DavinciIntel Fortran
|
Accounts and Charging on DavinciIntroduction to MPP HoursWhen a job runs on a NERSC MPP system, such as Franklin, Bassi, or Jacquard, charges accrue against one of the user's repository allocations. A parallel job is charged for exclusive use of each multi-core/CPU node allocated to the job. The unit of accounting for these systems at NERSC is "MPP Hours." The MPP charge for a given job is calculated as the product of: (1) the job's wall-clock time in hours, (2) the number of CPUs allocated to the job (regardless of the number actually used), (3) a machine charge factor based on typical performance of the machine relative to the historical basis of the 375 MHz IBM Power3 architecture (MCF=1.0), and (4) and a job priority factor. Computing MPP Hours on DavinciDavinci's machine charge factor (MCF) is 4.0, and Davinci has 32 CPU's; Davinci is not a multi-node system. Interactive, debug and regular priority batch jobs have a job priority factor of 1.0. Low priority batch jobs have a job priority factor of 0.5. Premium priority batch jobs have a job priority factor of 2.0. Interactive charges are based on CPU time (not wallclock time). Thus if a user uses 1 hour of CPU time on Davinci, their default repository will be charged 4 MPP hours. Batch jobs are charged at the regular priority rate, with an MCF of 4.0. The MPP Hours charge for a given batch job is calculated by multiplying the job's "wall-clock time" by the number of CPUs, and a "machine charge" factor of 4. For example, if you run an 8 CPUs regular priority job that begins at 12:00:00 and ends 2 hours later at 14:00:00, your MPP charge is: 2 hours * 8 CPUs * 4 = 64 MPP Hours |
![]() |
Page last modified: Fri, 05 Oct 2007 18:05:36 GMT Page URL: http://www.nersc.gov/nusers/systems/davinci/accounts.php Web contact: webmaster@nersc.gov Computing questions: consult@nersc.gov Privacy and Security Notice |
![]() |