Differences

This shows you the differences between two versions of the page.

--- grid [2017/09/27 14:01]
popel [Rules]
+++ grid [2024/10/02 15:21] (current)
popel
@@ Line 1: / Line 1: @@
 ====== ÚFAL Grid Engine (LRC) ======
-LRC (Linguistic Research Cluster) is a name of ÚFAL's computational grid/cluster, which has (as of 2017/09) about 1600 CPU cores (115 servers + 2 submission heads), with a total 10 TiB of RAM. It uses [[https://en.wikipedia.org/wiki/Oracle_Grid_Engine|(Sun/Oracle/Son of) Grid Engine]] software (SGE) for job scheduling etc. You can submit many computing tasks (jobs) at once, they will be placed in a queue and once a machine (slot) with the required capabilities (e.g. RAM, number of cores) is available, your job will be executed (scheduled) on this machine. This way we can maximize the usefulness of the computing resources and divide it among users in a fair way.
+LRC (Linguistic Research Cluster) is a name of ÚFAL's computational grid/cluster, which has (as of 2018/06) about 1728 CPU cores (65 servers + 10 submission heads), with a total 7.2 TiB of RAM. It uses [[https://en.wikipedia.org/wiki/Oracle_Grid_Engine|(Sun/Oracle/Son of) Grid Engine]] software (SGE) for job scheduling etc. You can submit many computing tasks (jobs) at once, they will be placed in a queue and once a machine (slot) with the required capabilities (e.g. RAM, number of cores) is available, your job will be executed (scheduled) on this machine. This way we can maximize the usefulness of the computing resources and divide it among users in a fair way.
-If you need GPU processing, see a special page about our [[:gpu|GPU cluster called DLL]] (which is actually a subsystem of LRC with independent queue ''gpu.q'').
+If you need GPU processing, see a special page about our [[:gpu|GPU cluster called DLL]] (which is actually a subsystem of LRC with an independent queue ''gpu-ms.q'').
+TODO: describe alternatives, e.g.: MetaCentrum / Cesnet cluster (all MFF students can use it), Amazon EC2, Microsoft Azure (some colleagues may have sometime free vouchers).
+**TODO: IN 2022 MOVING FROM SGE TO SLURM** (see [[slurm|slurm guidelines]]) -- **commands like ''qsub'' and ''qstat'' will stop working!**
+**IN 2024: Newly, all the documentation is at a dedicated wiki https://ufal.mff.cuni.cz/lrc (you need to use ufal and [[internal:welcome-at-ufal#small-linguistic-password|small-linguistic password]] to access the wiki from outside of the UFAL network).***
 ===== List of Machines =====
-Last update: 2017/09. All machines have Ubuntu 14.04.
+Last update: 2017/06. All machines have Ubuntu 18.04.
 Some machines are at Malá Strana (ground floor, new server room built from Lindat budget), some are at Troja (5 km north-east).
 If you need to quickly distinguish which machine is located where, you can use your knowledge of [[https://en.wikipedia.org/wiki/Trojan_War|Trojan war]]-related heroes, ''qhost -q'', or the tables below.
-==== Troja (troja-all.q) ====
+====== AVX instructions ======
-^ Name                ^ CPU type            ^ GHz ^cores ^RAM(GB)^ note ^
-| achilles[1-8]       | Intel               | 3.2 |   32 |  128 |  |
-| hector[1-8]         | Intel               | 1.3 |   32 |  128 |  |
-| helena[1-8]         | Intel               | 2.6 |   32 |  128 |  |
-| paris[1-8]          | Intel               | 2.4 |   32 |  128 |  |
-==== MS = Malá Strana (ms-all.q) ====
+==== Troja (cpu-troja.q) ====
+^ Name                ^ CPU type                                  ^ GHz ^cores ^RAM(GB)^ note ^
+| achilles[1-8]       | Intel(R) Xeon(R) CPU E5-2630 v3           | 2.4 |   31 |  123  |  AVX enabled    |
+| hector[1-8]         | Intel(R) Xeon(R) CPU E5-2630 v3           | 2.4 |   31 |  123  |  AVX enabled    |
+| helena[1-8]         | Intel(R) Xeon(R) CPU E5-2630 v3           | 2.4 |   31 |  123  |  AVX enabled    |
+| paris[1-8]          | Intel(R) Xeon(R) CPU E5-2630 v3           | 2.4 |   31 |  123  |  AVX enabled    |
-^ Name                ^ CPU type            ^ GHz ^cores ^RAM(GB)^ note ^
+==== MS = Malá Strana (cpu-ms.q) ====
-| andromeda[1-13]     | AMD 2xCore4 Opteron | 2.8 |    8 |   32 |  |
-| hydra[1-4]          | AMD                 | 2.6 |   16 |  128 |  |
-| fireball[1-10]      |Intel ??2xCore4 Xeon | 3.0 |    4 |   32 |  |
-| hyperion[1-9]       | Intel 2xCore2 Xeon  | 3.0 |    4 |   32 |  |
-| lucifer[1-10]       |Intel ??2xCore4 Xeon | 2.4 |   16 |  128 |  |
-| orion[1-6]          | Intel 2xCore4 Xeon  | 2.3 |    8 |   32 |  |
-| orion[7-10]         | Intel               | 2.0 |    4 |   32 |  |
-| tauri[1-10]         |Intel ??2xCore4 Xeon | 3.0 |    4 |   32 |  |
-| cosmos              | Intel 4xCore2 Xeon  | 2.9 |    8 |  256 |  |
-| belzebub            | Intel 8xCore4 Xeon  | 2.9 |   32 |  256 |  |
-| iridium             |Intel ??2xCore4 Xeon | 1.9 |   16 |  512 | also in ''gpu.q'' |
-| twister[1,2]        | Intel 2xCore4 Xeon  | 2.4 |    8 |   48 | also in ''gpu.q'' |
-=== Outside LRC cluster (but located as MS) ===
+^ Name                ^ CPU type and flags              ^ GHz ^cores ^RAM(GB)^ note ^
-^ Name                ^ CPU type            ^ GHz ^cores ^ RAM(GB)^ note  ^
+| hydra[1-4]          | AMD Opteron SSE4 AVX            | 2.6 |   15 |  122  |   AVX enabled   |
-| lrc[1,2]            | Intel               | 2.3 |    4 |   45 | **no computing here**, just submit jobs |
+| orion[1-8]          | Intel(R) Xeon(R) CPU E5-2630 v4 | 2.2 |   39 |  122  |   AVX enabled   |
-| pandora[1-10]       | 2xCore2 Intel Xeon  | 2.6 |      |   16 | special cluster&queue ''ms-guests.q''   |
+| belzebub            | Intel Xeon SSE4 AVX             | 2.9 |   31 |  249  |   AVX enabled   |
-| sol[1-5]            | Intel               | 2.6 |    4 |   16 | you can ssh here and compute |
+| iridium             | Intel Xeon SSE4                 | 1.9 |   15 |  501  |      |
-| sol[6-8]            | Intel               | 2.0 |    8 |   16 | you can ssh here and compute |
-The two **lrc machines** are so called heads of the cluster. **No computation is allowed here**, i.e. no CPU-intensive, disk-intensive nor RAM-intensive computation (very simple scripts are OK). You should just ssh to ''lrc1'' or ''lrc2'' and submit your jobs as described bellow.
+<html><!--
+sol5       fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow
+sol7       fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow
+andromeda4 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate npt lbrv svm_lock nrip_save vmmcall
+hydra1     fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid amd_dcm aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core perfctr_nb arat cpb hw_pstate npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold vmmcall
+fireball1  fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow
+hyperion1  fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow
+lucifer1   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 popcnt aes lahf_lm kaiser tpr_shadow vnmi flexpriority ept vpid dtherm ida arat
+orion1     fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow vnmi flexpriority
+orion7     fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow
+tauri1     fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow
+cosmos     fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca lahf_lm dtherm tpr_shadow vnmi flexpriority
+belzebub   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx lahf_lm epb kaiser tpr_shadow vnmi flexpriority ept vpid xsaveopt dtherm ida arat pln pts
+iridium    fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca sse4_1 sse4_2 x2apic popcnt lahf_lm dtherm tpr_shadow vnmi flexpriority ept vpid
+twister1   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni dtes64 monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 popcnt lahf_lm ida arat epb dtherm tpr_shadow vnmi flexpriority ept vpid
+-->
+</html>
+=== Submit hosts / test machines ===
+^ Name                ^ CPU type                        ^ GHz ^cores ^ RAM(GB) ^ note  ^
+| sol[1-8]           | Intel(R) Xeon(R) CPU E5-2630 v3 | 2.4 |    7 |   28    | you can ssh here and compute or submit jobs |
+| lrc[12]             | Intel(R) Xeon(R) CPU E5-2630 v4 | 2.2 |    4 |    4    | you can submit jobs here or monitor job execution - NO COMPUTATION IS ALLOWED HERE !!! |
-Alternatively, you can ssh to one of the **sol machines** and submit jobs from here. It is allowed to compute here, which is useful e.g. when you have a script which submits your jobs, but it also collects statistics from the jobs outputs (and possibly submits new jobs conditioned on the statistics). However, the sol machines are relatively slow and may be occupied by your colleagues, so for bigger (longer) tasks, always prefer submission as separate jobs.
+You can ssh to one of the **sol machines** and submit jobs from here. It is allowed to compute here, which is useful e.g. when you have a script which submits your jobs, but it also collects statistics from the jobs outputs (and possibly submits new jobs conditioned on the statistics). However, the sol machines are relatively slow and may be occupied by your colleagues, so for bigger (longer) tasks, always prefer submission as separate jobs.
-The **pandora machines** are in a special cluster (not accessible from lrc) and queue **ms-guests.q** available for our colleagues from KSVI and for students of [[http://ufal.mff.cuni.cz/courses/npfl102|Data intensive computing]] (see the 2016 handouts if you missed the course).
 ===== Installation =====
@@ Line 50: / Line 60: @@
 Add the following line into your '~/.bash_profile'.
-  source /net/projects/SGE/user/sge_profile
+  source /opt/LRC/sge_profile
+Or call one of these scripts directly:
+  /opt/LRC/common/settings.sh (for bash)
+  /opt/LRC/common/settings.csh (for tcsh/csh)
-This detects if you are on one of the cluster machines (including lrc and sol) and sets env variables accordingly. It also prints a status message.
+This detects if you are on one of the cluster machines (including sol) and sets env variables accordingly. It also prints a status message.
 Usually, this is the first line of your '~/.bash_profile' and the second-and-last line is
   [ -f ~/.bashrc ] && source ~/.bashrc
+Make sure you have correctly configured locale (otherwise ''qrsh'' may not show accented letters in ''less'' and you may get errors when printing utf8 on stdout/stderr from your script in ''qsub''). For example add the following line to your ''~/.bashrc'':
+  export LC_ALL=en_US.UTF-8
+If you are curious about purpose of .bashrc and .bash_profile and you need to know when they should be used you may read [[https://stackoverflow.com/a/415444|this]].
 ===== Basic usage =====
-First, you need to ssh to the cluster head (lrc1 or lrc2) or to one of the sol machines. The full address is ''lrc1.ufal.hide.ms.mff.cuni.cz'', but you can use just ''ssh lrc1'' ("hide" means it is accessible only from the ÚFAL network, not from outside; if working from home/Eduroam, you need to [[internal:remote-access|login/VPN]] to the ÚFAL network first).
+First, you need to ssh to one of the submit hosts (sol[1-10]). The full address is (for example) ''sol1.ufal.hide.ms.mff.cuni.cz'', but you can use just ''ssh sol1'' ("hide" means it is accessible only from the ÚFAL network, not from outside; if working from home/Eduroam, you need to [[internal:remote-access|login/VPN]] to the ÚFAL network first).
+In the following tutorial, we will prepare a wrapper shell script ''script.sh'' with a toy task. In practice you can name the script whatever you want and you can execute the real task, e.g. a Python/Perl/... script. It is recommended to use the wrapper shell scripts, but with ''-b y'' (see [[#advanced usage]]) you can execute a Python/Perl/... directly without any wrapper.
 <code>
-ssh lrc1
+ssh sol1
 echo 'hostname; pwd; echo The second parameter is $2' > script.sh
   # prepare a shell script describing your task
 qsub -cwd -j y script.sh Hello World
-  # This submits your job to the default queue, which is currently ''ms-all.q''.
+  # This submits your job to the default queue, which is currently ''cpu-*.q''.
   # Usually, there is a free slot, so the job will be scheduled within few seconds.
   # We have used two handy qsub parameters:
   #  -cwd  ... the script is executed in the current directory (the default is your home)
-  #  -j y  ... stdout and stderr outputs are merged and redirected to a file (''script.sh.o*'')
+  #  -j y  ... stdout and stderr outputs are merged and redirected to a file (''script.sh.o$JOB_ID'')
   # We have also provided two parameters for our script "Hello" and "World".
   # The qsub prints something like
@@ Line 112: / Line 134: @@
 The purpose of these rules is to prevent your jobs to damage the work of your colleagues and to divide the resources among users in a fair way.
-  * Read about our [[internal:linux-network|network]] first (so you know that e.g. reading big data from your home in 200 parallel jobs is not a good idea). Ask your colleagues (possibly via [[internal:mailing-lists|devel]]) if you are not sure, esp. if you plan to submit jobs with unusual/extreme disk/mem/CPU requirements.
+  * Read about our [[internal:linux-network|network]] first (so you know that e.g. reading big data from your home in 200 parallel jobs is not a good idea, but regular cleanup of your data is a good idea). Ask your colleagues (possibly via [[internal:mailing-lists|devel]]) if you are not sure, esp. if you plan to submit jobs with unusual/extreme disk/mem/CPU requirements.
   * While your jobs are running (or queued), check your jobs (esp. previously untested setups) and your email (esp. [[internal:mailing-lists|devel]]) regularly. If you really need to leave e.g. for two-week vacation offline, consult it first with it@ufal (whether they can kill your jobs if needed).
   * You can ssh to any cluster machine, which can be useful e.g. to diagnose what's happening there (using ''htop'' etc.).
@@ Line 118: / Line 140: @@
   * For interactive work, you can use ''qrsh'', but please try to end the job (exit with Ctrl+D) once finished with your work, especially if you ask for a lot of memory or CPUs (see below). One semi-permanent qrsh job (with non-extreme CPU/mem requirements) per user is acceptable.
   * **Specify the memory and CPU requirements** (if higher than the defaults) and **don't exceed them**.
-    * If your job needs more than one CPU (on a single machine) for most of the time, reserve the given number of CPU cores (and SGE slots) with <code>qsub -pe smp <number-of-CPU-cores></code> (As you can see in [[#List of Machines]], the maximum is 32 cores). If your job needs e.g. up to 110% CPU most of the time and just occasionally 200%, it is OK to reserve just one core (so you don't waste).
+    * If your job needs more than one CPU (on a single machine) for most of the time, reserve the given number of CPU cores (and SGE slots) with <code>qsub -pe smp <number-of-CPU-cores></code> As you can see in [[#List of Machines]], the maximum is 32 cores. If your job needs e.g. up to 110% CPU most of the time and just occasionally 200%, it is OK to reserve just one core (so you don't waste). TODO: when using ''-pe smp -l mf=8G,amf=8G,h_vmem=12G'', which memory limits are per machine and which are per core?
-    * If you are sure your job needs less than 1GB RAM, then you can skip this. Otherwise, if you need e.g. 8 GiB, you must always use ''qsub'' (or ''qrsh'') with ''-l mem_free=8G''. You should specify also ''act_mem_free'' with the same value and ''h_vmem'' with possibly a slightly bigger value. See [[#memory]] for details. TL;DR: <code>qsub -hard -l mem_free=8G,act_mem_free=8G,h_vmem=12G</code>
+    * If you are sure your job needs less than 1GB RAM, then you can skip this. Otherwise, if you need e.g. 8 GiB, you must always use ''qsub'' (or ''qrsh'') with ''-l mem_free=8G''. You should specify also ''act_mem_free'' with the same value and ''h_vmem'' with possibly a slightly bigger value. See [[#memory]] for details. TL;DR: <code>qsub -l mem_free=8G,act_mem_free=8G,h_vmem=12G</code>
-  * Be kind to your colleagues. If you are going to submit jobs that effectively take more than one fifth of our cluster for more than several hours, check if it is free (with ''qstat -g c'' or ''qstat -u \*''), ask your colleagues. Note that if you allocate one slot (CPU core) on a machine, but (almost) all its RAM, you have effectively occupied the whole machine and all its cores.
+  * Be kind to your colleagues. If you are going to submit jobs that effectively occupy **more than one fifth of our cluster for more than several hours**, check if the cluster is free (with ''qstat -g c'' or ''qstat -u \*'') and/or ask your colleagues if they don't plan to use the cluster intensively in the near future. Note that if you allocate one slot (CPU core) on a machine, but (almost) all its RAM, you have effectively occupied the whole machine and all its cores. If you are submitting **more than 100 jobs**, consider using setting them a low priority (e.g. ''-p -1024'', see below) or use [[#qunhold]] or (even better) [[#array jobs]].
+  * **Don't submit more than ca 5000 jobs at once**, even if you make sure that at most 100 are running/waiting and the rest is in the //hold// state (e.g. using ''qunhold''). More than 5000 jobs in the queue can overload the SGE, so then no one can execute ''qstat'' (or it takes too long).
-Další doporučení:
-  * Uklízet po sobě lokální data, protože jinak si tam už nikdo nic užitečného nepustí.
-  * Vyhnout se hodně divokému paralelnímu přístupu ke sdíleným diskům. NFS server to pak nepěkně zpomalí pro všechny. Distribuujte tedy i data.
-  * Pokud chci spouštět úlohy, které poběží dlouhou dobu (hodiny, dny), nepustím je všechny najednou, aby cluster mohli využívat i ostatní.
 === Memory ===
-mem_free (or mf): this is a 'consumable resource' tracked by SGE.
+  * There are three commonly used options for specifying memory requirements: ''mem_free, act_mem_free'' and ''h_vmem''. Each has a different purpose.
-  It affects job scheduling. Every machine as an initial value assigned.
+  * ''mem_free=1G'' means 1024×1024×1024 bytes, i.e. one [[https://en.wikipedia.org/wiki/Gibibyte|GiB (gibibyte)]]. ''mem_free=1g'' means 1000×1000×1000 bytes, i.e. one gigabyte. Similarly for the other options and other prefixes (k, K, m, M).
-  When you specify
+  * **mem_free** (or mf) specifies a //consumable resource// tracked by SGE and it affects job scheduling. Each machine has an initial value assigned (slightly lower than the real total physical RAM capacity). When you specify ''qsub -l mem_free=4G'', SGE finds a machine with mem_free >= 4GB, and subtracts 4GB from it. This limit is not enforced, so if a job exceeds this limit, **it is not automatically killed** and thus the SGE value of mem_free may not represent the real free memory. The default value is 1G. By not using this option and eating more than 1 GiB, you are breaking the rules.
-    qsub -l mem_free=4G
+  * **act_mem_free** (or amf) is a ÚFAL-specific option, which specifies the real amount of free memory (at the time of scheduling). You can specify it when submitting a job and it will be scheduled to a machine with at least this amount of memory free. In an ideal world, where no jobs are exceeding their ''mem_free'' requirements, we would not need this options. However, in real world it is recommended to use this option with the same value as ''mem_free'' to protect your job from failing with out-of-memory error (because of naughty jobs of other users).
-  SGE finds a machine with mem_free >= 4GB, and subtracts 4GB from it.
+  * **h_vmem** is equivalent to setting ''ulimit -v'', i.e. it is a hard limit on the size of virtual memory (see RLIMIT_AS in ''man setrlimit''). If your job exceeds this limit, memory allocation fails (i.e., malloc or mmap will return NULL), and your job will probably crash on SIGSEGV. TODO: according to ''man queue_conf'', the job is killed with SIGKILL, not with SIGSEGV. Note that ''h_vmem'' specifies the maximal size of **allocated_memory, not used_memory**, in other words it is the VIRT column in ''top'', not the RES column. SGE does not use this parameter in any other way. Notably, job scheduling is not affected by this parameter and therefore there is no guarantee that there will be this amount of memory on the chosen machine. The problem is that some programs (e.g. Java with the default setting) allocate much more (virtual) memory than they actually use in the end. If we want to be ultra conservative, we should set ''h_vmem'' to the same value as ''mem_free''. If we want to be only moderately conservative, we should specify something like h_vmem=1.5*mem_free, because some jobs will not use the whole mem_free requested, but still our job will be killed if it allocated much more than declared. The default effectively means that your job has no limits.
+    * For GPU jobs, it is usually better to use **h_data** instead of **h_vmem**. CUDA driver allocates a lot of "unused" virtual memory (tens of GB per card), which is counted in ''h_vmem'', but not in ''h_data''. All usual allocations (''malloc'', ''new'', Python allocations) seem to be included in ''h_data''.
+  * It is recommended to **profile your task first** (see [[#profiling]] below), so you can estimate reasonable memory requirements before submitting many jobs with the same task (varying in parameters which do not affect memory consumption). So for the first time, declare mem_free with much more memory than expected and ssh to a given machine and check ''htop'' (sum all processes of your job) or (if the job is done quickly) check the epilog. When running other jobs of this type, set ''mem_free'' (and ''act_mem_free'' and ''h_vmem'') so you are not wasting resources, but still have some reserve.
+  * **s_vmem** is similar to ''h_vmem'', but instead of SIGSEGV/SIGKILL, the job is sent a SIGXCPU signal which can be caught by the job and exit gracefully before it is killed. So if you need it, set ''s_vmem'' to a lower value than ''h_vmem'' and implement SIGXCPU handling and cleanup.
-  This limit is not enforced, so if a job exceeds this limit, the
-  SGE value of mem_free may not represent the real free memory.
-  Default value is 1GB.
 ===== Advanced usage =====
-Other useful commands:
+''qsub **-q** cpu-troja.q''
+This way your job is submitted to the Troja queue. The default is ''cpu-*.q''. You can also use e.g.
+''-q '(cpu-t*|cpu-m*)''' to submit on any machine in those two queues (but **don't use ''-q '*'''** as this includes also [[:gpu|gpu.q]]),
+''-q '*@hector[14]''' to submit on hector1 or hector4,
+''-q 'cpu-*@!(hector*|iridium)''' to submit on any troja/ms machine except hectors and iridium.
+However, usually you should specify just the queue (cpu-troja.q vs. cpu-ms.q), not a particular machine, and instead use ''-l'' to specify the needed resources in a general way.
-<code>
+''qsub **-l** ...''
-qsub -o LOG.stdout -e LOG.stderr skript.sh
+See ''man complex'' (run it on lrc or sol machines) for a list of possible resources you may require (in addition to ''mem_free'' etc. discussed above).
-  # když chcete přesměrovat výstup skriptu do určených souborů
-qsub -S /bin/bash
-  # když chcete, aby skript běžel v bashi
-qsub -V
-  # když chcete předat proměnné prostředí
-qdel \*
-  # když chcete zrušit všechny své joby (rušit cizí nesmíte)
-</code>
+''qsub **-p** -200''
+Define a priority of your job as a number between -1024 and 0. Only SGE admins may use a number higher than 0. In January 2018, we changed the default to -100 (it used to be 0). Please, do not use priority between -99 and 0 for jobs taking longer than a few hours, unless it is absolutely necessary for a deadline. In that case, please notify other GPU users. You should ask for lower priority (-1024..-101) if you submit many jobs at once or if the jobs are not urgent. SGE uses the priority to decide when to start which pending job in the queue (it computes a real number called ''prior'', which is reported in ''qstat'', which grows as the job is waiting in the queue). Note that once a job is started, you cannot "unschedule" it, so from that moment on, it is irrelevant what was its priority.
+''qsub **-o** LOG.stdout **-e** LOG.stderr''
+redirect std{out,err} to separate files with given names, instead of the defaults ''$JOB_NAME.o$JOB_ID'' and ''$JOB_NAME.e$JOB_ID''.
-==== ~bojar/tools/shell/qsubmit ====
+''qsub **-@** optionfile''
+Instead of specifying all the ''qsub'' options on the command line, you can store them in a file (you can use # comments in the file). See also [[#In-script options]].
-qsubmit je jako qsub, ale příjemnější:
+''qsub **-a** 12312359''
+Execute your job no sooner than at the given time (in ''[YY]MMDDhhmm'' format). An alternative to ''sleep 3600 && qsub ... &''.
-  * nemusíte vyrábět skript, vyrobí ho sám (pozn.: nemusíte vyrábět skript, když použijete přepínač ''-b y'')
+''qsub **-b** y''
-  * nemusíte připisovat ''-cwd -j y -S /bin/bash''
+Treat ''script.sh'' (or whatever is the name of the command you execute) as a binary, i.e. don't search for [[#in-script options]] within the file, don't transfer it to the qmaster and then to the execution node. This makes the execution a bit faster and it may prevent some rare but hard-to-detect errors caused SGE interpreting the script. The script must be available on the execution node via NFS, Lustre (which is our case), etc. With ''-b y'' (shortcut for ''-b yes''), ''script.sh'' can be an executable script or a binary (and you must provide full path, e.g. ''./script.sh''). With ''-b n'' (which is the default for ''qsub''), ''script.sh'' must be a script (text file).
-<code>
+''qsub **-M** popel@ufal.mff.cuni.cz,rosa@ufal.mff.cuni.cz **-m** beas''
-~bojar/tools/shell/qsubmit "bashovy_prikaz < prismeruj > presmeruj 2> atd..."
+Specify the emails where you want to be notified when the job has been **b** started, **e** ended, **a** aborted or rescheduled, **s** suspended.
-</code>
+The default is now ''-m a'' and the default email address is forwarded to you (so there is no need to use ''-M''). You can use ''-m n'' to override the defaults and send no emails.
-lépe funguje ''~{stepanek,pajas}/bin/qcmd'' (nemusí se kvotovat parametry, správně počítá čas běhu...)
+''qsub **-hold_jid** 121144,121145'' (or ''qsub **-hold_jid** get_src.sh,get_tgt.sh'')
+The current job is not executed before all the specified jobs are completed.
-===== Monitorování úloh =====
+''qsub **-now** y''
+Start the job immediately or not at all, i.e. don't put it as pending to the queue. This is the default for ''qrsh'', but you can change it with ''-now n'' (which is the default for ''qsub'').
-  * ''qstat [-u user]'' -- seznam úloh aktuálně běžících / ve frontě
+''qsub **-N** my-name''
-  * ''qhost'' -- dostupné zdroje
+By default the name of a job (which you can see e.g. in ''qstat'') is the name of the ''script.sh''. This way you can override it.
-  * ''/SGE/REPORTER/LRC-UFAL/bin/lrc_users_real_mem_usage -u user -w'' -- aktuální využití paměti uživatelem
-  * ''/SGE/REPORTER/LRC-UFAL/bin/lrc_users_limits_requested -w'' -- nárokované požadavky uživatelů
-  * ''/SGE/REPORTER/LRC-UFAL/bin/lrc_nodes_meminfo'' -- vypis vsech uzlu a stav vytiznosti pameti.
-    * mem_total: celkova pamet uzlu
-    * mem_free: tedy kolik je jeste volne pameti z pametove quoty uzlu
-    * act_mem_free: kolik uzlu OPRAVDU zbyva volne pameti
-    * mem_used: kolik je pameti skutecne pouzito
-  * ''/SGE/REPORTER/LRC-UFAL/bin/lrc_state_overview'' -- celkový přehled o clusteru
-    * celkovy pocet jader, pocet vyuzitych jader
-    * celkova velikost RAM, kolik je ji fyzicky nepouzite, kolik je ji jeste nerezervovane
-    * po jednotlivych uzivatelich (zrovna pocitajicich) -- kolik jim bezi uloh, kolik jich maji ve fronte a kolik z nich je ve stavu hold
-  * ''cat /SGE/REPORTER/LRC-UFAL/stats/userlist.weight'' -- seznam uživatelů clusteru seřazený podle dosavadní aktivity (počet odeslaných úloh × čas, který běžely), aktualizovaný každý den v&nbsp;noci
-  * [[https://ufaladm2.ufal.hide.ms.mff.cuni.cz/munin/ufal.hide.ms.mff.cuni.cz/lrc1.ufal.hide.ms.mff.cuni.cz/lrc_users.html|Munin: graf vytíženosti clusteru podle uživatelů]] (viditelný pouze ze sítě ÚFAL)
-===== Časté a záludné problémy =====
+''qsub **-S** /bin/bash''
+The hashbang (''!#/bin/bash'') in your ''script.sh'' is ignored, but you can change the interpreter with ''-S''. I think ''/bin/bash'' is now (2017/09) the default (but it used to be ''csh'').
+''qsub **-v** PATH[=value]''
+Export a given environment variable from the current shell to the job.
-==== Submitnutý job může znovu submitovat ====
+''qsub **-V**''
+Export all environment variables. (This is not so needed now, when bash is the default interpreter and it seems your ''~/.bashrc'' is always sourced.)
-Danovy starší zkušenosti s clusterem PBS (nikoli SGE) říkaly, že tohle nejde. Ale jde to, aspoň u nás. Příkazy ''qsub'' a spol. jsou kromě hlavy clusteru dostupné i na všech strojích clusteru, samozřejmě pokud váš soubor ''.bashrc'', ''.cshrc'' apod. zajistí, že se i na nich provede inicializace prostředí SGE.
+''qsub **-soft** -l ... **-hard** -l ... -q ...''
+By default, all the resource requirements (specified with ''-l'') and queue requirements (specified with ''-q'') are //hard//, i.e. your job won't be scheduled unless they can be fulfilled. You can use ''-soft'' to mark all following requirements as nice-to-have. And with ''-hard'' you can switch back to hard requirements.
+''qsub **-sync** y''
+This causes qsub to wait for the job to complete before exiting (with the same exit code as the job). Useful in scripts.
+''**qalter**''
+You can change some properties of already submitted jobs (both waiting in the queue and running). Changeable properties are listed in ''man qsub''.
-==== Proměnné prostředí, nastavení vlastního prostředí ====
+''[[https://gridscheduler.sourceforge.net/htmlman/htmlman1/qsub.html|man qsub]] qstat qalter qhold queue_conf sge_types complex''
+Find out all the gory details which are missing here. You'll have to do it one day anyway:-).
-SGE spouští skripty v čistém prostředí. Nebuďte proto překvapeni, když vám skript na konzoli poběží dobře, ale po submitnutí fungovat nebude. Třeba nenašel potřebné programy v ''$PATH''
+=== qsub wrappers ===
-Zatím nevím přesně, které ze souborů ''.login'', ''.bashrc'' ap. SGE spouští, jestli vůbec nějaké. Naopak, experimentálně jsem ověřil, že ''qsub -S /bin/bash skript'' nenačte žádný z ''.bashrc'', ''.bash_profile'', ''.login'', ani ''.profile''.
+If you often run (ad-hoc) bash commands via ''qsub'', check ''~bojar/tools/shell/**qsubmit**'' or ''~stepanek/bin/**qcmd**'', which allow you to enter the command on command line without creating any temp script files. The wrappers have also other features (some qsub options have changed default values). ''qcmd'' is older, but unlike ''qsubmit'' it has POD documentation, correct time computation and you don't need to quote the command.
-Z toho například také vyplývá, že bez ošetření se jako **Java** používá
+=== qunhold ===
+''~stepanek/bin/qunhold'' tries to keep the number of running SGE jobs under a given threshold: all jobs over the threshold are held. If the number of running jobs goes below the threshold (default: 100), 10 jobs (by default) are unheld. Beware: if your jobs submit new jobs, you can get far over the threshold!
-   java version "1.5.0"
+Don't submit more than ca 5000 jobs with qunhold - it overloads the SGE queue and slows done e.g. ''qstat'' (and qunhold uses ''qstat'' internally).
-   gij (GNU libgcj) version 4.1.2 20070502 (Red Hat 4.1.2-12)
-Pokud chcete submittovaný program pouštět ve svém oblíbeném prostředí (např. nastavení ''PATH''), musíte v obalujícím skriptu příslušné ''.bash*'' načíst. Vždy je ale bezpečnější všude psát plné cesty, než spoléhat na PATH.
+=== sshcwd ===
+This is useful not only when sshing to sol machines. Add the following lines to your ''~/.bashrc''.
-==== Jiný shell ====
+<code>
+function sshcwd () {
+  # save the current history so that it is available
+  # immediately on the remote machine
+  history -a;
+  # setup the working directory by setting WD, delete possible ".nfs/"
+  ssh -X -Y -C -t $@ "WD='${PWD/.nfs\//}' /bin/bash --login -i";
+}
-Abych mohl poslat nějakou úlohu do fronty, musím pro ni vyrobit vlastní skript. Budiž, vyrobil jsem vlastní skript:
+# use WD to setup the working directory
+if [ -n "$WD" ]; then
+  echo "Autochanging dir to $WD" >&2
+  cd $WD;
+fi
-<code>
+alias sol1="sshcwd sol1.ufal.hide.ms.mff.cuni.cz"
-#!/bin/bash
-program > log.out 2> log.err
 </code>
-Když tento skript spustím, stane se očekávané. Přesměrují se výstupy z daného programu do souborů a je to.
+=== In-script options ===
-Když takový skript submitnu, program se **nespustí**. V logu zjistím, že (standardní chybový) výstup shellu, který pouštěl můj skript praví kryptickou zprávu "Ambiguous redirect".
+If you don't use ''-b y'', you can write the ''qsub'' options into ''script.sh'' instead of providing them on the command line. So your ''script.sh'' can be e.g.
-Nebudu vás napínat, zde je vysvětlení: SGE ignoruje první řádek skriptu (ve skutečnosti je pravda horší, hledá v něm nějaké parametry pro sebe) a spouští skript v ''csh''. Rozdíl mezi bashem a csh se v primitivních skriptech na první pohled nepozná, pozná se až v konstrukci if-then-else, a také v přesměrovávání. csh nerozumí přesměrování ''2>''
+  #$ -l mem_free=10G,act_mem_free=10G,h_vmem=10G
+  #$ -pe smp 8
+  #$ -cwd -j y
+  #$ -N name-of-my-job
+  ./what/to/run
-Takto SGE přinutíte, aby použilo bash:
+and you execute it now simply with ''qsub script.sh''.
-<code>
+=== ~/.sge_request ===
-  qsub -S /bin/bash skript
-</code>
-Jinou možností je přesměrovat stderr a stdout pomocí syntaxe csh:
+You can change the defaults for any option by creating a personal configuration file ''~/.sge_request''. For example, you can add there a line ''-m n'', so you will get no email notifications (unless overridden from the command line or in-script options).
-<code>
+=== Array jobs ===
-( command >stdout_file ) >&stderr_file
-</code>
+If you have a set of tasks (of the same type) and want to run them on multiple machines, use ''qsub -t''.
+  * ''-t 1-n'' start array job with //n// tasks numbered //1 ... n//
+  * environmental variable ''SGE_TASK_ID''
+  * output and error files ''$JOB_NAME.[eo]$JOB_ID.$TASK_ID''
+  * ''-t m-n[:s]'' start array job with tasks //m, m + s, ..., n//
+  * environmental variables ''SGE_TASK_FIRST, SGE_TASK_LAST, SGE_TASK_STEPSIZE''
+  * ''-tc j'' run at most //j// tasks simultaneously
+  * ''-hold_jid_ad comma_separated_job_list'' array jobs that must finish before this job starts; task //i// of the current job depends only on task //i// of the specified jobs
-==== bashrc a podobné nesmí nic vypisovat na konzoli ====
+If you use ''-tc'', then SGE can handle array jobs of virtually any size. It only starts as many tasks as specified in ''-tc'' at any time, and each scheduling interval (15 seconds in our current configuration) it starts new tasks if less than the specified ''-tc'' limit are running. However, note that it means the maximum throughput is //4 * tc// tasks per minute, so the individual array job tasks need to run for at least tens of seconds for this to be effective.
-Opsáno z [[http://www.sara.nl/userinfo/lisa/usage/batch/index.html]].
+The advantage of array jobs over [[#qunhold]] is that it does not overload the SGE job queue. Also if you start an array job, the others can see that it is an array job, how many individual tasks there are and how many of them have already finished.
-It is important, that the files that are sourced during a login such as .bash_profile .profile .bashrc .login .cshrc don't produce any output when a non-interactive login is done. If they do, changes are that your job will run, but that the batch system is unable to deliver to you the standard output and error files. In that case the status of your job will be 'E' after the job is finished. Here is an example how you can test in a .bash_profile or .bashrc if this is an interactive login:
+=== Delete many jobs at once ===
+For deleting all your jobs use ''qdel \*''. For deleting array jobs (see above), you can use comma-separated ranges of task ids, e.g. ''qdel -t 1,10,50-100''.
+For deleting a range of (normal) job ids, you can use bash expansion (as an alternative to ''seq''), e.g. ''qdel {17979..18028}''.
+=== Ssh to random sol ===
+Ondřej Bojar suggests to add the following alias to your .bashrc (cf. [[#sshcwd]]):
+<code>alias cluster='comp=$(( (RANDOM % 10) +1)); ssh -o "StrictHostKeyChecking no" sol$comp'</code>
+===== Job monitoring =====
+  * ''qstat [-u user]'' -- print a list of running/waiting jobs of a given user
+  * ''qhost'' -- print available/total resources
+  * ''qacct -j job_id'' -- print info even for ended job (for which ''qstat -j job_id'' does not work). See ''man qacct'' for more.
+  * ''/opt/LRC/REPORTER/LRC-UFAL/bin/lrc_users_real_mem_usage -u user -w'' -- current memory usage of a given user
+  * ''/opt/LRC/REPORTER/LRC-UFAL/bin/lrc_users_limits_requested -w'' -- required resources of all users
+  * ''/opt/LRC/REPORTER/LRC-UFAL/bin/lrc_nodes_meminfo'' -- memory usage of all nodes
+    * mem_total:
+    * mem_free: total memory minus reserved memory (using ''qsub -l mem_free'') for each node
+    * act_mem_free: really free memory
+    * mem_used: really used memory
+  * ''/opt/LRC/REPORTER/LRC-UFAL/bin/lrc_state_overview'' -- overall summary (with per-user stats for users with running jobs)
+  * ''cat /opt/LRC/REPORTER/LRC-UFAL/stats/userlist.weight'' -- all users sorted according to their activity (number of submitted jobs × their average duration), updated each night
+  * [[https://ufaladm2.ufal.hide.ms.mff.cuni.cz/munin/ufal.hide.ms.mff.cuni.cz/lrc-master.ufal.hide.ms.mff.cuni.cz/index.html|Munin: graph of cluster usage by day and user]] and  [[https://ufaladm2.ufal.hide.ms.mff.cuni.cz/munin/ufal.hide.ms.mff.cuni.cz/nfs-core.ufal.hide.ms.mff.cuni.cz/index.html|Munin monitoring of disk storage]] (both accessible only from ÚFAL network)
+===== Profiling =====
+As stated above, you should always specify the exact memory limits when running your tasks, so that you neither waste RAM nor starve others of memory by using more than you requested. However, memory requirements can be difficult to estimate in advance. That's why you should profile your tasks first.
+A simple method is to run the task and observe the memory usage reported in the epilog, but SGE may not record transient allocations. As documented in ''man 5 accounting'' and observed in ''qconf -sconf'', SGE only collects stats every ''accounting_flush_time''. If this is not set, it defaults to ''flush_time'', which is preset to 15 seconds. But the kernel records all info immediately without polling, and you can view these exact stats by looking into /proc/$PID/status while the task is running.
+You can still miss allocations made shortly before the program exits – which often happens when trying to debug why your program gets killed by SGE after exhausting the reserved space. To record these, use ''/usr/bin/time -v'' (the actual binary, not the shell-builtin command ''time''). Be aware that unlike the builtin, it cannot measure shell functions and behaves differently on pipelines.
+Obtaining peak usage of multiprocess applications is trickier. Detached and backgrounded processes are ignored completely by ''time -v'' and you get the maximum footprint of any children, not the sum of all maximal footprints nor the largest footprint in any instant.
+If you program in C and need to know the peak memory usage of your children, you can also use the ''wait4()'' syscall and calculate the stats yourself.
+If your job is the only one on a given machine, you can also look how much free memory is left when running the job (e.g. with ''htop'' if you know when is the peak moment).
+===== Other =====
+  * There is a **great course [[http://ufal.mff.cuni.cz/courses/npfl102|Data intensive computing]]**, see the 2016 handouts if you missed the course. It covers the usage of [[http://spark.apache.org/|Spark]] (MapReduce/Hadoop alternative, but better) and HDFS (Hadoop filesystem).
+  * **Note:** some hadoop basics and a lot of NoSQL technologies are covered by [[https://is.cuni.cz/studium/predmety/index.php?do=predmet&kod=NDBI040|Big Data Management and NoSQL Databases]]
+  * There is a special cluster for Mgr (and Bc) students (but not for PhD and UFAL members): http://aic.ufal.mff.cuni.cz/
+  * You can use environment variables ''$JOB_ID'', ''$JOB_NAME''.
+  * One job can submit other jobs (but be careful with recursive:-)). A job submitted to the CPU cluster may submit GPU jobs (to the ''qpu.q'' queue).
+  * It is important, that the files that are sourced during a login such as .bash_profile, .profile, .bashrc, .login etc. don't produce any output when a non-interactive login is done. If they do, chances are that your job will run, but that the batch system is unable to deliver to you the standard output and error files. In that case the status of your job will be 'E' after the job is finished. Here is an example how you can test in a .bash_profile or .bashrc if this is an interactive login:
 <code>
 unset INTERACTIVE
@@ Line 254: / Line 336: @@
 fi
 </code>
+TODO: Is this restriction still true (for our cluster)? E.g. .bash_profile with /net/projects/SGE/user/sge_profile prints info messages on stderr and it is OK.
-==== Jak zjistit, jaké zdroje jsem pro svou úlohu požadoval ====
-<code>qstat -j 973884,982737,984029,984030,984031,984034,984036 | grep resource
-hard resource_list:         mem_free=50g
-hard resource_list:         mem_free=200g
-hard resource_list:         mem_free=16g
-hard resource_list:         mem_free=16g
-hard resource_list:         mem_free=16g
-hard resource_list:         mem_free=31g</code>
-==== Jak rezervovat více jader na stejném stroji pro 1 job ====
-<code>
-qsub -pe smp <pocet jader>
-</code>

[ Back to the navigation ] [ Back to the content ]

Institute of Formal and Applied Linguistics Wiki

Differences