Logo AND Algorithmique Numérique Distribuée

Public GIT Repository
convert platform-failures
authorFrederic Suter <frederic.suter@cc.in2p3.fr>
Sat, 7 Mar 2020 14:46:44 +0000 (15:46 +0100)
committerFrederic Suter <frederic.suter@cc.in2p3.fr>
Sat, 7 Mar 2020 14:46:44 +0000 (15:46 +0100)
MANIFEST.in
examples/c/CMakeLists.txt
examples/c/platform-failures/platform-failures.c [new file with mode: 0644]
examples/c/platform-failures/platform-failures.tesh [moved from examples/deprecated/msg/platform-failures/platform-failures.tesh with 94% similarity]
examples/deprecated/msg/CMakeLists.txt
examples/deprecated/msg/platform-failures/platform-failures.c [deleted file]

index 36e80cc..227d242 100644 (file)
@@ -89,6 +89,8 @@ include examples/c/io-disk-raw/io-disk-raw.tesh
 include examples/c/io-file-remote/io-file-remote.c
 include examples/c/io-file-remote/io-file-remote.tesh
 include examples/c/io-file-remote/io-file-remote_d.xml
+include examples/c/platform-failures/platform-failures.c
+include examples/c/platform-failures/platform-failures.tesh
 include examples/c/plugin-hostload/plugin-hostload.c
 include examples/c/plugin-hostload/plugin-hostload.tesh
 include examples/deprecated/java/app/bittorrent/Common.java
@@ -271,8 +273,6 @@ include examples/deprecated/msg/network-ns3/network-ns3.c
 include examples/deprecated/msg/network-ns3/network-ns3.tesh
 include examples/deprecated/msg/network-ns3/one_cluster_d.xml
 include examples/deprecated/msg/network-ns3/onelink_d.xml
-include examples/deprecated/msg/platform-failures/platform-failures.c
-include examples/deprecated/msg/platform-failures/platform-failures.tesh
 include examples/deprecated/msg/synchro-semaphore/synchro-semaphore.c
 include examples/deprecated/msg/synchro-semaphore/synchro-semaphore.tesh
 include examples/deprecated/msg/trace-categories/trace-categories.c
index 11d445b..dcf5c24 100644 (file)
@@ -10,6 +10,7 @@ foreach(x
         exec-async exec-basic exec-dvfs exec-remote exec-waitany
         energy-exec energy-exec-ptask energy-vm
         io-disk-raw io-file-remote
+        platform-failures
         plugin-hostload)
   add_executable       (${x}-c EXCLUDE_FROM_ALL ${x}/${x}.c)
   target_link_libraries(${x}-c simgrid)
@@ -65,6 +66,7 @@ foreach(x
         exec-async exec-basic exec-dvfs exec-remote exec-waitany
         energy-exec energy-exec-ptask energy-vm
         io-disk-raw io-file-remote
+        platform-failures
         plugin-hostload)
   ADD_TESH(c-${x} --setenv platfdir=${CMAKE_HOME_DIRECTORY}/examples/platforms
                   --setenv bindir=${CMAKE_BINARY_DIR}/examples/c/${x}
diff --git a/examples/c/platform-failures/platform-failures.c b/examples/c/platform-failures/platform-failures.c
new file mode 100644 (file)
index 0000000..725e31f
--- /dev/null
@@ -0,0 +1,139 @@
+/* Copyright (c) 2007-2020. The SimGrid Team. All rights reserved.          */
+
+/* This program is free software; you can redistribute it and/or modify it
+ * under the terms of the license (GNU LGPL) which comes with this package. */
+
+#include "simgrid/actor.h"
+#include "simgrid/comm.h"
+#include "simgrid/engine.h"
+#include "simgrid/exec.h"
+#include "simgrid/host.h"
+#include "simgrid/mailbox.h"
+
+#include "xbt/asserts.h"
+#include "xbt/log.h"
+#include "xbt/str.h"
+
+#include <stdio.h> /* snprintf */
+
+#define FINALIZE 221297 /* a magic number to tell people to stop working */
+
+XBT_LOG_NEW_DEFAULT_CATEGORY(platform_failures, "Messages specific for this example");
+
+static void master(int argc, char* argv[])
+{
+  xbt_assert(argc == 5);
+  long number_of_tasks  = xbt_str_parse_int(argv[1], "Invalid amount of tasks: %s");
+  double task_comp_size = xbt_str_parse_double(argv[2], "Invalid computational size: %s");
+  double task_comm_size = xbt_str_parse_double(argv[3], "Invalid communication size: %s");
+  long workers_count    = xbt_str_parse_int(argv[4], "Invalid amount of workers: %s");
+
+  XBT_INFO("Got %ld workers and %ld tasks to process", workers_count, number_of_tasks);
+
+  for (int i = 0; i < number_of_tasks; i++) {
+    char mailbox_name[256];
+    snprintf(mailbox_name, 255, "worker-%ld", i % workers_count);
+    sg_mailbox_t mailbox = sg_mailbox_by_name(mailbox_name);
+
+    XBT_INFO("Send a message to %s", mailbox_name);
+    double* payload = (double*)xbt_malloc(sizeof(double));
+    *payload        = task_comp_size;
+    sg_comm_t comm  = sg_mailbox_put_async(mailbox, payload, task_comm_size);
+
+    switch (sg_comm_wait_for(comm, 10.0)) {
+      case SG_OK:
+        XBT_INFO("Send to %s completed", mailbox_name);
+        break;
+      case SG_ERROR_NETWORK:
+        XBT_INFO("Mmh. The communication with '%s' failed. Nevermind. Let's keep going!", mailbox_name);
+        free(payload);
+        break;
+      case SG_ERROR_TIMEOUT:
+        XBT_INFO("Mmh. Got timeouted while speaking to '%s'. Nevermind. Let's keep going!", mailbox_name);
+        free(payload);
+        break;
+      default:
+        xbt_die("Unexpected behavior");
+    }
+  }
+
+  XBT_INFO("All tasks have been dispatched. Let's tell everybody the computation is over.");
+  for (int i = 0; i < workers_count; i++) {
+    char mailbox_name[256];
+    snprintf(mailbox_name, 255, "worker-%ld", i % workers_count);
+    sg_mailbox_t mailbox = sg_mailbox_by_name(mailbox_name);
+    double* payload      = (double*)xbt_malloc(sizeof(double));
+    *payload             = FINALIZE;
+    sg_comm_t comm       = sg_mailbox_put_async(mailbox, payload, 0);
+
+    switch (sg_comm_wait_for(comm, 1.0)) {
+      case SG_ERROR_NETWORK:
+        XBT_INFO("Mmh. Can't reach '%s'! Nevermind. Let's keep going!", mailbox_name);
+        free(payload);
+        break;
+      case SG_ERROR_TIMEOUT:
+        XBT_INFO("Mmh. Got timeouted while speaking to '%s'. Nevermind. Let's keep going!", mailbox_name);
+        free(payload);
+        break;
+      case SG_OK:
+        /* nothing */
+        break;
+      default:
+        xbt_die("Unexpected behavior with '%s'", mailbox_name);
+    }
+  }
+
+  XBT_INFO("Goodbye now!");
+}
+
+static void worker(int argc, char* argv[])
+{
+  xbt_assert(argc == 2);
+  char mailbox_name[80];
+  double* payload;
+  long id = xbt_str_parse_int(argv[1], "Invalid argument %s");
+
+  snprintf(mailbox_name, 79, "worker-%ld", id);
+  sg_mailbox_t mailbox = sg_mailbox_by_name(mailbox_name);
+
+  while (1) {
+    XBT_INFO("Waiting a message on %s", mailbox_name);
+    sg_comm_t comm     = sg_mailbox_get_async(mailbox, (void**)&payload);
+    sg_error_t retcode = sg_comm_wait(comm);
+    if (retcode == SG_OK) {
+      if (*payload == FINALIZE) {
+        free(payload);
+        break;
+      } else {
+        XBT_INFO("Start execution...");
+        sg_actor_execute(*payload);
+        XBT_INFO("Execution complete.");
+        free(payload);
+      }
+    } else if (retcode == SG_ERROR_NETWORK) {
+      XBT_INFO("Mmh. Something went wrong. Nevermind. Let's keep going!");
+    }
+  }
+  XBT_INFO("I'm done. See you!");
+}
+
+int main(int argc, char* argv[])
+{
+  simgrid_init(&argc, argv);
+  xbt_assert(argc > 2,
+             "Usage: %s platform_file deployment_file\n"
+             "\tExample: %s msg_platform.xml msg_deployment.xml\n",
+             argv[0], argv[0]);
+
+  simgrid_load_platform(argv[1]);
+
+  simgrid_register_function("master", master);
+  simgrid_register_function("worker", worker);
+  simgrid_load_deployment(argv[2]);
+
+  simgrid_run();
+
+  XBT_INFO("Simulation time %g", simgrid_get_clock());
+
+  return 0;
+}
@@ -3,7 +3,7 @@
 p Testing a simple master/worker example application handling failures TCP crosstraffic DISABLED
 
 ! output sort 19
-$ ${bindir:=.}/platform-failures --log=xbt_cfg.thres:critical --log=no_loc ${platfdir}/small_platform_failures.xml ${srcdir}/../app-masterworker/app-masterworker_d.xml --cfg=path:${srcdir} --cfg=network/crosstraffic:0 "--log=root.fmt:[%10.6r]%e(%i:%P@%h)%e%m%n" --log=surf_cpu.t:verbose
+$ ${bindir:=.}/platform-failures-c --log=xbt_cfg.thres:critical --log=no_loc ${platfdir}/small_platform_failures.xml ${srcdir:=.}/../../s4u/platform-failures/s4u-platform-failures_d.xml --cfg=path:${srcdir} --cfg=network/crosstraffic:0 "--log=root.fmt:[%10.6r]%e(%i:%P@%h)%e%m%n" --log=surf_cpu.t:verbose
 > [  0.000000] (0:maestro@) Cannot launch actor 'worker' on failed host 'Fafard'
 > [  0.000000] (0:maestro@) Deployment includes some initially turned off Hosts ... nevermind.
 > [  0.000000] (1:master@Tremblay) Got 5 workers and 20 tasks to process
@@ -108,7 +108,7 @@ $ ${bindir:=.}/platform-failures --log=xbt_cfg.thres:critical --log=no_loc ${pla
 p Testing a simple master/worker example application handling failures. TCP crosstraffic ENABLED
 
 ! output sort 19
-$ ${bindir:=.}/platform-failures --log=xbt_cfg.thres:critical --log=no_loc ${platfdir}/small_platform_failures.xml ${srcdir}/../app-masterworker/app-masterworker_d.xml --cfg=path:${srcdir} "--log=root.fmt:[%10.6r]%e(%i:%P@%h)%e%m%n" --log=surf_cpu.t:verbose
+$ ${bindir:=.}/platform-failures-c --log=xbt_cfg.thres:critical --log=no_loc ${platfdir}/small_platform_failures.xml ${srcdir:=.}/../../s4u/platform-failures/s4u-platform-failures_d.xml --cfg=path:${srcdir} "--log=root.fmt:[%10.6r]%e(%i:%P@%h)%e%m%n" --log=surf_cpu.t:verbose
 > [  0.000000] (0:maestro@) Cannot launch actor 'worker' on failed host 'Fafard'
 > [  0.000000] (0:maestro@) Deployment includes some initially turned off Hosts ... nevermind.
 > [  0.000000] (1:master@Tremblay) Got 5 workers and 20 tasks to process
@@ -217,4 +217,4 @@ p   complex with such an integration test. One day, we will setup a set of
 p   unit tests for the surf solver, and such issues will be addressable again.
 p For the time being, I just give up, sorry.
 
-p $ ${bindir:=.}/platform-failures --log=xbt_cfg.thres:critical --log=no_loc ${platfdir}/small_platform_failures.xml ${srcdir}/../app-masterworker/app-masterworker_d.xml --cfg=path:${srcdir} --cfg=cpu/optim:TI "--log=root.fmt:[%10.6r]%e(%i:%P@%h)%e%m%n" --log=surf_cpu.t:verbose
+p $ ${bindir:=.}/platform-failures-c --log=xbt_cfg.thres:critical --log=no_loc ${platfdir}/small_platform_failures.xml ${srcdir:=.}/../../s4u/platform-failures/s4u-platform-failures_d.xml --cfg=path:${srcdir} --cfg=cpu/optim:TI "--log=root.fmt:[%10.6r]%e(%i:%P@%h)%e%m%n" --log=surf_cpu.t:verbose
index 402f82c..cbbc1d5 100644 (file)
@@ -1,6 +1,6 @@
 # C examples
 foreach(x app-masterworker cloud-masterworker
-          dht-pastry platform-failures
+          dht-pastry
           synchro-semaphore trace-categories 
           trace-route-user-variables trace-link-user-variables trace-masterworker
           trace-process-migration trace-host-user-variables)
@@ -56,7 +56,7 @@ set(xml_files    ${xml_files}     ${CMAKE_CURRENT_SOURCE_DIR}/app-masterworker/a
 
 if(enable_msg)
   foreach(x app-masterworker cloud-masterworker
-            dht-pastry dht-kademlia platform-failures
+            dht-pastry dht-kademlia
             synchro-semaphore)
     ADD_TESH_FACTORIES(msg-${x} "thread;ucontext;raw;boost" 
                                 --setenv bindir=${CMAKE_BINARY_DIR}/examples/deprecated/msg/${x}
diff --git a/examples/deprecated/msg/platform-failures/platform-failures.c b/examples/deprecated/msg/platform-failures/platform-failures.c
deleted file mode 100644 (file)
index 542f034..0000000
+++ /dev/null
@@ -1,143 +0,0 @@
-/* Copyright (c) 2007-2020. The SimGrid Team. All rights reserved.          */
-
-/* This program is free software; you can redistribute it and/or modify it
- * under the terms of the license (GNU LGPL) which comes with this package. */
-
-#include "simgrid/msg.h"
-
-#include <stdio.h> /* snprintf */
-
-XBT_LOG_NEW_DEFAULT_CATEGORY(msg_test, "Messages specific for this msg example");
-
-#define FINALIZE ((void*)221297)        /* a magic number to tell people to stop working */
-
-static void task_cleanup_handler(void* task)
-{
-  MSG_task_destroy(task);
-}
-
-static int master(int argc, char *argv[])
-{
-  xbt_assert(argc == 5);
-  long number_of_tasks = xbt_str_parse_int(argv[1], "Invalid amount of tasks: %s");
-  double task_comp_size = xbt_str_parse_double(argv[2], "Invalid computational size: %s");
-  double task_comm_size = xbt_str_parse_double(argv[3], "Invalid communication size: %s");
-  long workers_count = xbt_str_parse_int(argv[4], "Invalid amount of workers: %s");
-
-  XBT_INFO("Got %ld workers and %ld tasks to process", workers_count, number_of_tasks);
-
-  for (int i = 0; i < number_of_tasks; i++) {
-    char mailbox[256];
-    snprintf(mailbox, 255, "worker-%ld", i % workers_count);
-    XBT_INFO("Send a message to %s", mailbox);
-    msg_task_t task = MSG_task_create("Task", task_comp_size, task_comm_size, NULL);
-
-    switch ( MSG_task_send_with_timeout(task,mailbox,10.0) ) {
-    case MSG_OK:
-      XBT_INFO("Send to %s completed", mailbox);
-      break;
-
-    case MSG_TRANSFER_FAILURE:
-      XBT_INFO("Mmh. The communication with '%s' failed. Nevermind. Let's keep going!", mailbox);
-      MSG_task_destroy(task);
-      break;
-
-    case MSG_TIMEOUT:
-      XBT_INFO ("Mmh. Got timeouted while speaking to '%s'. Nevermind. Let's keep going!", mailbox);
-      MSG_task_destroy(task);
-      break;
-
-    default:
-      xbt_die( "Unexpected behavior");
-    }
-  }
-
-  XBT_INFO("All tasks have been dispatched. Let's tell everybody the computation is over.");
-  for (int i = 0; i < workers_count; i++) {
-    char mailbox[256];
-    snprintf(mailbox, 255, "worker-%ld", i % workers_count);
-    msg_task_t task = MSG_task_create("finalize", 0, 0, FINALIZE);
-
-    switch (MSG_task_send_with_timeout(task,mailbox,1.0)) {
-
-    case MSG_TRANSFER_FAILURE:
-      XBT_INFO("Mmh. Can't reach '%s'! Nevermind. Let's keep going!", mailbox);
-      MSG_task_destroy(task);
-      break;
-
-    case MSG_TIMEOUT:
-      XBT_INFO("Mmh. Got timeouted while speaking to '%s'. Nevermind. Let's keep going!", mailbox);
-      MSG_task_destroy(task);
-      break;
-
-    case MSG_OK:
-      /* nothing */
-      break;
-
-    default:
-      xbt_die("Unexpected behavior with '%s'", mailbox);
-    }
-  }
-
-  XBT_INFO("Goodbye now!");
-  return 0;
-}
-
-static int worker(int argc, char *argv[])
-{
-  xbt_assert(argc == 2);
-  char mailbox[80];
-
-  long id= xbt_str_parse_int(argv[1], "Invalid argument %s");
-
-  snprintf(mailbox, 79,"worker-%ld", id);
-
-  while (1) {
-    msg_task_t task = NULL;
-    XBT_INFO("Waiting a message on %s", mailbox);
-    int retcode = MSG_task_receive( &(task), mailbox);
-    if (retcode == MSG_OK) {
-      if (MSG_task_get_data(task) == FINALIZE) {
-        MSG_task_destroy(task);
-        break;
-      }
-      XBT_INFO("Start execution...");
-      MSG_process_set_data(MSG_process_self(), task);
-      retcode = MSG_task_execute(task);
-      MSG_process_set_data(MSG_process_self(), NULL);
-      if (retcode == MSG_OK) {
-        XBT_INFO("Execution complete.");
-        MSG_task_destroy(task);
-      } else {
-        XBT_INFO("Hey ?! What's up ? ");
-        xbt_die("Unexpected behavior");
-      }
-    } else if (retcode == MSG_TRANSFER_FAILURE) {
-      XBT_INFO("Mmh. Something went wrong. Nevermind. Let's keep going!");
-    } else {
-      xbt_die("Unexpected behavior");
-    }
-  }
-  XBT_INFO("I'm done. See you!");
-  return 0;
-}
-
-int main(int argc, char *argv[])
-{
-  MSG_init(&argc, argv);
-  xbt_assert(argc > 2, "Usage: %s platform_file deployment_file\n"
-             "\tExample: %s msg_platform.xml msg_deployment.xml\n", argv[0], argv[0]);
-
-  MSG_create_environment(argv[1]);
-
-  MSG_function_register("master", master);
-  MSG_function_register("worker", worker);
-  MSG_process_set_data_cleanup(task_cleanup_handler);
-  MSG_launch_application(argv[2]);
-
-  msg_error_t res = MSG_main();
-
-  XBT_INFO("Simulation time %g", MSG_get_clock());
-
-  return res != MSG_OK;
-}