Changes between Version 15 and Version 16 of AppPlan


Ignore:
Timestamp:
Dec 14, 2009, 10:33:43 PM (15 years ago)
Author:
davea
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • AppPlan

    v15 v16  
    3030   double ncudas;     // number of NVIDIA GPUs used
    3131   double natis;      // number of ATI GPUs used
     32   double gpu_ram;    // max amount of GPU RAM used
    3233   double avg_ncpus;  // avg #CPUs used by app (may be fractional)
    3334   double max_ncpus;  // max #CPUs used (not currently used for anything)
     
    4445The client uses '''flops''' to estimate job completion times.
    4546
     47If '''gpu_ram''' is nonzero, the BOINC client (6.10.25+)
     48won't start the app unless that much RAM is available on the allocated GPU.
     49
    4650You are free to define your own set of plan classes,
    4751and to link your own '''app_plan()''' function with the scheduler.
     
    5256 '''cuda''':: A CUDA application that requires 254MB of GPU RAM, and that uses .5% as many CPU FLOPS as GPU FLOPS.
    5357 '''cuda23''':: Similar, but requires a driver version supporting CUDA 2.3.
     58 '''ati''':: ATI GPU with Catalyst 8.12+ and DLLS named amd*
     59 '''ati13amd''':: ATI GPU  with Catalyst 9.1+ and DLLS named amd*
     60 '''ati13ati''':: ATI GPU  with Catalyst 9.1+ and DLLS named ati*
     61 '''ati14''':: ATI GPU with Catalyst 9.7+ and DLLS named ati*
    5462 '''nci''':: A non-CPU-intensive application that uses 1% of a CPU (this will cause the BOINC client 6.7+ to run it at non-idle priority).
    5563 '''sse3''':: A CPU app that requires the SSE3 CPU feature.