From c42c95b3996ab11cdd6af604ba5ed6fc1de01887 Mon Sep 17 00:00:00 2001 From: Jin Lee Date: Fri, 19 Aug 2016 18:49:42 -0700 Subject: [PATCH] small BDS tasks get queued faster on cluster engine (SGE) --- modules/general.bds | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/modules/general.bds b/modules/general.bds index 742dbd9..891e6b3 100644 --- a/modules/general.bds +++ b/modules/general.bds @@ -74,7 +74,7 @@ void init_general() { if ( use_system != "" ) system = use_system if ( system.toLower() == "slurm" ) system = "generic" - cpus = 1 + cpus = -1 // With cpus==-1, BDS does not pass number of threads to cluster engine (SGE, SLURM, ...), which means single-threaded mem = get_res_mem(memory,1) timeout = get_res_wt(wt) walltimeout = 3600*24*100 // timeout var. in BigDataScript (100 days, jobs will never be stopped by BDS due to BDS timeout)