diff --git a/benchmarks/dispatch.b b/benchmarks/bench-dispatch.b
similarity index 100%
rename from benchmarks/dispatch.b
rename to benchmarks/bench-dispatch.b
diff --git a/benchmarks/bench-list.b b/benchmarks/bench-list.b
new file mode 100644
index 00000000..4a9c5063
--- /dev/null
+++ b/benchmarks/bench-list.b
@@ -0,0 +1,75 @@
+# This code is derived from the SOM benchmarks
+# 
+# Based on the JavaScript implementation at 
+# https://github.com/smarr/are-we-fast-yet/blob/master/benchmarks/JavaScript/list.js
+
+class Element {
+  Element(v){
+    self.val = v
+    self.next = nil
+  }
+
+  length() {
+    if self.next == nil {
+      return 1
+    }
+
+    return 1 + self.next.length()
+  }
+}
+
+class List {
+  benchmark() {
+    var result = self.tail(
+      self.make_list(15),
+      self.make_list(10),
+      self.make_list(6)
+    )
+
+    return result.length()
+  }
+
+  make_list(length) {
+    if length == 0 {
+      return nil
+    }
+
+    var e = Element(length)
+    e.next = self.make_list(length - 1)
+    return e
+  }
+
+  is_shorter_than(x, y) {
+    var x_tail = x
+    var y_tail = y
+
+    while y_tail != nil {
+      if x_tail == nil {
+        return true
+      }
+
+      x_tail = x_tail.next
+      y_tail = y_tail.next
+    }
+
+    return false
+  }
+
+  tail(x, y, z) {
+    if self.is_shorter_than(y, x) {
+      return self.tail(
+        self.tail(x.next, y, z),
+        self.tail(y.next, z, x),
+        self.tail(z.next, x, y)
+      )
+    }
+
+    return z
+  }
+}
+
+var start = microtime()
+echo List().benchmark() == 10
+var end = microtime()
+
+echo 'Time taken = ${(end - start) / 1000000} seconds'
\ No newline at end of file
diff --git a/benchmarks/bench-sieve.b b/benchmarks/bench-sieve.b
index 772dea13..233aa20d 100644
--- a/benchmarks/bench-sieve.b
+++ b/benchmarks/bench-sieve.b
@@ -1,4 +1,4 @@
-# The sieve benchmark counts the number of primes below 600000. 
+# The sieve benchmark counts the number of primes below 600000 based on the sieve of Eratosthenes.
 # The correct result is 49098.
 #
 # based on the JavaScript implementation at
diff --git a/benchmarks/bench-spectral-norm-2.b b/benchmarks/bench-spectral-norm-2.b
new file mode 100644
index 00000000..a937cf7f
--- /dev/null
+++ b/benchmarks/bench-spectral-norm-2.b
@@ -0,0 +1,101 @@
+/**
+ * The Computer Language Benchmarks Game
+ * https://salsa.debian.org/benchmarksgame-team/benchmarksgame/
+ * 
+ * contributed by Richard Ore
+ * based on contributed work by Ziad Hatahet
+ * based on the Go entry by K P anonymous
+ * 
+ * This implementation of the spectral norm employs multithreading 
+ * to achieve more than 10x the speed over the ordinary version 
+ * (depending on how many cores your cpu has).
+ */ 
+
+import thread
+
+def spectralnorm(n) {
+  if !n n = 100
+
+  var u = [1] * n
+  var v = [0] * n
+
+  for i in 0..10 {
+    a_times_transp(v, u)
+    a_times_transp(u, v)
+  }
+
+  var vbv = 0, vv = 0
+  for i in 0..n {
+    var vi = v[i]
+    vbv += u[i] * vi
+    vv += vi ** 2
+  }
+
+  return ((vbv / vv) ** 0.5)
+}
+
+def a_times_transp(v, u) {
+  var x = [0] * u.length()
+  var t = []
+  var ncpu = thread.cpu_count
+
+  for i in 0..ncpu {
+    t.append(thread.start(
+      times, 
+      [
+        x, 
+        i * v.length() / ncpu, 
+        (i + 1) * v.length() / ncpu, 
+        u, 
+        false,
+      ]
+    ))
+  }
+
+  for i in 0..ncpu {
+    t[i].await()
+  }
+
+  for i in 0..ncpu {
+    t[i] = thread.start(
+      times, 
+      [
+        v, 
+        i * v.length() / ncpu, 
+        (i + 1) * v.length() / ncpu, 
+        x, 
+        true,
+      ]
+    )
+  }
+
+  for i in 0..ncpu {
+    t[i].await()
+  }
+}
+
+def times(_, v, ii, n, u, transpose) {
+  var ul = u.length()
+  for i in ii..n {
+    var vi = 0
+    for j in 0..ul {
+      if transpose {
+        vi += u[j] / a(j, i)
+      } else {
+        vi += u[j] / a (i, j)
+      }
+    }
+
+    v[i] = vi
+  }
+}
+
+def a(i, j) {
+  return (i + j) * (i + j + 1) / 2 + i + 1
+}
+
+var start = microtime()
+echo spectralnorm(5500)
+var end = microtime()
+
+echo '\nTime taken = ${(end - start) / 1000000} seconds'
diff --git a/benchmarks/bench-spectral-norm.b b/benchmarks/bench-spectral-norm.b
index 8ad9547e..bb0f7638 100644
--- a/benchmarks/bench-spectral-norm.b
+++ b/benchmarks/bench-spectral-norm.b
@@ -1,15 +1,16 @@
 /**
  * The Computer Language Benchmarks Game
  * https://salsa.debian.org/benchmarksgame-team/benchmarksgame/
- * contributed by Ian Osgood
+ * 
+ * contributed by Richard Ore
+ * based on contributed work by Ian Osgood
  * modified for Node.js by Isaac Gouy
- * modified for Blade by Ore Richard
  */ 
 
 def A(i, j) {
-    return 1 /(
-        (i + j) * (i + j + 1 ) / 2 + i + 1
-    )
+  return 1 / (
+    (i + j) * (i + j + 1 ) / 2 + i + 1
+  )
 }
 
 def Au(u, v) {
@@ -37,7 +38,7 @@ def AtAu(u, v, w) {
 
 def spectralnorm(n) {
   var i, u = [0] * n, v = [0] * n, 
-        w = [0] * n, vv = 0, vBv = 0
+      w = [0] * n, vv = 0, vBv = 0
 
   for i in 0..n {
     u[i] = 1
diff --git a/libs/thread.b b/libs/thread.b
index eccaac70..7045c183 100644
--- a/libs/thread.b
+++ b/libs/thread.b
@@ -91,19 +91,27 @@
  */
 
 import _thread
+import _process
 
 import reflect
 import os
 
+
 var _MIN_STACK_SIZE = 16384 # 16kb
 var _DEFAULT_STACK_SIZE = 65536 # 64kb
-
 var _main_thread_id = _thread.get_id()
 
 def _is_not_main_thread(id) {id
   return id != _main_thread_id
 }
 
+
+/**
+ * The number of CPU cores available on the current device.
+ * @type number
+ */
+var cpu_count = _process.cpu_count
+
 /**
  * The thread class exposes methods to manage creating, running, 
  * and controlling threads.