Last modified 5 months ago Last modified on 12/03/13 15:53:23

Specifying plan classes in XML

You can specify plan classes using an XML configuration file with the format

      ... specification of a plan class
   ... more plan class specifications

Name this file plan_class_spec.xml and put it in your project directory. This will replace the built in plan classes!


An example configuration file is here. This file specifies the predefined plan classes.

Specification format

The specification of a plan class has the following format. All elements except <name> are optional. In version numbers, M is major, m is minor, R is release.

the name of the plan class
multiply projected FLOPS by this factor. Use this to favor one class over another. For example, if you have both SSE and non-SSE versions, use 1.1 and 1.0 respectively.
use only hosts whose operating system version matches the given regular expression
a required CPU feature (such as sse3). You can include more than one.
only sends the app version to hosts with host.serialnum field that matches the given regular expression.

The following lets you use a project preference to decide whether to use the app version:

the name of the tag
the contents must match this regular expression

Fields for multicore apps:

run only on hosts with at least X usable CPUs.
use at most this many CPUs

Fields for GPU apps:

the GPU type (nvidia, amd, or intel)
the fraction of total FLOPs that are done by the CPU. This is used to calculate CPU usage and estimated FLOPS.
The minimum amount of GPU RAM. This is needed because older clients report total RAM but not available RAM.
require this much available GPU RAM
scale GPU peak speed by this (default 1).
how many GPUs will be used
minimum display driver version. AMD driver versions are represented as MMmmRRRR. NVIDIA driver versions are represented as MMMmm.
maximum display driver version
CUDA application (NVIDIA)
CAL application (AMD)
OpenCL application
you can use a project-specific preferences to let users scale the # of GPUs used. This is the tag name.
send this version only to hosts without OpenCL capability

Fields for AMD/ATI GPU apps:

Require libraries named "ati", not "amd".
Require libraries named "amd". You can verify which DLLs your application is linked against using Dependency Walker against your application. If your executable contains DLL names prefixed with 'ati', use this option. These flags are usually not needed for OpenCL apps.
Min and max CAL targets:
typedef enum CALtargetEnum {
    CAL_TARGET_600,                /**< R600 GPU ISA */     
    CAL_TARGET_610,                /**< RV610 GPU ISA */    
    CAL_TARGET_630,                /**< RV630 GPU ISA */
    CAL_TARGET_670,                /**< RV670 GPU ISA */
    CAL_TARGET_7XX,                /**< R700 class GPU ISA */
    CAL_TARGET_770,                /**< RV770 GPU ISA */
    CAL_TARGET_710,                /**< RV710 GPU ISA */
    CAL_TARGET_730,                /**< RV730 GPU ISA */
    CAL_TARGET_CYPRESS,            /**< CYPRESS GPU ISA */
    CAL_TARGET_JUNIPER,            /**< JUNIPER GPU ISA */
    CAL_TARGET_REDWOOD,            /**< REDWOOD GPU ISA */
    CAL_TARGET_CEDAR,               /**< CEDAR GPU ISA */
    CAL_TARGET_CAYMAN,              /**< CAYMAN GPU ISA */
    CAL_TARGET_KAUAI,           /** added by me **/
    CAL_TARGET_BARTS,               /**< BARTS GPU ISA */
    CAL_TARGET_TURKS,          /** added by me **/
    CAL_TARGET_CAICOS   /** added by me **/
} CALtarget;

Fields for NVIDIA GPU apps:

minimum compute capability
maximum compute capability

Fields for CUDA apps

minimum CUDA version
maximum CUDA version

Fields for OpenCL apps

minimum OpenCL version
maximum OpenCL version

Fields for OpenCL apps for AMD

minimum OpenCL driver revision
maximum OpenCL driver revision

Fields for VirtualBox apps

VirtualBox application; send only to hosts with VirtualBox installed
minimum VirtualBox version
maximum VirtualBox version
64-bit application.
send only to hosts with enabled VM hardware acceleration.

Note: VirtualBox apps can be multicore (set <min_ncpus> and <max_threads>). However, when sent to hosts without enabled VM hardware acceleration, they'll be run single-core.

Fields for non-compute-intensive apps:

average # CPUs used. Use for non-compute-intensive apps; for others it's calculated for you.