Browse Source

add link between off-line performance feedback and defining a performance model

Samuel Thibault 13 years ago
parent
commit
4a55fb873c
1 changed files with 4 additions and 3 deletions
  1. 4 3
      doc/starpu.texi

+ 4 - 3
doc/starpu.texi

@@ -1349,8 +1349,9 @@ for further executions, and can be observed by using the
 @code{starpu_perfmodel_display} command, or drawn by using
 @code{starpu_perfmodel_display} command, or drawn by using
 the @code{starpu_perfmodel_plot}.  The models are indexed by machine name. To
 the @code{starpu_perfmodel_plot}.  The models are indexed by machine name. To
 share the models between machines (e.g. for a homogeneous cluster), use
 share the models between machines (e.g. for a homogeneous cluster), use
-@code{export STARPU_HOSTNAME=some_global_name}.  The following is a small code
-example.
+@code{export STARPU_HOSTNAME=some_global_name}. Measurements are only done when using a task scheduler which makes use of it, such as @code{heft} or @code{dmda}.
+
+The following is a small code example.
 
 
 If e.g. the code is recompiled with other compilation options, or several
 If e.g. the code is recompiled with other compilation options, or several
 variants of the code are used, the symbol string should be changed to reflect
 variants of the code are used, the symbol string should be changed to reflect
@@ -2214,7 +2215,7 @@ schedulable yet are shown in grey.
 @node Codelet performance
 @node Codelet performance
 @section Performance of codelets
 @section Performance of codelets
 
 
-The performance model of codelets can be examined by using the
+The performance model of codelets (described in @ref{Performance model example}) can be examined by using the
 @code{starpu_perfmodel_display} tool:
 @code{starpu_perfmodel_display} tool:
 
 
 @example
 @example