU info: > Number of GPUs detected: 1 > #0: NVIDIA GeForce GTX 1050 Ti, compute cap.: 6.1, ECC: no, stat: > compatible > > On 1 MPI rank, each using 6 OpenMP threads > > Core t (s) Wall t (s)(%) > Time: 1102.853183.809 600.0 > (ns/day)(hour/ns) > Performance: 235.0270.102 > Finished mdrun on rank 0 Tue Sep 11 01:53:36 2018 How would the 8600K perform without GPU (using the "-nb cpu" option on the mdrun binary)? The Ryzen 1700X does the following (-nb cpu): Core t (s) Wall t (s)(%) Time: 4886.335 305.396 1600.0 (ns/day)(hour/ns) Performance: 141.4560.170 Finished mdrun on rank 0 Mon Sep 10 22:21:40 2018 That's still a factor 82 compared to Bratin Kumar's setup. M. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
> > > (50 ts instead of 5000 ts, otherwise it's too fast) > > > That would be a factor of about 227! On an available machine (not used for simulations) self-compiled 2018.3, CUDA 9.1 the similar result, factor is about 136: Working dir: /home/user/dev/gromacs/gmxbench-3.0/d.poly-ch2 Command line: gmx mdrun -ntmpi 1 -nt 6 Running on 1 node with total 6 cores, 6 logical cores, 1 compatible GPU Hardware detected: CPU info: Vendor: Intel Brand: Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz GPU info: Number of GPUs detected: 1 #0: NVIDIA GeForce GTX 1050 Ti, compute cap.: 6.1, ECC: no, stat: compatible On 1 MPI rank, each using 6 OpenMP threads Core t (s) Wall t (s)(%) Time: 1102.853183.809 600.0 (ns/day)(hour/ns) Performance: 235.0270.102 Finished mdrun on rank 0 Tue Sep 11 01:53:36 2018 -- Best regards, Olga Selyutina -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
On 10.09.2018 20:22, Benson Muite wrote: > Some results (probably suboptimal) for d.poly-ch2 on a desktop running > Fedora 28 and using Gromacs-Opencl from Fedora repositories: > ... > Running on 1 node with total 8 cores, 8 logical cores, 1 compatible GPU > Hardware detected: > CPU info: > Vendor: AMD > Brand: AMD FX(tm)-8350 Eight-Core Processor > GPU info: > Number of GPUs detected: 1 > #0: name: Radeon RX 560 Series (POLARIS11 / DRM 3.23.0 / > > ... > Average per-step force GPU/CPU evaluation time ratio: 2.170 ms/17.768 ms > = 0.122 > > Core t (s) Wall t (s) (%) > Time: 2001.585 250.198 800.0 > (ns/day) (hour/ns) > Performance: 1.727 13.897 > Finished mdrun on rank 0 Mon Sep 10 21:04:37 2018 For your record: self-compiled 2018.3, CUDA 9.2, R-1700X, GTX-1080, (50 ts instead of 5000 ts, otherwise it's too fast) CPU info: Vendor: AMD Brand: AMD Ryzen 7 1700X Eight-Core Processor GPU info: Number of GPUs detected: 1 #0: NVIDIA GeForce GTX 1080, compute cap.: 6.1, ECC: no, stat: ... Core t (s) Wall t (s)(%) Time: 1764.209 110.263 1600.0 (ns/day)(hour/ns) Performance: 391.7910.061 Finished mdrun on rank 0 Mon Sep 10 20:42:22 2018 That would be a factor of about 227! M. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Some results (probably suboptimal) for d.poly-ch2 on a desktop running Fedora 28 and using Gromacs-Opencl from Fedora repositories: Log file opened on Mon Sep 10 21:00:25 2018 Host: mikihir pid: 32669 rank ID: 0 number of ranks: 1 :-) GROMACS - gmx mdrun, 2018.2 (-: GROMACS is written by: Emile Apol Rossen Apostolov Paul Bauer Herman J.C. Berendsen Par Bjelkmar Aldert van Buuren Rudi van Drunen Anton Feenstra Gerrit Groenhof Aleksei Iupinov Christoph Junghans Anca Hamuraru Vincent Hindriksen Dimitrios Karkoulis Peter Kasson Jiri Kraus Carsten Kutzner Per Larsson Justin A. Lemkul Viveca Lindahl Magnus Lundborg Pieter Meulenhoff Erik Marklund Teemu Murtola Szilard Pall Sander Pronk Roland Schulz Alexey Shvetsov Michael Shirts Alfons Sijbers Peter Tieleman Teemu Virolainen Christian Wennberg Maarten Wolf and the project leaders: Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel Copyright (c) 1991-2000, University of Groningen, The Netherlands. Copyright (c) 2001-2017, The GROMACS development team at Uppsala University, Stockholm University and the Royal Institute of Technology, Sweden. check out http://www.gromacs.org for more information. GROMACS is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. GROMACS: gmx mdrun, version 2018.2 Executable: /usr/bin/gmx Data prefix: /usr Working dir: /home/benson/Projects/GromacsBench/d.poly-ch2 Command line: gmx mdrun GROMACS version: 2018.2 Precision: single Memory model: 64 bit MPI library: thread_mpi OpenMP support: enabled (GMX_OPENMP_MAX_THREADS = 64) GPU support: OpenCL SIMD instructions: SSE2 FFT library: fftw-3.3.5-sse2-avx RDTSCP usage: disabled TNG support: enabled Hwloc support: hwloc-1.11.6 Tracing support: disabled Built on: 2018-07-19 19:45:21 Built by: mockbuild@ [CMAKE] Build OS/arch: Linux 4.17.3-200.fc28.x86_64 x86_64 Build CPU vendor: Intel Build CPU brand: Intel Core Processor (Haswell, no TSX) Build CPU family: 6 Model: 60 Stepping: 1 Build CPU features: aes apic avx avx2 clfsh cmov cx8 cx16 f16c fma intel lahf mmx msr pcid pclmuldq popcnt pse rdrnd rdtscp sse2 sse3 sse4.1 sse4.2 ssse3 tdt x2apic C compiler: /usr/bin/cc GNU 8.1.1 C compiler flags: -msse2 -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -DNDEBUG -funroll-all-loops -fexcess-precision=fast C++ compiler: /usr/bin/c++ GNU 8.1.1 C++ compiler flags: -msse2 -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -std=c++11 -DNDEBUG -funroll-all-loops -fexcess-precision=fast OpenCL include dir: /usr/include OpenCL library: /usr/lib64/libOpenCL.so OpenCL version: 2.0 Running on 1 node with total 8 cores, 8 logical cores, 1 compatible GPU Hardware detected: CPU info: Vendor: AMD Brand: AMD FX(tm)-8350 Eight-Core Processor Family: 21 Model: 2 Stepping: 0 Features: aes amd apic avx clfsh cmov cx8 cx16 f16c fma fma4 htt lahf misalignsse mmx msr nonstop_tsc pclmuldq pdpe1gb popcnt pse rdtscp sse2 sse3 sse4a sse4.1 sse4.2 ssse3 xop Hardware topology: Full, with devices Sockets, cores, and logical processors: Socket 0: [ 0] [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] Numa nodes: Node 0 (16714620928 bytes mem): 0 1 2 3 4 5 6 7 Latency: 0 0 1.00 Caches: L1: 16384 bytes, linesize 64 bytes, assoc. 4, shared 1 ways L2: 2097152 bytes, linesize 64 bytes, assoc. 16, shared 2 ways L3: 8388608 bytes, linesize 64 bytes, assoc. 64, shared 8 ways PCI devices: :01:00.0 Id: 1002:67ef Class: 0x0300 Numa: 0
Re: [gmx-users] Workstation choice
It takes a bit of time to benchmark, and probably more maintenance needed. May want to see what kind of performance you get with laptop processor - may require some development of OpenCL to better use integrated GPUs. Are you able to provide a log file for a typical run, these indicate time spent in each part of the computation. On 09/09/2018 01:32 PM, Olga Selyutina wrote: Thank you, an idea to build own low-price cluster is very interesting. Seven Ryzen 3 2200 with hardware cost ~$2000 and their total performance for some purposes could be higher than performance of one high-performance CPU with few GPUs (configurations CPU+GPUs mentioned in my previous letter) . But I’m not sure that they provide higher performance for GROMACS simulations. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Thank you, an idea to build own low-price cluster is very interesting. Seven Ryzen 3 2200 with hardware cost ~$2000 and their total performance for some purposes could be higher than performance of one high-performance CPU with few GPUs (configurations CPU+GPUs mentioned in my previous letter) . But I’m not sure that they provide higher performance for GROMACS simulations. 2018-09-09 12:59 GMT+07:00 Benson Muite : > This is old, but seems to indicate Beowulf clusters work quite well: > > https://docs.uabgrid.uab.edu/wiki/Gromacs_Benchmark > > Szilárd had helped create a benchmark data set available at: > http://www.gromacs.org/About_Gromacs/Benchmarks > http://www.gromacs.org/@api/deki/files/240/=gromacs-5.0-benchmarks.pdf > ftp://ftp.gromacs.org/pub/benchmarks/gmxbench-3.0.tar.gz > > Does your use case involves a large number of ensemble simulations which > can be done in single precision without error correction? If so might you > be better building a small Beowulf cluster with lower spec processors that > have integrated GPUs? For example a Ryzen 3 with integrated graphics is > about $100. Motherboard, RAM, power supply would probably get you to about > $300. Intel core I3 bundle would be about $350. Setup could be done using > OpenHPC stack: > http://www.openhpc.community/ > > This would get you a personal 5-7 node in house cluster. However, ability > to do maintenance, have local support for repair may also be important in > considering system lifetime cost, not just initial purchase price. Gromacs > current and future support for OpenCl, and likely also important here. > > At least one computer store in my region has allowed benchmarking. > > On 09/07/2018 09:40 PM, Olga Selyutina wrote: > >> Hi, >> A lot of thanks for valuable information. >> If it isn’t difficult for you, could you answer how the growth of >> performance under using the second GPU on the single simulation was >> changed >> in GROMACS 2018 vs older versions (2016, 5.1, it was 20-30% higher)? >> >> >> 2018-09-07 23:25 GMT+07:00 Szilárd Páll : >> >> Are you intending to use it mostly/only for running simulations or also as >>> a desktop computer? >>> >>> Yes, it will be mostly used for simulations. >>> >> >> I'm not on the top of pricing details so you should probably look at some >>> configs and get back with concrete CPU + GPU (+price) combinations and we >>> might be able to guesstimate what's best. >>> >>> >>> These sets of CPU and GPU are suitable for price (in our region): >> *GPU* >> GTX 1070 ~1700MHz, cuda 1920 - $514 >> GTX 1080 ~1700MHz, cuda 2560 - $615 >> GTX 1070Ti ~1700MHz, cuda 2432 - $615 >> GTX 1080Ti ~1600MHz, cuda 3584 - $930 >> >> *CPU* >> Ryzen 7 2700X - $357 >> 4200MHz, 8/16 cores/threads, cache L1/L2/L3 768KB/4MB/16MB, 105W, max.T >> 85C >> >> Threadripper 1950X - $930 >> 4000MHz, 16/32 cores/threads, cache L1/L2/L3 1.5/8/32MB, 180W, max.T 68C >> >> i7 8086K - $515 >> 4800MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C >> >> i7 8700K - $442 >> 4600MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C >> >> The most suitable combinations CPU+GPU are as follows: >> 1) Ryzen 7 2700X + two GTX 1080 - $1587 >> 1.1) Ryzen 7 2700X + one GTX 1080 + one GTX 1080*Ti* - $1900 (maybe?) >> 2) Threadripper 1950X + one GTX 1080Ti - $1860 >> 3) i7 8700K + two GTX 1080 - $1672 >> 4) Ryzen 7 2700X + three GTX 1070 - $1900 >> My suggestions: >> Variant 1 seems to be the most suitable. >> Variant 2 seems to be suitable only if the single simulation is running on >> workstation >> It’s a bit confusing that in synthetic tests/games performance of i7 8700 >> is higher than Ryzen 7 2700. >> Thanks a lot again for your advice, it has already clarified a lot! >> >> >> >> > -- > Gromacs Users mailing list > > * Please search the archive at http://www.gromacs.org/Support > /Mailing_Lists/GMX-Users_List before posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. > -- С уважением, Ольга Селютина -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
As an example on the d.poly-ch2 benchmark from ftp://ftp.gromacs.org/pub/benchmarks/gmxbench-3.0.tar.gz, I get: Log file opened on Sun Sep 9 09:03:49 2018 Host: localhost.localdomain pid: 14194 rank ID: 0 number of ranks: 1 :-) GROMACS - gmx mdrun, 2018.3 (-: GROMACS is written by: Emile Apol Rossen Apostolov Paul Bauer Herman J.C. Berendsen Par Bjelkmar Aldert van Buuren Rudi van Drunen Anton Feenstra Gerrit Groenhof Aleksei Iupinov Christoph Junghans Anca Hamuraru Vincent Hindriksen Dimitrios Karkoulis Peter Kasson Jiri Kraus Carsten Kutzner Per Larsson Justin A. Lemkul Viveca Lindahl Magnus Lundborg Pieter Meulenhoff Erik Marklund Teemu Murtola Szilard Pall Sander Pronk Roland Schulz Alexey Shvetsov Michael Shirts Alfons Sijbers Peter Tieleman Teemu Virolainen Christian Wennberg Maarten Wolf and the project leaders: Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel Copyright (c) 1991-2000, University of Groningen, The Netherlands. Copyright (c) 2001-2017, The GROMACS development team at Uppsala University, Stockholm University and the Royal Institute of Technology, Sweden. check out http://www.gromacs.org for more information. GROMACS is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. GROMACS: gmx mdrun, version 2018.3 Executable: /home/benson/Projects/GromacsTest/gmx-bench3/d.poly-ch2/../../gromacsinstall/bin/gmx Data prefix: /home/benson/Projects/GromacsTest/gmx-bench3/d.poly-ch2/../../gromacsinstall Working dir: /home/benson/Projects/GromacsTest/gmx-bench3/d.poly-ch2 Command line: gmx mdrun -deffnm mdbench GROMACS version: 2018.3 Precision: single Memory model: 64 bit MPI library: thread_mpi OpenMP support: enabled (GMX_OPENMP_MAX_THREADS = 64) GPU support: OpenCL SIMD instructions: AVX2_256 FFT library: fftw-3.3.8-sse2-avx-avx2-avx2_128-avx512 RDTSCP usage: enabled TNG support: enabled Hwloc support: disabled Tracing support: disabled Built on: 2018-08-24 19:14:49 Built by: benson@localhost.localdomain [CMAKE] Build OS/arch: Linux 4.17.14-202.fc28.x86_64 x86_64 Build CPU vendor: Intel Build CPU brand: Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz Build CPU family: 6 Model: 78 Stepping: 3 Build CPU features: aes apic avx avx2 clfsh cmov cx8 cx16 f16c fma htt intel lahf mmx msr nonstop_tsc pcid pclmuldq pdcm p dpe1gb popcnt pse rdrnd rdtscp sse2 sse3 sse4.1 sse4.2 ssse3 tdt x2apic C compiler: /usr/bin/cc GNU 8.1.1 C compiler flags: -march=core-avx2 -O3 -DNDEBUG -funroll-all-loops -fexcess-precision=fast C++ compiler: /usr/bin/c++ GNU 8.1.1 C++ compiler flags: -march=core-avx2 -std=c++11 -O3 -DNDEBUG -funroll-all-loops -fexcess-precision=fast OpenCL include dir: /usr/include OpenCL library: /usr/lib64/libOpenCL.so OpenCL version: 2.0 Running on 1 node with total 2 cores, 4 logical cores, 0 compatible GPUs Hardware detected: CPU info: Vendor: Intel Brand: Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz Family: 6 Model: 78 Stepping: 3 Features: aes apic avx avx2 clfsh cmov cx8 cx16 f16c fma htt intel lahf mmx msr nonstop_tsc pcid pclmuldq pdcm pdpe1gb popcnt pse rdrnd rdtscp sse2 sse3 sse4.1 sse4.2 ssse3 tdt x2apic Hardware topology: Basic Sockets, cores, and logical processors: Socket 0: [ 0 2] [ 1 3] GPU info: Number of GPUs detected: 1 #0: name: Intel(R) HD Graphics Skylake ULT GT2, vendor: Intel, device version: OpenCL 2.0 beignet 1.3, stat: incompati ble PLEASE READ AND CITE THE FOLLOWING REFERENCE M. J. Abraham, T. Murtola, R. Schulz, S. Páll, J. C. Smith, B. Hess, E. Lindahl GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers SoftwareX 1 (2015) pp. 19-25 --- Thank You --- PLEASE READ AND CITE THE FOLLOWING REFERENCE S. Páll, M. J. Abraham, C. Kutzner, B. Hess, E. Lindahl Tackling Exascale Software Challenges in Molecular Dynamics Simulations with GROMACS In S. Markidis & E. Laure (Eds.), Solving Software Challenges for Exascale 8759 (2015) pp. 3-27 --- Thank You --- PLEASE READ AND CITE THE FOLLOWING REFERENCE S. Pronk, S. Páll, R. Schulz, P. Larsson, P. Bjelkmar, R. Apostolov, M. R. Shirts, J. C. Smith, P. M. Kasson, D. van der Spoel, B. Hess, and E. Lindahl GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit Bioinformatics 29 (2013) pp. 845-54 --- Thank You ---
Re: [gmx-users] Workstation choice
This is old, but seems to indicate Beowulf clusters work quite well: https://docs.uabgrid.uab.edu/wiki/Gromacs_Benchmark Szilárd had helped create a benchmark data set available at: http://www.gromacs.org/About_Gromacs/Benchmarks http://www.gromacs.org/@api/deki/files/240/=gromacs-5.0-benchmarks.pdf ftp://ftp.gromacs.org/pub/benchmarks/gmxbench-3.0.tar.gz Does your use case involves a large number of ensemble simulations which can be done in single precision without error correction? If so might you be better building a small Beowulf cluster with lower spec processors that have integrated GPUs? For example a Ryzen 3 with integrated graphics is about $100. Motherboard, RAM, power supply would probably get you to about $300. Intel core I3 bundle would be about $350. Setup could be done using OpenHPC stack: http://www.openhpc.community/ This would get you a personal 5-7 node in house cluster. However, ability to do maintenance, have local support for repair may also be important in considering system lifetime cost, not just initial purchase price. Gromacs current and future support for OpenCl, and likely also important here. At least one computer store in my region has allowed benchmarking. On 09/07/2018 09:40 PM, Olga Selyutina wrote: Hi, A lot of thanks for valuable information. If it isn’t difficult for you, could you answer how the growth of performance under using the second GPU on the single simulation was changed in GROMACS 2018 vs older versions (2016, 5.1, it was 20-30% higher)? 2018-09-07 23:25 GMT+07:00 Szilárd Páll : Are you intending to use it mostly/only for running simulations or also as a desktop computer? Yes, it will be mostly used for simulations. I'm not on the top of pricing details so you should probably look at some configs and get back with concrete CPU + GPU (+price) combinations and we might be able to guesstimate what's best. These sets of CPU and GPU are suitable for price (in our region): *GPU* GTX 1070 ~1700MHz, cuda 1920 - $514 GTX 1080 ~1700MHz, cuda 2560 - $615 GTX 1070Ti ~1700MHz, cuda 2432 - $615 GTX 1080Ti ~1600MHz, cuda 3584 - $930 *CPU* Ryzen 7 2700X - $357 4200MHz, 8/16 cores/threads, cache L1/L2/L3 768KB/4MB/16MB, 105W, max.T 85C Threadripper 1950X - $930 4000MHz, 16/32 cores/threads, cache L1/L2/L3 1.5/8/32MB, 180W, max.T 68C i7 8086K - $515 4800MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C i7 8700K - $442 4600MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C The most suitable combinations CPU+GPU are as follows: 1) Ryzen 7 2700X + two GTX 1080 - $1587 1.1) Ryzen 7 2700X + one GTX 1080 + one GTX 1080*Ti* - $1900 (maybe?) 2) Threadripper 1950X + one GTX 1080Ti - $1860 3) i7 8700K + two GTX 1080 - $1672 4) Ryzen 7 2700X + three GTX 1070 - $1900 My suggestions: Variant 1 seems to be the most suitable. Variant 2 seems to be suitable only if the single simulation is running on workstation It’s a bit confusing that in synthetic tests/games performance of i7 8700 is higher than Ryzen 7 2700. Thanks a lot again for your advice, it has already clarified a lot! -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Hi, A lot of thanks for valuable information. If it isn’t difficult for you, could you answer how the growth of performance under using the second GPU on the single simulation was changed in GROMACS 2018 vs older versions (2016, 5.1, it was 20-30% higher)? 2018-09-07 23:25 GMT+07:00 Szilárd Páll : > > Are you intending to use it mostly/only for running simulations or also as > a desktop computer? > > Yes, it will be mostly used for simulations. > I'm not on the top of pricing details so you should probably look at some > configs and get back with concrete CPU + GPU (+price) combinations and we > might be able to guesstimate what's best. > > These sets of CPU and GPU are suitable for price (in our region): *GPU* GTX 1070 ~1700MHz, cuda 1920 - $514 GTX 1080 ~1700MHz, cuda 2560 - $615 GTX 1070Ti ~1700MHz, cuda 2432 - $615 GTX 1080Ti ~1600MHz, cuda 3584 - $930 *CPU* Ryzen 7 2700X - $357 4200MHz, 8/16 cores/threads, cache L1/L2/L3 768KB/4MB/16MB, 105W, max.T 85C Threadripper 1950X - $930 4000MHz, 16/32 cores/threads, cache L1/L2/L3 1.5/8/32MB, 180W, max.T 68C i7 8086K - $515 4800MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C i7 8700K - $442 4600MHz, 6/12 cores/threads, cache L2/L3 1.5/12MB, 95W, max.T 100C The most suitable combinations CPU+GPU are as follows: 1) Ryzen 7 2700X + two GTX 1080 - $1587 1.1) Ryzen 7 2700X + one GTX 1080 + one GTX 1080*Ti* - $1900 (maybe?) 2) Threadripper 1950X + one GTX 1080Ti - $1860 3) i7 8700K + two GTX 1080 - $1672 4) Ryzen 7 2700X + three GTX 1070 - $1900 My suggestions: Variant 1 seems to be the most suitable. Variant 2 seems to be suitable only if the single simulation is running on workstation It’s a bit confusing that in synthetic tests/games performance of i7 8700 is higher than Ryzen 7 2700. Thanks a lot again for your advice, it has already clarified a lot! -- Best regards, Olga Selyutina -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
On Fri, Sep 7, 2018 at 4:15 PM Benson Muite wrote: > Check if the routines you will use have been ported to use GPUs. Time > and profile a typical run you will perform on your current hardware to > determine the bottlenecks, and then choose hardware that will perform > best on these bottlenecks and is within your budget. > > Note that benchmarking can be quite time consuming - even without > performance optimization. If you have an idea of the simulations you > will run and have profiled your code, and will primarily do these > simulations, then it is probably easiest to go to a computer store where > they have several available workstations, run your benchmark yourself > and choose the best value system. > Not a bad suggestion, but these days people shop online. Even if one would opt for a physical store, benchmarking each candidate doesn't sound feasible (or easily done). You'd need to have multiple systems assembled, OS and benchmarks scripts ready on a bootable drive, and measure performance of each system before buying. > > On 09/06/2018 08:09 PM, Olga Selyutina wrote: > > Thank you, > > > > to use these tests isn’t a problem, but available workstation are too > old, > > so test results won’t be helpful in choosing actual components. I have an > > access to high-performance cluster, but it also couldn’t be the basis > > because of its high cost. If it is hard to give the certain > recommendation, > > I would to know at least how do performance correlate with the number of > > cores/threads and core frequency? What is more important the core > frequency > > or number of cores? What is preferable, Intel or AMD? How should be > > correlated GPU and CPU performance? > > > > In the previous letters in mail-list I found recommendations on purchase > of > > Threadripper 1950X+GeForce GTX-1080Ti, I think that more cheap and > similar > > effective choice is TR 1950X+GeForce GTX-*1080*. Will there be > significant > > differences with 1)TR 19*20*X+GeForce GTX-1080 2) Intel 7900x+GeForce > > GTX-1080? > > > > 2018-09-05 20:53 GMT+07:00 Benson Muite : > > > >> Hi Olga, > >> > >> The authors of: > >> > >> https://github.com/bio-phys/MDBenchmark/tree/version-1.3.2 > >> > >> https://zenodo.org/record/1318123 > >> > >> May be helpful. If your data is not confidential, you may consider > running > >> Gromacs remotely on a cloud high performance computing resource, or > >> benchmarking remotely and then purchasing a suitable configuration. > >> > >> Some other information: > >> http://manual.gromacs.org/documentation/current/user-guide/ > >> mdrun-performance.html# > >> http://www.gromacs.org/GPU_acceleration > >> https://extras.csc.fi/chem/courses/gmx2007/Erik_Talks/buildi > >> ng_clusters.pdf > >> http://www.gromacs.org/Documentation/Performance_checklist > >> > >> On 09/05/2018 12:02 PM, Olga Selyutina wrote: > >> > >>>Hello, > >>> > >>> I need help in choice of the workstation for MD simulations using > GROMACS. > >>> It is supposed to study systems consisting from 30-50k atoms, in > >>> particular, lipid bilayer models. Since the last generation of Intel > and > >>> Ryzen CPU has made a big leap in performance, available workstation > can’t > >>> be taken as a basis. Please, help to choose components of the > workstation, > >>> particularly, GPU and CPU with total cost about $2000. What is better, > to > >>> buy two GPUs(sli) or one GPU but more effective? > >>> > >>> > >> -- > >> Gromacs Users mailing list > >> > >> * Please search the archive at http://www.gromacs.org/Support > >> /Mailing_Lists/GMX-Users_List before posting! > >> > >> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > >> > >> * For (un)subscribe requests visit > >> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > >> send a mail to gmx-users-requ...@gromacs.org. > > > > > > > > -- > Gromacs Users mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before > posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Hi, Are you intending to use it mostly/only for running simulations or also as a desktop computer? Starting with the 2018 release we offload more work to the GPU to account for the increase in the gap between the performance of CPUs and GPUs and the prevalence of (especially workstations) with >1 GPU per CPU (socket). As a result, fewer CPU cores are needed to keep a GPU busy. For "regular" biomolecular simulations (decent amount of water, no pulling, no free energy, etc.), the rule of thumb is that 3-4 workstation cores are enough to get up to 90% of the peak performance. If you have more non-offloaded work (i.e. not pair-interactions or PME), you might need a a few more cores to avoid CPU-bound simulations. Given the above, and assuming you generally have multiple simulations you can do concurrently (and you don't need to scale these to multiple GPUs), you are generally better off with two GPUs (e.g. second or third from the highest-end consumer GPU, e.g. a GTX 1080 or 1070) and a decent 8+ core CPU. That will also result in a more cost-effective setup as well as higher aggregate performance. Now, when it comes to the best absolute performance, the Intel i9's or the Threadripper 2's are hard to beat, but for your case those are probably too pricey and an overkill for a 1-2 GPU setup (fast enough to drive 3+ GPUs). I'd suggest looking at Ryzen 7's or possibly Threadripper 1's; the Intel i7 8600k/8700k are also good options, but likely more pricey and less value for money than the Ryzen 7's. I'm not on the top of pricing details so you should probably look at some configs and get back with concrete CPU + GPU (+price) combinations and we might be able to guesstimate what's best. Cheers, -- Szilárd On Thu, Sep 6, 2018 at 7:10 PM Olga Selyutina wrote: > Thank you, > > to use these tests isn’t a problem, but available workstation are too old, > so test results won’t be helpful in choosing actual components. I have an > access to high-performance cluster, but it also couldn’t be the basis > because of its high cost. If it is hard to give the certain recommendation, > I would to know at least how do performance correlate with the number of > cores/threads and core frequency? What is more important the core frequency > or number of cores? What is preferable, Intel or AMD? How should be > correlated GPU and CPU performance? > > In the previous letters in mail-list I found recommendations on purchase of > Threadripper 1950X+GeForce GTX-1080Ti, I think that more cheap and similar > effective choice is TR 1950X+GeForce GTX-*1080*. Will there be significant > differences with 1)TR 19*20*X+GeForce GTX-1080 2) Intel 7900x+GeForce > GTX-1080? > > 2018-09-05 20:53 GMT+07:00 Benson Muite : > > > Hi Olga, > > > > The authors of: > > > > https://github.com/bio-phys/MDBenchmark/tree/version-1.3.2 > > > > https://zenodo.org/record/1318123 > > > > May be helpful. If your data is not confidential, you may consider > running > > Gromacs remotely on a cloud high performance computing resource, or > > benchmarking remotely and then purchasing a suitable configuration. > > > > Some other information: > > http://manual.gromacs.org/documentation/current/user-guide/ > > mdrun-performance.html# > > http://www.gromacs.org/GPU_acceleration > > https://extras.csc.fi/chem/courses/gmx2007/Erik_Talks/buildi > > ng_clusters.pdf > > http://www.gromacs.org/Documentation/Performance_checklist > > > > On 09/05/2018 12:02 PM, Olga Selyutina wrote: > > > >> Hello, > >> > >> I need help in choice of the workstation for MD simulations using > GROMACS. > >> It is supposed to study systems consisting from 30-50k atoms, in > >> particular, lipid bilayer models. Since the last generation of Intel and > >> Ryzen CPU has made a big leap in performance, available workstation > can’t > >> be taken as a basis. Please, help to choose components of the > workstation, > >> particularly, GPU and CPU with total cost about $2000. What is better, > to > >> buy two GPUs(sli) or one GPU but more effective? > >> > >> > > -- > > Gromacs Users mailing list > > > > * Please search the archive at http://www.gromacs.org/Support > > /Mailing_Lists/GMX-Users_List before posting! > > > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > > > * For (un)subscribe requests visit > > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > > send a mail to gmx-users-requ...@gromacs.org. > > > > > -- > Best regards, Olga Selyutina > -- > Gromacs Users mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before > posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
Re: [gmx-users] Workstation choice
Check if the routines you will use have been ported to use GPUs. Time and profile a typical run you will perform on your current hardware to determine the bottlenecks, and then choose hardware that will perform best on these bottlenecks and is within your budget. Note that benchmarking can be quite time consuming - even without performance optimization. If you have an idea of the simulations you will run and have profiled your code, and will primarily do these simulations, then it is probably easiest to go to a computer store where they have several available workstations, run your benchmark yourself and choose the best value system. On 09/06/2018 08:09 PM, Olga Selyutina wrote: Thank you, to use these tests isn’t a problem, but available workstation are too old, so test results won’t be helpful in choosing actual components. I have an access to high-performance cluster, but it also couldn’t be the basis because of its high cost. If it is hard to give the certain recommendation, I would to know at least how do performance correlate with the number of cores/threads and core frequency? What is more important the core frequency or number of cores? What is preferable, Intel or AMD? How should be correlated GPU and CPU performance? In the previous letters in mail-list I found recommendations on purchase of Threadripper 1950X+GeForce GTX-1080Ti, I think that more cheap and similar effective choice is TR 1950X+GeForce GTX-*1080*. Will there be significant differences with 1)TR 19*20*X+GeForce GTX-1080 2) Intel 7900x+GeForce GTX-1080? 2018-09-05 20:53 GMT+07:00 Benson Muite : Hi Olga, The authors of: https://github.com/bio-phys/MDBenchmark/tree/version-1.3.2 https://zenodo.org/record/1318123 May be helpful. If your data is not confidential, you may consider running Gromacs remotely on a cloud high performance computing resource, or benchmarking remotely and then purchasing a suitable configuration. Some other information: http://manual.gromacs.org/documentation/current/user-guide/ mdrun-performance.html# http://www.gromacs.org/GPU_acceleration https://extras.csc.fi/chem/courses/gmx2007/Erik_Talks/buildi ng_clusters.pdf http://www.gromacs.org/Documentation/Performance_checklist On 09/05/2018 12:02 PM, Olga Selyutina wrote: Hello, I need help in choice of the workstation for MD simulations using GROMACS. It is supposed to study systems consisting from 30-50k atoms, in particular, lipid bilayer models. Since the last generation of Intel and Ryzen CPU has made a big leap in performance, available workstation can’t be taken as a basis. Please, help to choose components of the workstation, particularly, GPU and CPU with total cost about $2000. What is better, to buy two GPUs(sli) or one GPU but more effective? -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support /Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org. -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Thank you, to use these tests isn’t a problem, but available workstation are too old, so test results won’t be helpful in choosing actual components. I have an access to high-performance cluster, but it also couldn’t be the basis because of its high cost. If it is hard to give the certain recommendation, I would to know at least how do performance correlate with the number of cores/threads and core frequency? What is more important the core frequency or number of cores? What is preferable, Intel or AMD? How should be correlated GPU and CPU performance? In the previous letters in mail-list I found recommendations on purchase of Threadripper 1950X+GeForce GTX-1080Ti, I think that more cheap and similar effective choice is TR 1950X+GeForce GTX-*1080*. Will there be significant differences with 1)TR 19*20*X+GeForce GTX-1080 2) Intel 7900x+GeForce GTX-1080? 2018-09-05 20:53 GMT+07:00 Benson Muite : > Hi Olga, > > The authors of: > > https://github.com/bio-phys/MDBenchmark/tree/version-1.3.2 > > https://zenodo.org/record/1318123 > > May be helpful. If your data is not confidential, you may consider running > Gromacs remotely on a cloud high performance computing resource, or > benchmarking remotely and then purchasing a suitable configuration. > > Some other information: > http://manual.gromacs.org/documentation/current/user-guide/ > mdrun-performance.html# > http://www.gromacs.org/GPU_acceleration > https://extras.csc.fi/chem/courses/gmx2007/Erik_Talks/buildi > ng_clusters.pdf > http://www.gromacs.org/Documentation/Performance_checklist > > On 09/05/2018 12:02 PM, Olga Selyutina wrote: > >> Hello, >> >> I need help in choice of the workstation for MD simulations using GROMACS. >> It is supposed to study systems consisting from 30-50k atoms, in >> particular, lipid bilayer models. Since the last generation of Intel and >> Ryzen CPU has made a big leap in performance, available workstation can’t >> be taken as a basis. Please, help to choose components of the workstation, >> particularly, GPU and CPU with total cost about $2000. What is better, to >> buy two GPUs(sli) or one GPU but more effective? >> >> > -- > Gromacs Users mailing list > > * Please search the archive at http://www.gromacs.org/Support > /Mailing_Lists/GMX-Users_List before posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. -- Best regards, Olga Selyutina -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Workstation choice
Hi Olga, The authors of: https://github.com/bio-phys/MDBenchmark/tree/version-1.3.2 https://zenodo.org/record/1318123 May be helpful. If your data is not confidential, you may consider running Gromacs remotely on a cloud high performance computing resource, or benchmarking remotely and then purchasing a suitable configuration. Some other information: http://manual.gromacs.org/documentation/current/user-guide/mdrun-performance.html# http://www.gromacs.org/GPU_acceleration https://extras.csc.fi/chem/courses/gmx2007/Erik_Talks/building_clusters.pdf http://www.gromacs.org/Documentation/Performance_checklist On 09/05/2018 12:02 PM, Olga Selyutina wrote: Hello, I need help in choice of the workstation for MD simulations using GROMACS. It is supposed to study systems consisting from 30-50k atoms, in particular, lipid bilayer models. Since the last generation of Intel and Ryzen CPU has made a big leap in performance, available workstation can’t be taken as a basis. Please, help to choose components of the workstation, particularly, GPU and CPU with total cost about $2000. What is better, to buy two GPUs(sli) or one GPU but more effective? -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] Workstation choice
Hello, I need help in choice of the workstation for MD simulations using GROMACS. It is supposed to study systems consisting from 30-50k atoms, in particular, lipid bilayer models. Since the last generation of Intel and Ryzen CPU has made a big leap in performance, available workstation can’t be taken as a basis. Please, help to choose components of the workstation, particularly, GPU and CPU with total cost about $2000. What is better, to buy two GPUs(sli) or one GPU but more effective? -- Best rerards, Olga Selyutina -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.