diff --git a/gns3server/controller/base_gns3_vm.py b/gns3server/controller/base_gns3_vm.py
index 4d86ac9b..458c99e4 100644
--- a/gns3server/controller/base_gns3_vm.py
+++ b/gns3server/controller/base_gns3_vm.py
@@ -15,10 +15,13 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
+from .gns3_vm_error import GNS3VMError
+
 import os
 import sys
 import json
 import asyncio
+import psutil
 
 import logging
 log = logging.getLogger(__name__)
@@ -43,6 +46,16 @@ class BaseGNS3VM:
         else:
             config_path = os.path.join(os.path.expanduser("~"), ".config", "GNS3")
         self._config_file = os.path.join(config_path, "gns3_vm.conf")
+
+        # limit the number of vCPUs to the number of physical cores (hyper thread CPUs are excluded)
+        # because this is likely to degrade performances.
+        self._vcpus = psutil.cpu_count(logical=False)
+        # we want to allocate half of the available physical memory
+        ram = int(psutil.virtual_memory().total / (1024 * 1024) / 2)
+        # value must be a multiple of 4 (VMware requirement)
+        ram -= ram % 4
+        self._ram = ram
+
         self.load()
 
     def __json__(self):
diff --git a/gns3server/controller/vmware_gns3_vm.py b/gns3server/controller/vmware_gns3_vm.py
index b32cd591..c9e71b1c 100644
--- a/gns3server/controller/vmware_gns3_vm.py
+++ b/gns3server/controller/vmware_gns3_vm.py
@@ -18,6 +18,7 @@
 import os
 import logging
 import asyncio
+import psutil
 
 from gns3server.compute.vmware import (
     VMware,
@@ -56,6 +57,18 @@ class VMwareGNS3VM(BaseGNS3VM):
         :param ram: amount of RAM
         """
 
+        available_ram = int(psutil.virtual_memory().available / (1024 * 1024))
+        if ram > available_ram:
+            raise GNS3VMError("You have allocated too much memory for the GNS3 VM! (available memory is {} MB)".format(available_ram))
+
+        # memory must be a multiple of 4 (VMware requirement)
+        if ram % 4 != 0:
+            raise GNS3VMError("Allocated memory for the GNS3 VM must be a multiple of 4".format(available_ram))
+
+        available_vcpus = psutil.cpu_count(logical=False)
+        if vcpus > available_vcpus:
+            raise GNS3VMError("You have allocated too many vCPUs for the GNS3 VM! (max available is {} vCPUs)".format(available_vcpus))
+
         try:
             pairs = VMware.parse_vmware_file(self._vmx_path)
             pairs["numvcpus"] = str(vcpus)
@@ -91,8 +104,8 @@ class VMwareGNS3VM(BaseGNS3VM):
         if not os.path.exists(self._vmx_path):
             raise GNS3VMError("VMware VMX file {} doesn't exist".format(self._vmx_path))
 
-        # set the number of vCPUs and amount of RAM  # FIXME
-        # yield from self._set_vcpus_ram(self.vcpus, self.ram)
+        # set the number of vCPUs and amount of RAM
+        yield from self._set_vcpus_ram(self.vcpus, self.ram)
 
         # start the VM
         args = [self._vmx_path]
diff --git a/gns3server/web/route.py b/gns3server/web/route.py
index e956ed59..ddd8297f 100644
--- a/gns3server/web/route.py
+++ b/gns3server/web/route.py
@@ -31,6 +31,7 @@ log = logging.getLogger(__name__)
 from ..compute.error import NodeError, ImageMissingError
 from ..controller.controller_error import ControllerError
 from ..ubridge.ubridge_error import UbridgeError
+from ..controller.gns3_vm_error import GNS3VMError
 from .response import Response
 from ..crash_report import CrashReport
 from ..config import Config
@@ -205,7 +206,7 @@ class Route(object):
                     response = Response(request=request, route=route)
                     response.set_status(e.status)
                     response.json({"message": e.text, "status": e.status})
-                except (ControllerError) as e:
+                except (ControllerError, GNS3VMError) as e:
                     log.error("Controller error detected: {type}".format(type=type(e)), exc_info=1)
                     response = Response(request=request, route=route)
                     response.set_status(409)