Changes between Version 23 and Version 24 of AppMultiThread

Timestamp:

May 29, 2009, 2:48:12 PM (15 years ago)

Author:

davea

Comment:

--

AppMultiThread

@@ -1 @@
-= API for multi-thread apps =
-
-== Why write a multi-threaded app? ==
+= Multi-thread apps =
 
 The average number of cores per PC will increase over the next few years,
 possibly at a faster rate than the average amount of available RAM.
 
-Depending on your application and project, it may be desirable
-to develop a multi-threaded application.
+Depending on your application and project,
+it may be desirable to develop a multi-thread application.
 Possible reasons to do this:
 
@@ -25 +23 @@
 To deploy it:
 
- * Choose a "planning class" name for the program, say "par16" (see below).
- * Create an [UpdateVersions app version].  Include a file '''app_plan''' containing "par16".
- * Link the following function into your scheduler:
+ * Choose a "planning class" name for the program, say "par64" (see below).
+ * Create an [UpdateVersions app version], specifying its plan class as "par64".
+ * Link the following function into your scheduler (customized as needed):
 {{{
-bool app_plan(SCHEDULER_REQUEST& sreq, const char* plan_class, HOST_USAGE& hu) {
+int app_plan(SCHEDULER_REQUEST& sreq, const char* plan_class, HOST_USAGE& hu) {
     if (!strcmp(plan_class, "par64")) {
+        // the following is for an app that can use anywhere
+        // from 1 to 64 threads, can control this exactly,
+        // and whose speedup is .95N
+        // (on a uniprocessor, we'll use a sequential app if one is available)
+        //
         int ncpus, nthreads;
         bool bounded;
@@ -41 +44 @@
         sprintf(hu.cmdline, "--nthreads %d", nthreads);
         hu.flops = 0.95*sreq.host.p_fpops*nthreads;
-        return true;
-
+        if (config.debug_version_select) {
+            log_messages.printf(MSG_NORMAL,
+                "[version] Multi-thread app estimate %.2f GFLOPS\n",
+                hu.flops/1e9
+            );
+        }
+        return 0;
     }
-    return false;
+    return PLAN_REJECT_UNKNOWN;
 }
 }}}
+
+The BOINC client will schedule applications based
+on the average CPU usage returned by this function.