Skip to content

ShmemTrainMatcher

Base classes: TrainMatcher

Enable Karabo channel (enableKaraboOutput)

Type
BOOL
Allowed in states
PASSIVE
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Default value
True

Processing time (processingTime)

Type
DOUBLE
Access mode
READONLY
Assignment
OPTIONAL
Default value
0.0

processingThreads

Type
UINT32
Allowed in states
PASSIVE
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Default value
16

useInfiniband

Type
BOOL
Access mode
INITONLY
Assignment
OPTIONAL
Description
If enabled, device will during initialization try to bind its data output channel (output) to its node's infiniband interface. Default interface is used if no 'ib0' interface is found.
Default value
True

Frame selection (frameSelector)

Enabled (frameSelector.enable)

Type
BOOL
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Default value
False

Arbiter source (frameSelector.arbiterSource)

Type
STRING
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Description
Source name to pull the frame selection pattern from, must be part of matched sources.
Default value
empty string

Data source pattern (frameSelector.dataSourcePattern)

Type
STRING
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Description
Source name pattern to apply frame selection to. Should match subset of matched sources.
Default value
empty string

Data keys (frameSelector.dataKeys)

Type
VECTOR_STRING
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Description
Keys in data sources to apply frame selection to.
Default value
[]

Array stacking (merge)

Type
VECTOR_HASH
Allowed in states
PASSIVE
Access mode
RECONFIGURABLE
Assignment
OPTIONAL
Description
Specify which source(s) or key(s) to stack or interleave.When stacking sources, the 'Source pattern' is interpreted as a regular expression and the 'Key pattern' is interpreted as an ordinary string. From all sources matching the source pattern, the data under this key (should be array with same dimensions across all stacked sources) is stacked in the same order as the sources are listed in 'Data sources' and the result is under the same key name in a new source named by 'Replacement'. When stacking keys, both the 'Source pattern' and the 'Key pattern' are regular expressions. Within each source matching the source pattern, all keys matching the key pattern are stacked and the result is put under the key named by 'Replacement'. While source stacking is optimized and can use thread pool, key stacking will iterate over all paths in matched sources and naively call np.stack for each key pattern. In either case, data that is used for stacking is removed from its original location (e.g. key is erased from hash). In both cases, the data can alternatively be interleaved. This is essentially equivalent to stacking except followed by a reshape such that the output is shaped like concatenation.
Table schema and default value
Select Source pattern Key pattern Replacement Group type Merge method Axis Missing value default