diff --git a/src/lib/uninitialised-storage.hpp b/src/lib/uninitialised-storage.hpp
index 3f28f0d11..466582bba 100644
--- a/src/lib/uninitialised-storage.hpp
+++ b/src/lib/uninitialised-storage.hpp
@@ -137,7 +137,7 @@ namespace lib {
     {
       using _Arr = T[];
       
-      T* buff_{nullptr};
+      void*  buff_{nullptr};
       size_t size_{0};
       
     public:
@@ -146,9 +146,9 @@ namespace lib {
         {
           if (buff_) discard();
           size_ = cnt;
-          buff_ = cnt? static_cast<T*> (std::aligned_alloc (std::alignment_of<T>(), cnt * sizeof(T)))
+          buff_ = cnt? std::aligned_alloc (std::alignment_of<T>(), cnt * sizeof(T))
                      : nullptr;
-          return buff_;
+          return front();
         }
       
       void
@@ -216,6 +216,13 @@ namespace lib {
         }
       
       
+      T *      front()       { return &array()[0]; }
+      T const* front() const { return &array()[0]; }
+      T *      after()       { return &array()[size_];}
+      T const* after() const { return &array()[size_];}
+      T *      back ()       { return after() - 1; }
+      T const* back () const { return after() - 1; }
+      
       T &      operator[] (size_t idx)        { return array()[idx]; }
       T const& operator[] (size_t idx)  const { return array()[idx]; }
       
diff --git a/tests/vault/gear/test-chain-load-test.cpp b/tests/vault/gear/test-chain-load-test.cpp
index a54fc5835..1192822e2 100644
--- a/tests/vault/gear/test-chain-load-test.cpp
+++ b/tests/vault/gear/test-chain-load-test.cpp
@@ -872,6 +872,7 @@ SHOW_EXPR(testLoad.getHash())
         {
           ComputationalLoad cpuLoad;
           CHECK (cpuLoad.timeBase == 100us);
+          cpuLoad.useAllocation = true;
           
           double micros = cpuLoad.invoke();
 SHOW_EXPR(micros)
@@ -882,13 +883,13 @@ SHOW_EXPR(micros)
           
           micros = cpuLoad.invoke();
 SHOW_EXPR(micros)
-          CHECK (micros < 123);
+          CHECK (micros < 133);
           CHECK (micros > 80);
           
           micros = cpuLoad.benchmark();
 SHOW_EXPR(micros)
-          CHECK (micros < 105);
-          CHECK (micros > 95);
+          CHECK (micros < 110);
+          CHECK (micros > 90);
         }
       
       
diff --git a/tests/vault/gear/test-chain-load.hpp b/tests/vault/gear/test-chain-load.hpp
index ae7675b15..679786451 100644
--- a/tests/vault/gear/test-chain-load.hpp
+++ b/tests/vault/gear/test-chain-load.hpp
@@ -139,6 +139,7 @@ namespace test {
   using lib::meta::_FunRet;
   using lib::test::Transiently;
 
+  using std::make_pair;
   using std::forward;
   using std::string;
   using std::swap;
@@ -1079,23 +1080,29 @@ namespace test {
    */
   class ComputationalLoad
     {
-      lib::UninitialisedDynBlock<size_t> memBlock_{};
+      using Sink = volatile size_t;
+      
+      lib::UninitialisedDynBlock<Sink> memBlock_{};
       
       static double&
-      computationSpeed()    ///< in iterations/µs
+      computationSpeed (bool mem)     ///< in iterations/µs
         {
-          static double speed{LOAD_SPEED_BASELINE};
-          return speed;
+          static double cpuSpeed{LOAD_SPEED_BASELINE};
+          static double memSpeed{LOAD_SPEED_BASELINE};
+          return mem? memSpeed : cpuSpeed;
         }
       
     public:
       microseconds timeBase = 100us;
+      size_t       sizeBase = 1000;
       bool useAllocation = false;
       
       double
       invoke (uint scaleStep =1)
         {
-          return benchmarkTime ([this,scaleStep]{ causeComputationLoad(scaleStep); });
+          if (scaleStep == 0) return 0;
+          return useAllocation? benchmarkTime ([this,scaleStep]{ causeMemProcessLoad (scaleStep); })
+                              : benchmarkTime ([this,scaleStep]{ causeComputationLoad(scaleStep); });
         }
       
       double
@@ -1109,13 +1116,13 @@ namespace test {
       void
       calibrate()
         {
-cout<<">CAL: speed="<<computationSpeed()<<" rounds:"<<roundsNeeded(1)<<endl;
+cout<<">CAL: speed="<<computationSpeed(useAllocation)<<" rounds:"<<roundsNeeded(1)<<endl;
           auto speed = determineSpeed();
 cout<<".CAL: speed="<<speed<<endl;          
           speed = determineSpeed();
 cout<<".CAL: speed="<<speed<<endl;          
-          computationSpeed() = determineSpeed();
-cout<<"<CAL: speed="<<computationSpeed()<<" rounds:"<<roundsNeeded(1)<<endl;
+          computationSpeed(useAllocation) = determineSpeed();
+cout<<"<CAL: speed="<<computationSpeed(useAllocation)<<" rounds:"<<roundsNeeded(1)<<endl;
         }
       
       static void
@@ -1131,26 +1138,42 @@ cout<<"<CAL: speed="<<computationSpeed()<<" rounds:"<<roundsNeeded(1)<<endl;
       roundsNeeded (uint scaleStep)
         {
           auto desiredMicros = scaleStep*timeBase.count();
-          return uint64_t(desiredMicros*computationSpeed());
+          return uint64_t(desiredMicros*computationSpeed(useAllocation));
+        }
+      
+      auto
+      allocNeeded (uint scaleStep)
+        {
+          auto cnt = roundsNeeded(scaleStep);
+          auto siz = scaleStep * sizeBase;
+          // increase size to fit
+          siz = cnt /(cnt/siz);
+          cnt /= siz;
+          return make_pair (siz,cnt);
         }
       
       void
       causeComputationLoad (uint scaleStep)
         {
-          auto round = roundsNeeded(scaleStep);
-          volatile size_t sink;
+          auto round = roundsNeeded (scaleStep);
+          Sink sink;
           size_t scree;
           for ( ; 0 < round; --round)
-            scree = compute (scree);
+            boost::hash_combine (scree,scree);
           sink = scree;
           sink++;
         }
       
-      size_t
-      compute (size_t input)
+      void
+      causeMemProcessLoad (uint scaleStep)
         {
-          boost::hash_combine (input,input);
-          return input;
+          auto [siz,round] = allocNeeded (scaleStep);
+          memBlock_.allocate(siz);
+          ++*memBlock_.front();
+          for ( ; 0 < round; --round)
+            for (size_t i=0; i<memBlock_.size()-1; ++i)
+              memBlock_[i+1] += memBlock_[i];
+          ++*memBlock_.back();
         }
       
       double
@@ -1158,8 +1181,8 @@ cout<<"<CAL: speed="<<computationSpeed()<<" rounds:"<<roundsNeeded(1)<<endl;
         {
           uint step4gauge = 1;
           double micros   = benchmark (step4gauge);
-          auto roundsDone = roundsNeeded (step4gauge);
-          return roundsDone / micros;
+          auto stepsDone  = roundsNeeded (step4gauge);
+          return stepsDone / micros;
         }
     };
   
diff --git a/wiki/thinkPad.ichthyo.mm b/wiki/thinkPad.ichthyo.mm
index 702171525..03aff9f96 100644
--- a/wiki/thinkPad.ichthyo.mm
+++ b/wiki/thinkPad.ichthyo.mm
@@ -99550,15 +99550,16 @@ Date:&#160;&#160;&#160;Thu Apr 20 18:53:17 2023 +0200<br/>
 <node COLOR="#6b3e25" CREATED="1702180522766" ID="ID_316625699" MODIFIED="1702245071598" TEXT="ComputationLoad">
 <font NAME="Monospaced" SIZE="13"/>
 </node>
-<node BACKGROUND_COLOR="#f0d5c5" COLOR="#990033" CREATED="1702238322430" ID="ID_891650660" MODIFIED="1702238331269" TEXT="verwendete Techologie?">
+<node COLOR="#435e98" CREATED="1702238322430" ID="ID_891650660" MODIFIED="1702260278758" TEXT="verwendete Techologie?">
 <icon BUILTIN="help"/>
 <node COLOR="#338800" CREATED="1702238333843" ID="ID_1192309298" MODIFIED="1702245044607" TEXT="reine computation-Loop">
 <icon BUILTIN="button_ok"/>
 </node>
-<node BACKGROUND_COLOR="#eee5c3" COLOR="#990000" CREATED="1702238371167" ID="ID_767333201" MODIFIED="1702238496567" TEXT="oder eine etwas realistischere Last">
-<icon BUILTIN="flag-yellow"/>
+<node COLOR="#338800" CREATED="1702238371167" ID="ID_767333201" MODIFIED="1702260275694" TEXT="oder eine etwas realistischere Last">
+<icon BUILTIN="button_ok"/>
 <node CREATED="1702238388101" ID="ID_1772254921" MODIFIED="1702238391640" TEXT="Heap-Allokation"/>
 <node CREATED="1702238392652" ID="ID_1394905064" MODIFIED="1702238396495" TEXT="Speicher-Zugriffe"/>
+<node CREATED="1702260259872" ID="ID_1938750699" MODIFIED="1702260272219" TEXT="realisiere hier scaleStep durch gr&#xf6;&#xdf;ere Allokation"/>
 </node>
 <node BACKGROUND_COLOR="#ccb59b" COLOR="#6e2a38" CREATED="1702238410562" ID="ID_1357630846" MODIFIED="1702238431601" TEXT="kann ich nicht rein-theoretisch enscheiden">
 <font ITALIC="true" NAME="SansSerif" SIZE="14"/>
@@ -99579,6 +99580,23 @@ Date:&#160;&#160;&#160;Thu Apr 20 18:53:17 2023 +0200<br/>
 <icon BUILTIN="forward"/>
 </node>
 </node>
+<node CREATED="1702252767113" ID="ID_688715635" MODIFIED="1702252791529">
+<richcontent TYPE="NODE"><html>
+  <head>
+    
+  </head>
+  <body>
+    <p>
+      legt die Eichungs-Konstanten <i><font color="#152eb7">statisch </font></i>ab
+    </p>
+  </body>
+</html>
+</richcontent>
+<node CREATED="1702252795573" ID="ID_1426973621" MODIFIED="1702252808439" TEXT="wichtig f&#xfc;r effizienten Einsatz in Lasttest-Szenarien"/>
+<node CREATED="1702252814273" ID="ID_184752238" MODIFIED="1702252838764" TEXT="mu&#xdf; daf&#xfc;r eine gewisse Koh&#xe4;renz der Eichung gew&#xe4;hrleisten">
+<icon BUILTIN="yes"/>
+</node>
+</node>
 </node>
 <node CREATED="1702176870348" ID="ID_1843992678" MODIFIED="1702176878242" TEXT="soll statisch kalibrierbar sein">
 <node CREATED="1702176883018" ID="ID_1063649357" MODIFIED="1702177162816" TEXT="auf 1 LoadUnit &#x2259; 500ms"/>
@@ -99600,6 +99618,9 @@ Date:&#160;&#160;&#160;Thu Apr 20 18:53:17 2023 +0200<br/>
 <node CREATED="1702232438440" ID="ID_1878954665" MODIFIED="1702232455960" TEXT="man mu&#xdf; f&#xfc;r die jeweiligen Einstellungen (re)kalibrieren"/>
 <node CREATED="1702232412226" HGAP="34" ID="ID_43541079" MODIFIED="1702232463171" TEXT="damit werden nicht-Linearit&#xe4;ten ausgeschaltet" VSHIFT="-7"/>
 </node>
+<node BACKGROUND_COLOR="#eee5c3" COLOR="#990000" CREATED="1702260328321" ID="ID_1651629984" MODIFIED="1702260339078" TEXT="beide Varianten inkrementell kalibrieren">
+<icon BUILTIN="flag-yellow"/>
+</node>
 </node>
 <node BACKGROUND_COLOR="#eee5c3" COLOR="#990000" CREATED="1702180295927" ID="ID_773261601" MODIFIED="1702180308206" TEXT="Komponente testen">
 <icon BUILTIN="flag-yellow"/>
@@ -99610,7 +99631,7 @@ Date:&#160;&#160;&#160;Thu Apr 20 18:53:17 2023 +0200<br/>
 <node BACKGROUND_COLOR="#eef0c5" COLOR="#990000" CREATED="1702180446234" ID="ID_257805938" MODIFIED="1702245405826" TEXT="Last und Eichung">
 <arrowlink COLOR="#c26a36" DESTINATION="ID_168687387" ENDARROW="Default" ENDINCLINATION="-214;715;" ID="Arrow_ID_917319622" STARTARROW="None" STARTINCLINATION="-169;-474;"/>
 <icon BUILTIN="pencil"/>
-<node BACKGROUND_COLOR="#f0d5c5" COLOR="#990033" CREATED="1702245098652" ID="ID_1449908056" MODIFIED="1702245403037" TEXT="Fehler im Test?">
+<node BACKGROUND_COLOR="#f0d5c5" COLOR="#990033" CREATED="1702245098652" ID="ID_1449908056" MODIFIED="1702260306418" TEXT="Fehler im Test?">
 <richcontent TYPE="NOTE"><html>
   <head>
     
@@ -99626,7 +99647,7 @@ Date:&#160;&#160;&#160;Thu Apr 20 18:53:17 2023 +0200<br/>
       
     </p>
     <p>
-      Stelle schon mal fest: der Einzel-Lauf streut ganz deutlich. Und Cache-Effekte k&#246;nnten auch noch ein gewisses Problem darstellen (wenngleich auch der vorangegangene Test grunds&#228;tzlich den Code schon <i>vorgew&#228;rmt</i>&#160; hat. Also setze mal&#160;&#160;&#177;20% f&#252;r den Einzeltest an, aber nur &#177;5% f&#252;r den Benchmark
+      Stelle schon mal fest: der Einzel-Lauf streut ganz deutlich. Und Cache-Effekte k&#246;nnten auch noch ein gewisses Problem darstellen (wenngleich auch der vorangegangene Test grunds&#228;tzlich den Code schon <i>vorgew&#228;rmt</i>&#160; hat. Also setze mal&#160;&#160;&#177;30% f&#252;r den Einzeltest an, aber nur &#177;10% f&#252;r den Benchmark
     </p>
   </body>
 </html>