diff --git a/tests/test_helper.tcl b/tests/test_helper.tcl
index 598a39291..7874256da 100644
--- a/tests/test_helper.tcl
+++ b/tests/test_helper.tcl
@@ -404,6 +404,18 @@ for {set j 0} {$j < [llength $argv]} {incr j} {
     }
 }
 
+# With the parallel test running multiple Redis instances at the same time
+# we need a fast enough computer, otherwise a lot of tests may generate
+# false positives.
+# If the computer is too slow we revert the sequetial test without any
+# parallelism, that is, clients == 1.
+proc is_a_slow_computer {} {
+    set start [clock milliseconds]
+    for {set j 0} {$j < 1000000} {incr j} {}
+    set elapsed [expr [clock milliseconds]-$start]
+    expr {$elapsed > 200}
+}
+
 if {$::client} {
     if {[catch { test_client_main $::test_server_port } err]} {
         set estr "Executing test client: $err.\n$::errorInfo"
@@ -413,6 +425,11 @@ if {$::client} {
         exit 1
     }
 } else {
+    if {[is_a_slow_computer]} {
+        puts "** SLOW COMPUTER ** Using a single client to avoid false positives."
+        set ::numclients 1
+    }
+
     if {[catch { test_server_main } err]} {
         if {[string length $err] > 0} {
             # only display error when not generated by the test suite