Merge branch 'rickard/scheduler-wakeup-threshold/OTP-8811' into dev

* rickard/scheduler-wakeup-threshold/OTP-8811:
  Add scheduler wakup threshold as command line argument
  Lower the scheduler wakeup threshold

5 files changed, 66 insertions, 10 deletions

diff --git a/erts/doc/src/erl.xml b/erts/doc/src/erl.xml
index f477280a6f..e36d0adb0d 100644
--- a/erts/doc/src/erl.xml
+++ b/erts/doc/src/erl.xml
@@ -831,14 +831,28 @@
 	    <p>For more information, see
 	    <seealso marker="erlang#system_flag_cpu_topology">erlang:system_flag(cpu_topology, CpuTopology)</seealso>.</p>
           </item>
+	  <tag><marker id="+swt"><c>+swt very_low|low|medium|high|very_high</c></marker></tag>
+	  <item>
+             <p>Set scheduler wakeup threshold. Default is <c>medium</c>.
+	        The threshold determines when to wake up sleeping schedulers
+		when more work than can be handled by currently awake schedulers
+		exist. A low threshold will cause earlier wakeups, and a high
+		threshold will cause later wakeups. Early wakeups will
+		distribute work over multiple schedulers faster, but work will
+		more easily bounce between schedulers.
+	     </p>
+             <p><em>NOTE:</em> This flag may be removed or changed at any time
+	        without prior notice.
+	     </p>
+	  </item>
+	  <tag><marker id="sched_thread_stack_size"><c><![CDATA[+sss size]]></c></marker></tag>
+	  <item>
+	    <p>Suggested stack size, in kilowords, for scheduler threads.
+            Valid range is 4-8192 kilowords. The default stack size
+            is OS dependent.</p>
+	  </item>
         </taglist>
       </item>
-      <tag><marker id="sched_thread_stack_size"><c><![CDATA[+sss size]]></c></marker></tag>
-      <item>
-        <p>Suggested stack size, in kilowords, for scheduler threads.
-          Valid range is 4-8192 kilowords. The default stack size
-          is OS dependent.</p>
-      </item>
       <tag><marker id="+t"><c><![CDATA[+t size]]></c></marker></tag>
       <item>
         <p>Set the maximum number of atoms the VM can handle. Default is 1048576.</p>
diff --git a/erts/emulator/beam/erl_init.c b/erts/emulator/beam/erl_init.c
index 14bd10b42c..4ae656a3ad 100644
--- a/erts/emulator/beam/erl_init.c
+++ b/erts/emulator/beam/erl_init.c
@@ -512,6 +512,8 @@ void erts_usage(void)
     erts_fprintf(stderr, "            u|ns|ts|ps|s|nnts|nnps|tnnps|db\n");
     erts_fprintf(stderr, "-sct cput   set cpu topology,\n");
     erts_fprintf(stderr, "            see the erl(1) documentation for more info.\n");
+    erts_fprintf(stderr, "-swt val    set scheduler wakeup threshold, valid values are:\n");
+    erts_fprintf(stderr, "            very_low|low|medium|high|very_high.\n");
     erts_fprintf(stderr, "-sss size   suggested stack size in kilo words for scheduler threads,\n");
     erts_fprintf(stderr, "            valid range is [%d-%d]\n",
 		 ERTS_SCHED_THREAD_MIN_STACK_SIZE,
@@ -1176,10 +1178,20 @@ erl_start(int argc, char **argv)
 	    }
 	    else if (sys_strcmp("mrq", sub_param) == 0)
 		use_multi_run_queue = 1;
-	    else if (sys_strcmp("srq", sub_param) == 0)
-		use_multi_run_queue = 0;
 	    else if (sys_strcmp("nsp", sub_param) == 0)
 		erts_use_sender_punish = 0;
+	    else if (sys_strcmp("srq", sub_param) == 0)
+		use_multi_run_queue = 0;
+	    else if (sys_strcmp("wt", sub_param) == 0) {
+		arg = get_arg(sub_param+2, argv[i+1], &i);
+		if (erts_sched_set_wakeup_limit(arg) != 0) {
+		    erts_fprintf(stderr, "scheduler wakeup threshold: %s\n",
+				 arg);
+		    erts_usage();
+		}
+		VERBOSE(DEBUG_SYSTEM,
+			("scheduler wakup threshold: %s\n", arg));
+	    }
 	    else if (has_prefix("ss", sub_param)) {
 		/* suggested stack size (Kilo Words) for scheduler threads */
 		arg = get_arg(sub_param+2, argv[i+1], &i);
diff --git a/erts/emulator/beam/erl_process.c b/erts/emulator/beam/erl_process.c
index 2609457415..901167a315 100644
--- a/erts/emulator/beam/erl_process.c
+++ b/erts/emulator/beam/erl_process.c
@@ -54,7 +54,12 @@
   (ERTS_SCHED_SYS_SLEEP_SPINCOUNT*ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT)
 #define ERTS_SCHED_SUSPEND_SLEEP_SPINCOUNT 0
 
-#define ERTS_WAKEUP_OTHER_LIMIT (100*CONTEXT_REDS/2)
+#define ERTS_WAKEUP_OTHER_LIMIT_VERY_HIGH (200*CONTEXT_REDS)
+#define ERTS_WAKEUP_OTHER_LIMIT_HIGH (50*CONTEXT_REDS)
+#define ERTS_WAKEUP_OTHER_LIMIT_MEDIUM (10*CONTEXT_REDS)
+#define ERTS_WAKEUP_OTHER_LIMIT_LOW (CONTEXT_REDS)
+#define ERTS_WAKEUP_OTHER_LIMIT_VERY_LOW (CONTEXT_REDS/10)
+
 #define ERTS_WAKEUP_OTHER_DEC 10
 #define ERTS_WAKEUP_OTHER_FIXED_INC (CONTEXT_REDS/10)
 
@@ -112,6 +117,8 @@ Uint erts_no_schedulers;
 Uint erts_max_processes = ERTS_DEFAULT_MAX_PROCESSES;
 Uint erts_process_tab_index_mask;
 
+static int wakeup_other_limit;
+
 #ifdef ERTS_SMP
 Uint erts_max_main_threads;
 #endif
@@ -2373,7 +2380,27 @@ void
 erts_early_init_scheduling(void)
 {
     early_cpu_bind_init();
+    wakeup_other_limit = ERTS_WAKEUP_OTHER_LIMIT_MEDIUM;
+}
+
+int
+erts_sched_set_wakeup_limit(char *str)
+{
+    if (sys_strcmp(str, "very_high") == 0)
+	wakeup_other_limit = ERTS_WAKEUP_OTHER_LIMIT_VERY_HIGH;
+    else if (sys_strcmp(str, "high") == 0)
+	wakeup_other_limit = ERTS_WAKEUP_OTHER_LIMIT_HIGH;
+    else if (sys_strcmp(str, "medium") == 0)
+	wakeup_other_limit = ERTS_WAKEUP_OTHER_LIMIT_MEDIUM;
+    else if (sys_strcmp(str, "low") == 0)
+	wakeup_other_limit = ERTS_WAKEUP_OTHER_LIMIT_LOW;
+    else if (sys_strcmp(str, "very_low") == 0)
+	wakeup_other_limit = ERTS_WAKEUP_OTHER_LIMIT_VERY_LOW;
+    else
+	return EINVAL;
+    return 0;
 }
+	
 
 void
 erts_init_scheduling(int mrq, int no_schedulers, int no_schedulers_online)
@@ -7169,7 +7196,7 @@ Process *schedule(Process *p, int calls)
 		    if (rq->wakeup_other < 0)
 			rq->wakeup_other = 0;
 		}
-		else if (rq->wakeup_other < ERTS_WAKEUP_OTHER_LIMIT)
+		else if (rq->wakeup_other < wakeup_other_limit)
 		    rq->wakeup_other += rq->len*wo_reds + ERTS_WAKEUP_OTHER_FIXED_INC;
 		else {
 		    if (erts_common_run_queue) {
diff --git a/erts/emulator/beam/erl_process.h b/erts/emulator/beam/erl_process.h
index e49710a7ed..4365e409e5 100644
--- a/erts/emulator/beam/erl_process.h
+++ b/erts/emulator/beam/erl_process.h
@@ -1038,6 +1038,8 @@ ErtsProcList *erts_proclist_create(Process *);
 void erts_proclist_destroy(ErtsProcList *);
 int erts_proclist_same(ErtsProcList *, Process *);
 
+int erts_sched_set_wakeup_limit(char *str);
+
 #ifdef DEBUG
 void erts_dbg_multi_scheduling_return_trap(Process *, Eterm);
 #endif
diff --git a/erts/etc/common/erlexec.c b/erts/etc/common/erlexec.c
index 76ce0a7e3c..c1fc2aebee 100644
--- a/erts/etc/common/erlexec.c
+++ b/erts/etc/common/erlexec.c
@@ -120,6 +120,7 @@ static char *plusM_other_switches[] = {
 static char *pluss_val_switches[] = {
     "bt",
     "ct",
+    "wt",
     "ss",
     NULL
 };