diff --git a/gc.c b/gc.c
index 213e5d83..7d4500b0 100644
--- a/gc.c
+++ b/gc.c
@@ -71,8 +71,6 @@ void *gc_try_alloc(gc_heap *h, size_t size)
         } else { /* Take the whole chunk */
           f1->next = f2->next;
         }
-        // zero-out the header
-        //memset((object)f2, 0, sizeof(gc_header_type));
         return f2;
       }
     }
@@ -216,7 +214,9 @@ size_t gc_sweep(gc_heap *h, size_t *sum_freed_ptr)
   object p, end;
   gc_free_list *q, *r, *s;
   for (; h; h = h->next) { // All heaps
-fprintf(stdout, "sweep heap %p, size = %d\n", h, h->size);
+#if GC_DEBUG_CONCISE_PRINTFS
+    fprintf(stdout, "sweep heap %p, size = %d\n", h, h->size);
+#endif
     p = gc_heap_first_block(h);
     q = h->free_list;
     end = gc_heap_end(h);
@@ -234,7 +234,7 @@ fprintf(stdout, "sweep heap %p, size = %d\n", h, h->size);
       size = gc_heap_align(gc_allocated_bytes(p));
 //fprintf(stdout, "check object %p, size = %d\n", p, size);
       
-//#if GC_DEBUG_PRINTFS
+#if GC_DEBUG_CONCISE_PRINTFS
       // DEBUG
       if (!is_object_type(p))
         fprintf(stderr, "sweep: invalid object at %p", p);
@@ -243,7 +243,7 @@ fprintf(stdout, "sweep heap %p, size = %d\n", h, h->size);
       if (r && ((char *)p) + size > (char *)r)
         fprintf(stderr, "sweep: bad size at %p + %d > %p", p, size, r);
       // END DEBUG
-//#endif
+#endif
 
       if (!mark(p)) {
 #if GC_DEBUG_PRINTFS
@@ -311,7 +311,9 @@ void gc_thr_grow_move_buffer(gc_thread_data *d)
   }
 
   d->moveBuf = realloc(d->moveBuf, d->moveBufLen * sizeof(void *));
+#if GC_DEBUG_CONCISE_PRINTFS
   printf("grew moveBuffer, len = %d\n", d->moveBufLen);
+#endif
 }
 
 void gc_thr_add_to_move_buffer(gc_thread_data *d, int *alloci, object obj)
diff --git a/include/cyclone/types.h b/include/cyclone/types.h
index 9768aec1..ecf58721 100644
--- a/include/cyclone/types.h
+++ b/include/cyclone/types.h
@@ -83,6 +83,7 @@ void gc_thr_add_to_move_buffer(gc_thread_data *d, int *alloci, object obj);
 /* GC debugging flags */
 //#define DEBUG_GC 0
 #define GC_DEBUG_PRINTFS 0
+#define GC_DEBUG_CONCISE_PRINTFS 0
 
 /* Show diagnostic information for the GC when program terminates */
 #define DEBUG_SHOW_DIAG 0
@@ -93,13 +94,14 @@ void gc_thr_add_to_move_buffer(gc_thread_data *d, int *alloci, object obj);
 /* Which way does the CPU grow its stack? */
 #define STACK_GROWS_DOWNWARD 1
 
-/* Size of the stack buffer, in bytes.           */
+/* Size of the stack buffer, in bytes.
+   This is used as the first generation of the GC. */
 #define STACK_SIZE 250000
 
-/* Size of the 2nd generation, in bytes. */
+/* Size of a "page" on the heap (the 2nd generation), in bytes. */
 #define HEAP_SIZE 6000000
 
-/* Define size of object tags.  Options are "short" or "long". */
+/* Define size of object tags */
 typedef long tag_type;
 
 #ifndef CLOCKS_PER_SEC
@@ -150,7 +152,7 @@ typedef long tag_type;
    have extra least significant bits that can be used to mark them as
    values instead of objects (IE, pointer to a tagged object).
    On many machines, addresses are multiples of four, leaving the two
-   least significant bits free - according to lisp in small pieces.
+   least significant bits free - from lisp in small pieces.
 */
 #define obj_is_char(x)  ((unsigned long)(x) & (unsigned long)1)
 #define obj_obj2char(x) (char)((long)(x)>>1)
diff --git a/runtime.c b/runtime.c
index a3c8477e..cd47312f 100644
--- a/runtime.c
+++ b/runtime.c
@@ -10,7 +10,7 @@
 #include "cyclone/runtime.h"
 
 //int JAE_DEBUG = 0;
-int gcMoveCountsDEBUG[20] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
+//int gcMoveCountsDEBUG[20] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0};
 
 /* Error checking section - type mismatch, num args, etc */
 /* Type names to use for error messages */
@@ -2350,22 +2350,21 @@ void Cyc_apply_from_buf(int argc, object prim, object *buf) {
 // longjmp(jmp_main,1); /* Return globals gc_cont, gc_ans. */
 //}
 
-// NEW GC algorithm
-//
-// TODO: not quite sure when to call this function. might want to set a flag
-// if the heap was expanded during alloc, and after all the allocs are done
-// after a minor GC (or during the minor GC), call into this function to
-// free up unused space
+// Collect garbage using mark&sweep algorithm
+// Note non-global roots should be marked prior to calling this function.
 size_t gc_collect(gc_heap *h, size_t *sum_freed) 
 {
+#if GC_DEBUG_CONCISE_PRINTFS
   printf("(heap: %p size: %d)\n", h, (unsigned int)gc_heap_total_size(h));
+#endif
   // Mark global variables
-  gc_mark(h, Cyc_global_variables); /* Internal global used by the runtime */
+  gc_mark(h, Cyc_global_variables); // Internal global used by the runtime
+                                    // Marking it ensures all glos are marked
   {
     list l = global_table;
     for(; !nullp(l); l = cdr(l)){
      cvar_type *c = (cvar_type *)car(l);
-     gc_mark(h, *(c->pvar)); // Mark global, not the pvar
+     gc_mark(h, *(c->pvar)); // Mark actual object the global points to
     }
   }
   // TODO: what else to mark? gc_mark(
@@ -2373,8 +2372,7 @@ size_t gc_collect(gc_heap *h, size_t *sum_freed)
   // weak refs?
   // finalize?
   return gc_sweep(h, sum_freed);
-  // debug print free stats
-  // return value from sweep??
+  // debug print free stats?
 }
 
 // TODO: move globals to thread-specific structures.
@@ -2383,7 +2381,7 @@ size_t gc_collect(gc_heap *h, size_t *sum_freed)
 char *gc_move(char *obj, gc_thread_data *thd, int *alloci, int *heap_grown) {
   if (!is_object_type(obj)) return obj;
 
-gcMoveCountsDEBUG[type_of(obj)]++;
+//gcMoveCountsDEBUG[type_of(obj)]++;
 
 //printf("DEBUG gc_move type = %ld\n", type_of(obj)); // JAE DEBUG
   switch(type_of(obj)){
@@ -2715,19 +2713,24 @@ void GC(cont, args, num_args) closure cont; object *args; int num_args;
 
   // Check if we need to do a major GC
   if (heap_grown) {
-    time_t majorStart = time(NULL);
     size_t freed = 0, max_freed = 0;
-fprintf(stdout, "DEBUG, starting major mark/sweep GC\n"); // JAE DEBUG
+#if GC_DEBUG_CONCISE_PRINTFS
+    time_t majorStart = time(NULL);
+    fprintf(stdout, "DEBUG, starting major mark/sweep GC\n"); // JAE DEBUG
+#endif
     gc_mark(Cyc_heap, cont);
     for (i = 0; i < num_args; i++){ 
       gc_mark(Cyc_heap, args[i]);
     }
     max_freed = gc_collect(Cyc_heap, &freed);
-printf("done, freed = %d, max_freed = %d, elapsed = %ld\n", freed, max_freed, time(NULL) - majorStart);
-//JAE_DEBUG++;
-//if (JAE_DEBUG == 2) exit(1); // JAE DEBUG
-for (i = 0; i < 20; i++){
-  printf("gcMoveCountsDEBUG[%d] = %d\n", i, gcMoveCountsDEBUG[i]);}
+#if GC_DEBUG_CONCISE_PRINTFS
+    printf("done, freed = %d, max_freed = %d, elapsed = %ld\n", freed, max_freed, time(NULL) - majorStart);
+    //JAE_DEBUG++;
+    //if (JAE_DEBUG == 2) exit(1); // JAE DEBUG
+    for (i = 0; i < 20; i++){
+      printf("gcMoveCountsDEBUG[%d] = %d\n", i, gcMoveCountsDEBUG[i]);
+    }
+#endif
   }
 
 //fprintf(stdout, "DEBUG, finished minor GC\n"); // JAE DEBUG