The HPC installation at CBS is a part of the
Danish Center for Scientific
Computing (DCSC). The resources available here
(see
picture)
comprise four multiprocessor,
shared memory servers by
Silicon Graphics
and a RAID storage system connected to the servers by a 1 Gbit fiberchannel
network (Storage Area Network, SAN). The detailed description is as follows.
| Name |
Model |
Specification |
Comments |
| genome |
Origin 300 |
32 500 MHZ IP35 Processors (MIPS)
CPU: R14000 Processor Chip Revision: 2.4
FPU: R14010 Floating Point Chip Revision: 2.4
Main memory size: 16384 Mbytes (shared memory)
Instruction cache size: 32 Kbytes
Data cache size: 32 Kbytes
Secondary unified instruction/data cache size: 2 Mbytes
Operating system: IRIX 6.5
|
CBS WWW server, offering a large number of public bioinformatics
prediction and computation services.
System administration tasks (NIS, backup, CXFS etc.) are run here as well.
The system is not available to users.
|
| cell |
Origin 300 |
18 500 MHZ IP35 Processors (MIPS)
CPU: R14000 Processor Chip Revision: 2.4
FPU: R14010 Floating Point Chip Revision: 2.4
Main memory size: 16384 Mbytes (shared memory)
Instruction cache size: 32 Kbytes
Data cache size: 32 Kbytes
Secondary unified instruction/data cache size: 2 Mbytes
Operating system: IRIX 6.5
|
This server is mainly used for running large parallel
memory intensive jobs.
This architecture is being phazed out as a bioinformatics platform!
|
| organism |
Origin 300 |
32 500 MHZ IP35 Processors (MIPS)
CPU: R14000 Processor Chip Revision: 2.4
FPU: R14010 Floating Point Chip Revision: 2.4
Main memory size: 32768 Mbytes (shared memory)
Instruction cache size: 32 Kbytes (on chip)
Data cache size: 32 Kbytes (on chip)
Secondary unified instruction/data cache size: 2 Mbytes
Operating system: IRIX 6.5
|
This server is mainly used for running large parallel
memory intensive jobs.
This architecture is being phazed out as a bioinformatics platform!
|
| life |
Altix 3000 |
64 900 MHz Itanium 2 model 0 Rev. 7 Processor (Intel)
CPU: Itanium 2 model 0 Rev. 7
FPU: Itanium 2 model 0 Rev. 7
Main memory size: 181.89 Gb (shared memory)
Instruction cache size: 32 Kbytes
Data cache size: 32 Kbytes
Secondary unified instruction/data cache size: 8 Mbytes
Operating system: SuSE Linux |
This server is mainly used for running large parallel
and extremely memory intensive jobs.
This server is binary compatible with Linux; however, the performance
of many programs can be improved significantly by compiling and
optimizing fusing the native Intel compilers available.
PBS
(Portable Batch System, professional edition) is installed here.
|
| ibiology |
Altix 3000 |
64 1300 MHz Itanium 2 model 1 Rev. 5 Processor (Intel)
CPU: Itanium 2 model 1 Rev. 5
FPU: Itanium 2 model 1 Rev. 5
Main memory size: 192 Gb (shared memory)
Instruction cache size: 32 Kbytes
Data cache size: 32 Kbytes
Secondary unified instruction/data cache size: 8 Mbytes
Operating system: SuSE Linux |
This server is mainly used for running large parallel
and memory intensive jobs.
This server is binary compatible with Linux; however, the performance
of many programs can be improved significantly by compiling and
optimizing fusing the native Intel compilers available.
PBS
(Portable Batch System, professional edition) is installed here.
|
| sbiology |
Altix BX2 |
128 1500 MHz Itanium 2 model 2 Rev. 1 Processor (Intel)
CPU: Itanium 2 model 2 Rev. 1
FPU: Itanium 2 model 2 Rev. 1
Main memory size: 768 Gb (shared memory)
Instruction cache size: 32 Kbytes
Data cache size: 32 Kbytes
Secondary unified instruction/data cache size: 8 Mbytes
Operating system: SuSE Linux |
This server is mainly used for running large parallel
and memory intensive jobs.
This server is binary compatible with Linux; however, the performance
of many programs can be improved significantly by compiling and
optimizing fusing the native Intel compilers available.
PBS
(Portable Batch System, professional edition) is installed here.
|
| interaction |
Altix 3000 |
32 1500 MHz Itanium 2 model 1 Rev. 5 Processor (Intel)
CPU: Itanium 2 model 1 Rev. 5
FPU: Itanium 2 model 1 Rev. 5
Main memory size: 118.60 Gb (shared memory)
Instruction cache size: 32 Kbytes
Data cache size: 32 Kbytes
Secondary unified instruction/data cache size: 8 Mbytes
Operating system: SuSE Linux |
This server is mainly used for interactive testing and compiling, as well
as for updatine and maintenance of the available data resources (datbases and databanks).
It is not supposed to be used for large batch jobs.
This server is binary compatible with Linux; however, the performance
of many programs can be improved significantly by compiling and
optimizing fusing the native Intel compilers available. The server
functions as MySQL server, and is used for web based services.
|
| /home/databases |
TP9600 RAID 5 |
TP9500 RAID 5 storage rack
122 x 120 Gb disks
2 x RAID 4 port 1 Gbit FC ctrl
13 Tb
Connected to the servers mentioned above by two 2 Gbit
fiberchannel network (Storage Area Network, SAN). |
The various non-relational databases (like BLAST'able sequence data) are stored here.
|
| /home/mysql |
TP9400 RAID 5 |
TP9400 RAID 5 storage rack
77 x 73 Gb disks
4 x RAID 4 port 1 Gbit FC ctrl
6.5 Tb
Connected directly to the mysql relational database server by two 1 Gbit
fiberchannel. |
The relational databases are stored here.
|
| /home/projects |
IS4000 RAID 5 |
IS4000 RAID 5 storage rack
300 Gb disks
4 x RAID 4 port 2 Gbit FC ctrl
50 Tb
Connected to the servers mentioned above by two 2 Gbit
fiberchannel network (Storage Area Network, SAN). |
The user (projects) data and the databases are stored here.
|
| /home/people |
RAID 5 |
RAID 5 storage rack
12 x 73 Gb disks
1 x RAID 1 port 1 Gbit FC ctrl
0.5 Tb
Connected to dedicated NFS server by a 1 Gbit fiberchannel. |
The home directories of all users are located here.
|
CBS operates a growing datawarehouse service. The page linked above describes
the current status of the databases, where to find them, and how they are
structured. The datawarehouse effort includes both storage of project data
and a large number of externally sourced databases, updated regularly.
Bioinformatics infrastructure
A large selection of bioinformatics software is installed on the servers; this
includes the standard bioinformatics tools available in the public domain and
the tools developed at CBS. The key programs are updated on regular basis.
All the software is available to the external (DCSC) users.