diff --git a/src/engine/engine.h b/src/engine/engine.h
index 3a5b1b044..ba2cf567b 100644
--- a/src/engine/engine.h
+++ b/src/engine/engine.h
@@ -39,6 +39,8 @@
 #include <unordered_map>
 #include <deque>
 
+class DivWorkPool;
+
 #define addWarning(x) \
   if (warnings.empty()) { \
     warnings+=x; \
@@ -485,6 +487,8 @@ class DivEngine {
 
   size_t totalProcessed;
 
+  DivWorkPool* renderPool;
+
   // MIDI stuff
   std::function<int(const TAMidiMessage&)> midiCallback=[](const TAMidiMessage&) -> int {return -2;};
 
diff --git a/src/engine/playback.cpp b/src/engine/playback.cpp
index 270dd7f79..0b21dc57f 100644
--- a/src/engine/playback.cpp
+++ b/src/engine/playback.cpp
@@ -22,6 +22,7 @@
 #define _USE_MATH_DEFINES
 #include "dispatch.h"
 #include "engine.h"
+#include "workPool.h"
 #include "../ta-log.h"
 #include <math.h>
 
@@ -1788,6 +1789,11 @@ void DivEngine::nextBuf(float** in, float** out, int inChans, int outChans, unsi
 
   std::chrono::steady_clock::time_point ts_processBegin=std::chrono::steady_clock::now();
 
+  if (renderPool==NULL) {
+    // TODO: test this
+    renderPool=new DivWorkPool(0);
+  }
+
   // process MIDI events (TODO: everything)
   if (output) if (output->midiIn) while (!output->midiIn->queue.empty()) {
     TAMidiMessage& msg=output->midiIn->queue.front();
@@ -2061,20 +2067,28 @@ void DivEngine::nextBuf(float** in, float** out, int inChans, int outChans, unsi
         // 5. tick the clock and fill buffers as needed
         if (cycles<runLeftG) {
           for (int i=0; i<song.systemLen; i++) {
-            int total=(cycles*disCont[i].runtotal)/(size<<MASTER_CLOCK_PREC);
-            disCont[i].acquire(disCont[i].runPos,total);
-            disCont[i].runLeft-=total;
-            disCont[i].runPos+=total;
+            renderPool->push([this,size](void* d) {
+              DivDispatchContainer* dc=(DivDispatchContainer*)d;
+              int total=(cycles*dc->runtotal)/(size<<MASTER_CLOCK_PREC);
+              dc->acquire(dc->runPos,total);
+              dc->runLeft-=total;
+              dc->runPos+=total;
+            },&disCont[i]);
           }
+          renderPool->wait();
           runLeftG-=cycles;
           cycles=0;
         } else {
           cycles-=runLeftG;
           runLeftG=0;
           for (int i=0; i<song.systemLen; i++) {
-            disCont[i].acquire(disCont[i].runPos,disCont[i].runLeft);
-            disCont[i].runLeft=0;
+            renderPool->push([](void* d) {
+              DivDispatchContainer* dc=(DivDispatchContainer*)d;
+              dc->acquire(dc->runPos,dc->runLeft);
+              dc->runLeft=0;
+            },&disCont[i]);
           }
+          renderPool->wait();
         }
       }
     }
diff --git a/src/engine/workPool.cpp b/src/engine/workPool.cpp
index a62f4ff7a..fa67aeb50 100644
--- a/src/engine/workPool.cpp
+++ b/src/engine/workPool.cpp
@@ -50,7 +50,9 @@ void DivWorkThread::run() {
 
       task.func(task.funcArg);
 
-      parent->busyCount--;
+      if (--parent->busyCount<0) {
+        logE("oh no PROBLEM...");
+      }
       parent->notify.notify_one();
     }
   }
@@ -125,8 +127,8 @@ void DivWorkPool::wait() {
   }
 
   // wait
-  while (busyCount!=0) {
-    if (notify.wait_for(unique,std::chrono::milliseconds(100))==std::cv_status::timeout) {
+  while (busyCount>0) {
+    if (notify.wait_for(unique,std::chrono::milliseconds(30))==std::cv_status::timeout) {
       logW("DivWorkPool: wait() timed out!");
     }
   }