Unity Partitions

General Access	Partition	# Nodes	Maximum Time Limit	Default Time Limit	Notes
check	arm-gpu	1	14 days	1 hour	7
check	arm-preempt	3	14 days	1 hour	4 6 7
check	cpu	179	2 days	1 hour	7
check	cpu-preempt	134	2 days	1 hour	6 7
check	gpu	197	2 days	1 hour	7
check	gpu-preempt	209	2 days	1 hour	6 7
check	power9	2	30 days	1 hour	5
check	power9-gpu	16	30 days	1 hour	5 7
check	power9-gpu-preempt	16	2 days	1 hour	5 6 7
	astroth-cpu	9	30 days	1 hour
	astroth-gpu	3	30 days	1 hour
	cbio-cpu	5	30 days	1 hour
	cbio-gpu	1	30 days	1 hour
	ceewater_casey-cpu	10	30 days	1 hour
	ceewater_cjgleason-cpu	10	30 days	1 hour
	ceewater_kandread-cpu	10	30 days	1 hour
	ece-gpu	2	5 days	1 hour
	fsi-lab	1	30 days	1 hour
	gaoseismolab-cpu	5	30 days	1 hour
	gypsum-1080ti	42	7 days	1 hour	3
	gypsum-2080ti	21	7 days	1 hour	3
	gypsum-m40	23	7 days	1 hour	3
	gypsum-rtx8000	8	7 days	1 hour	3
	gypsum-titanx	51	7 days	1 hour	3
	superpod-a100	12	30 days	1 hour
	gpupod-l40s	7	14 days	1 hour
	ials-gpu	33	14 days	1 hour
	lan	1	30 days	1 hour
	mpi	5	1 day	1 hour
	power9-gpu-osg	2	30 days	1 hour	5
	toltec-cpu	7	30 days	1 hour
	umd-cscdr-arm	3	30 days	1 hour	1 4
	umd-cscdr-cpu	35	30 days	1 hour	1
	umd-cscdr-gpu	2	30 days	1 hour	1
	uri-cpu	30	30 days	1 hour	2
	uri-gpu	6	30 days	1 hour	2
	uri-richamp	5	Unlimited	1 hour
	visterra	8	30 days	1 hour
	zhoulin-cpu	6	30 days	1 hour

Restricted to members of CSCDR at UMass Dartmouth.
Restricted to users affiliated with the University of Rhode Island.
Restricted to members of CICS at UMass Amherst. Additional resource restriction apply based on status in CICS, see the Gypsum page for details.
Nodes have the ARM architecture, most software and modules will not work unless built for ARM.
Nodes have the POWER9 architecture, most software and modules will not work unless built for POWER9.
Nodes in preempt partitions belong to specific groups which would normally be inaccessible to general users. Priority users reserve the right to kill (preempt) any job running on their nodes to make room for their own jobs. When you schedule a job on a preempt partition, your job can be killed after two hours.
Use --qos=long to get up to 14 days.