diff --git a/bochs/cpu/fetchdecode.cc b/bochs/cpu/fetchdecode.cc
index 48a8a66fa..318b825c6 100644
--- a/bochs/cpu/fetchdecode.cc
+++ b/bochs/cpu/fetchdecode.cc
@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////
-// $Id: fetchdecode.cc,v 1.193 2008-06-23 02:56:31 sshwarts Exp $
+// $Id: fetchdecode.cc,v 1.194 2008-06-25 10:34:20 sshwarts Exp $
 /////////////////////////////////////////////////////////////////////////
 //
 //  Copyright (C) 2001  MandrakeSoft S.A.
@@ -2931,7 +2931,10 @@ modrm_done:
   i->setB1(b1);
   i->setILen(ilen);
 
+#if BX_INSTRUMENTATION
+  i->ia_opcode = ia_opcode;
   BX_INSTR_FETCH_DECODE_COMPLETED(BX_CPU_ID, i);
+#endif
 
   return(1);
 }
diff --git a/bochs/cpu/fetchdecode64.cc b/bochs/cpu/fetchdecode64.cc
index ad5e653a4..1d634b80e 100644
--- a/bochs/cpu/fetchdecode64.cc
+++ b/bochs/cpu/fetchdecode64.cc
@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////
-// $Id: fetchdecode64.cc,v 1.202 2008-06-23 02:56:31 sshwarts Exp $
+// $Id: fetchdecode64.cc,v 1.203 2008-06-25 10:34:20 sshwarts Exp $
 /////////////////////////////////////////////////////////////////////////
 //
 //  Copyright (C) 2001  MandrakeSoft S.A.
@@ -3852,7 +3852,10 @@ modrm_done:
   i->setB1(b1);
   i->setILen(ilen);
 
+#if BX_INSTRUMENTATION
+  i->ia_opcode = ia_opcode;
   BX_INSTR_FETCH_DECODE_COMPLETED(BX_CPU_ID, i);
+#endif
 
   return(1);
 }
diff --git a/bochs/cpu/instr.h b/bochs/cpu/instr.h
index da4b1b092..8b1368b24 100755
--- a/bochs/cpu/instr.h
+++ b/bochs/cpu/instr.h
@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////
-// $Id: instr.h,v 1.11 2008-04-14 21:48:35 sshwarts Exp $
+// $Id: instr.h,v 1.12 2008-06-25 10:34:21 sshwarts Exp $
 /////////////////////////////////////////////////////////////////////////
 //
 //   Copyright (c) 2008 Stanislav Shwartsman
@@ -45,6 +45,9 @@ public:
   // the memory address (if any).
   BxExecutePtr_tR ResolveModrm;
   BxExecutePtr_tR execute;
+#if BX_INSTRUMENTATION
+  Bit16u ia_opcode;
+#endif
 
   struct {
     //  7...2 (unused)
diff --git a/bochs/cpu/io.cc b/bochs/cpu/io.cc
index 631eec45f..2c745d52f 100644
--- a/bochs/cpu/io.cc
+++ b/bochs/cpu/io.cc
@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////
-// $Id: io.cc,v 1.61 2008-06-12 20:12:25 sshwarts Exp $
+// $Id: io.cc,v 1.62 2008-06-25 10:34:21 sshwarts Exp $
 /////////////////////////////////////////////////////////////////////////
 //
 //  Copyright (C) 2001  MandrakeSoft S.A.
@@ -341,59 +341,25 @@ void BX_CPP_AttrRegparmN(1) BX_CPU_C::INSW16_YwDX(bxInstruction_c *i)
 {
   Bit16u value16=0;
   Bit16u di = DI;
-  unsigned incr = 2;
 
   if (! BX_CPU_THIS_PTR allow_io(DX, 2)) {
     BX_DEBUG(("INSW16_YwDX: I/O access not allowed !"));
     exception(BX_GP_EXCEPTION, 0, 0);
   }
 
-#if (BX_SupportRepeatSpeedups) && (BX_DEBUGGER == 0)
-  /* If conditions are right, we can transfer IO to physical memory
-   * in a batch, rather than one instruction at a time.
-   */
-  if (i->repUsedL() && !BX_CPU_THIS_PTR async_event)
-  {
-    Bit32u wordCount = CX;
-    BX_ASSERT(wordCount > 0);
-    wordCount = FastRepINSW(i, di, DX, wordCount);
-    if (wordCount) {
-      // Decrement the ticks count by the number of iterations, minus
-      // one, since the main cpu loop will decrement one.  Also,
-      // the count is predecremented before examined, so defintely
-      // don't roll it under zero.
-      BX_TICKN(wordCount-1);
-      CX -= (wordCount-1);
-      incr = wordCount << 1; // count * 2
-    }
-    else {
-      // Write a zero to memory, to trigger any segment or page
-      // faults before reading from IO port.
-      write_virtual_word_32(BX_SEG_REG_ES, di, value16);
+  // Write a zero to memory, to trigger any segment or page
+  // faults before reading from IO port.
+  write_virtual_word_32(BX_SEG_REG_ES, di, value16);
 
-      value16 = BX_INP(DX, 2);
+  value16 = BX_INP(DX, 2);
 
-      /* no seg override allowed */
-      write_virtual_word_32(BX_SEG_REG_ES, di, value16);
-    }
-  }
-  else
-#endif
-  {
-    // Write a zero to memory, to trigger any segment or page
-    // faults before reading from IO port.
-    write_virtual_word_32(BX_SEG_REG_ES, di, value16);
-
-    value16 = BX_INP(DX, 2);
-
-    /* no seg override allowed */
-    write_virtual_word_32(BX_SEG_REG_ES, di, value16);
-  }
+  /* no seg override allowed */
+  write_virtual_word_32(BX_SEG_REG_ES, di, value16);
 
   if (BX_CPU_THIS_PTR get_DF())
-    DI -= incr;
+    DI -= 2;
   else
-    DI += incr;
+    DI += 2;
 }
 
 // 16-bit operand size, 32-bit address size
@@ -703,40 +669,13 @@ void BX_CPP_AttrRegparmN(1) BX_CPU_C::OUTSW16_DXXw(bxInstruction_c *i)
     exception(BX_GP_EXCEPTION, 0, 0);
   }
 
-  Bit16u value16;
-  Bit16u si = SI;
-  unsigned incr = 2;
-
-#if (BX_SupportRepeatSpeedups) && (BX_DEBUGGER == 0)
-  /* If conditions are right, we can transfer IO to physical memory
-   * in a batch, rather than one instruction at a time.
-   */
-  if (i->repUsedL() && !BX_CPU_THIS_PTR async_event) {
-    Bit32u wordCount = CX;
-    wordCount = FastRepOUTSW(i, i->seg(), si, DX, wordCount);
-    if (wordCount) {
-      // Decrement eCX.  Note, the main loop will decrement 1 also, so
-      // decrement by one less than expected, like the case above.
-      BX_TICKN(wordCount-1); // Main cpu loop also decrements one more.
-      CX -= (wordCount-1);
-      incr = wordCount << 1; // count * 2.
-    }
-    else {
-      value16 = read_virtual_word_32(i->seg(), si);
-      BX_OUTP(DX, value16, 2);
-    }
-  }
-  else
-#endif
-  {
-    value16 = read_virtual_word_32(i->seg(), si);
-    BX_OUTP(DX, value16, 2);
-  }
+  Bit16u value16 = read_virtual_word_32(i->seg(), SI);
+  BX_OUTP(DX, value16, 2);
 
   if (BX_CPU_THIS_PTR get_DF())
-    SI = SI - incr;
+    SI -= 2;
   else
-    SI = SI + incr;
+    SI += 2;
 }
 
 // 16-bit operand size, 32-bit address size
diff --git a/bochs/cpu/proc_ctrl.cc b/bochs/cpu/proc_ctrl.cc
index 14c0b93b1..1aa47abd9 100644
--- a/bochs/cpu/proc_ctrl.cc
+++ b/bochs/cpu/proc_ctrl.cc
@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////
-// $Id: proc_ctrl.cc,v 1.245 2008-06-25 02:28:31 sshwarts Exp $
+// $Id: proc_ctrl.cc,v 1.246 2008-06-25 10:34:21 sshwarts Exp $
 /////////////////////////////////////////////////////////////////////////
 //
 //  Copyright (C) 2001  MandrakeSoft S.A.
@@ -1923,7 +1923,8 @@ void BX_CPP_AttrRegparmN(1) BX_CPU_C::MONITOR(bxInstruction_c *i)
 
   // check if we could access the memory segment
   if (!(seg->cache.valid & SegAccessROK4G)) {
-    read_virtual_checks(&BX_CPU_THIS_PTR sregs[i->seg()], offset, 1);
+    if (! read_virtual_checks(&BX_CPU_THIS_PTR sregs[i->seg()], offset, 1))
+      exception(int_number(i->seg()), 0, 0);
   }
 
   // set MONITOR
diff --git a/bochs/cpu/string.cc b/bochs/cpu/string.cc
index d68d5fd14..ed2b30646 100644
--- a/bochs/cpu/string.cc
+++ b/bochs/cpu/string.cc
@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////
-// $Id: string.cc,v 1.61 2008-06-12 19:14:39 sshwarts Exp $
+// $Id: string.cc,v 1.62 2008-06-25 10:34:21 sshwarts Exp $
 /////////////////////////////////////////////////////////////////////////
 //
 //  Copyright (C) 2001  MandrakeSoft S.A.
@@ -609,50 +609,18 @@ void BX_CPP_AttrRegparmN(1) BX_CPU_C::REP_MOVSQ_XqYq(bxInstruction_c *i)
 // 16 bit address size
 void BX_CPP_AttrRegparmN(1) BX_CPU_C::MOVSB16_XbYb(bxInstruction_c *i)
 {
-  Bit8u temp8;
-
-  Bit32u incr = 1;
-
-#if (BX_SupportRepeatSpeedups) && (BX_DEBUGGER == 0)
-  /* If conditions are right, we can transfer IO to physical memory
-   * in a batch, rather than one instruction at a time */
-  if (i->repUsedL() && !BX_CPU_THIS_PTR async_event)
-  {
-    Bit32u byteCount = FastRepMOVSB(i, i->seg(), SI, BX_SEG_REG_ES, DI, CX);
-    if (byteCount) {
-      // Decrement the ticks count by the number of iterations, minus
-      // one, since the main cpu loop will decrement one.  Also,
-      // the count is predecremented before examined, so defintely
-      // don't roll it under zero.
-      BX_TICKN(byteCount-1);
-
-      // Decrement eCX. Note, the main loop will decrement 1 also, so
-      // decrement by one less than expected, like the case above.
-      CX -= (byteCount-1);
-
-      incr = byteCount;
-    }
-    else {
-      temp8 = read_virtual_byte(i->seg(), SI);
-      write_virtual_byte_32(BX_SEG_REG_ES, DI, temp8);
-    }
-  }
-  else
-#endif
-  {
-    temp8 = read_virtual_byte_32(i->seg(), SI);
-    write_virtual_byte_32(BX_SEG_REG_ES, DI, temp8);
-  }
+  Bit8u temp8 = read_virtual_byte_32(i->seg(), SI);
+  write_virtual_byte_32(BX_SEG_REG_ES, DI, temp8);
 
   if (BX_CPU_THIS_PTR get_DF()) {
     /* decrement SI, DI */
-    SI -= incr;
-    DI -= incr;
+    SI--;
+    DI--;
   }
   else {
     /* increment SI, DI */
-    SI += incr;
-    DI += incr;
+    SI++;
+    DI++;
   }
 }
 
@@ -735,54 +703,21 @@ void BX_CPP_AttrRegparmN(1) BX_CPU_C::MOVSB64_XbYb(bxInstruction_c *i)
 /* 16 bit opsize mode, 16 bit address size */
 void BX_CPP_AttrRegparmN(1) BX_CPU_C::MOVSW16_XwYw(bxInstruction_c *i)
 {
-  Bit16u temp16;
-
-  Bit32u incr = 2;
-
   Bit16u si = SI;
   Bit16u di = DI;
 
-#if (BX_SupportRepeatSpeedups) && (BX_DEBUGGER == 0)
-  /* If conditions are right, we can transfer IO to physical memory
-   * in a batch, rather than one instruction at a time.
-   */
-  if (i->repUsedL() && !BX_CPU_THIS_PTR async_event)
-  {
-    Bit32u wordCount = FastRepMOVSW(i, i->seg(), si, BX_SEG_REG_ES, di, CX);
-    if (wordCount) {
-      // Decrement the ticks count by the number of iterations, minus
-      // one, since the main cpu loop will decrement one.  Also,
-      // the count is predecremented before examined, so defintely
-      // don't roll it under zero.
-      BX_TICKN(wordCount-1);
-
-      // Decrement eCX. Note, the main loop will decrement 1 also, so
-      // decrement by one less than expected, like the case above.
-      CX -= (wordCount-1);
-
-      incr = wordCount << 1; // count * 2
-    }
-    else {
-      temp16 = read_virtual_word_32(i->seg(), si);
-      write_virtual_word_32(BX_SEG_REG_ES, di, temp16);
-    }
-  }
-  else
-#endif
-  {
-    temp16 = read_virtual_word_32(i->seg(), si);
-    write_virtual_word_32(BX_SEG_REG_ES, di, temp16);
-  }
+  Bit16u temp16 = read_virtual_word_32(i->seg(), si);
+  write_virtual_word_32(BX_SEG_REG_ES, di, temp16);
 
   if (BX_CPU_THIS_PTR get_DF()) {
     /* decrement SI, DI */
-    si -= incr;
-    di -= incr;
+    si -= 2;
+    di -= 2;
   }
   else {
     /* increment SI, DI */
-    si += incr;
-    di += incr;
+    si += 2;
+    di += 2;
   }
 
   SI = si;