ref-manual: Added ICECC_PARALLEL_MAKE to glossary.

Fixes [YOCTO #5801]

First draft of this description.

(From yocto-docs rev: 412d0cc7dbb74ece45940c40fbb617217ab88d2d)

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

documentation/ref-manual/ref-variables.xml

@@ -2405,6 +2405,46 @@
             </glossdef>
         </glossentry>
 
+        <glossentry id='var-ICECC_PARALLEL_MAKE'><glossterm>ICECC_PARALLEL_MAKE</glossterm>
+            <glossdef>
+                <para>
+                    Extra options passed to the <filename>make</filename>
+                    command during the <filename>do_compile</filename> task
+                    that specify parallel compilation.
+                    This variable usually takes the form of
+                    <filename>-j 4</filename>, where the number
+                    represents the maximum number of parallel threads
+                    <filename>make</filename> can run.
+                    <note>
+                        The options passed affect builds on all enabled
+                        machines on the network, which are machines running the
+                        <filename>iceccd</filename> daemon.
+                    </note>
+                </para>
+
+                <para>
+                    If your enabled machines support multiple cores,
+                    coming up with the maximum number of parallel threads
+                    that gives you the best performance could take some
+                    experimentation since machine speed, network lag,
+                    available memory, and existing machine loads can all
+                    affect build time.
+                    Consequently, unlike the
+                    <link linkend='var-PARALLEL_MAKE'><filename>PARALLEL_MAKE</filename></link>
+                    variable, there is no rule-of-thumb for setting
+                    <filename>ICECC_PARALLEL_MAKE</filename> to achieve
+                    optimal performance.
+                </para>
+
+                <para>
+                    If you do not set <filename>ICECC_PARALLEL_MAKE</filename>,
+                    the build system does not use it (i.e. the system does
+                    not detect and assign the number of cores as is done with
+                    <filename>PARALLEL_MAKE</filename>).
+                </para>
+            </glossdef>
+        </glossentry>
+
         <glossentry id='var-ICECC_PATH'><glossterm>ICECC_PATH</glossterm>
             <glossdef>
                 <para>