refactor: optimize truncate_stack procedure

stdlib/asm/sys.masm

@@ -4,27 +4,23 @@
 #! function at the end will reduce the size of the public inputs that are shared with the verifier.
 #! Input: Stack with 16 or more elements.
 #! Output: Stack with only the original top 16 elements.
-export.truncate_stack.4
-    loc_storew.0
-    dropw
-    loc_storew.1
-    dropw
-    loc_storew.2
-    dropw
-    loc_storew.3
-    dropw
-    sdepth
-    neq.16
+#! Cycles: 17 + 11 * overflow_words, where `overflow_words` is the number of words needed to drop.
+export.truncate_stack.1
+    # save the first word to memory and bring elements to be dropped to the top of the stack
+    loc_storew.0 dropw movupw.3
+    # => [X, B, C, D, ...]
+
+    # until stack depth greater than 16, keep dropping extra elements
+    sdepth neq.16
     while.true
-        dropw
-        sdepth
-        neq.16
+        dropw movupw.3
+        # => [X, B, C, D, ...]
+
+        sdepth neq.16
     end
-    loc_loadw.3
-    swapw.3
-    loc_loadw.2
-    swapw.2
-    loc_loadw.1
-    swapw.1
+    # => [X, B, C, D, ...]
+
+    # bring the previously saved word back onto the stack
     loc_loadw.0
+    # => [A, B, C, D, ...]
 end

stdlib/docs/sys.md

@@ -1,4 +1,4 @@
-Removes elements deep in the stack until the depth of the stack is exactly 16. The elements<br />are removed in such a way that the top 16 elements of the stack remain unchanged. If the stack<br />would otherwise contain more than 16 elements at the end of execution, then adding a call to this<br />function at the end will reduce the size of the public inputs that are shared with the verifier.<br />Input: Stack with 16 or more elements.<br />Output: Stack with only the original top 16 elements.<br />
+Removes elements deep in the stack until the depth of the stack is exactly 16. The elements<br />are removed in such a way that the top 16 elements of the stack remain unchanged. If the stack<br />would otherwise contain more than 16 elements at the end of execution, then adding a call to this<br />function at the end will reduce the size of the public inputs that are shared with the verifier.<br />Input: Stack with 16 or more elements.<br />Output: Stack with only the original top 16 elements.<br />Cycles: 17 + 11 * overflow_words, where `overflow_words` is the number of words needed to drop.<br />
 ## std::sys
 | Procedure | Description |
 | ----------- | ------------- |