- update to current root trunc's version

- update to root trunc's UNICODE support

git-svn-id: svn://ttmath.org/publicrep/ttmath/branches/chk@182 e52654a7-88a9-db11-a3e9-0013d4bc506e

CHANGELOG

@@ -1,3 +1,86 @@
+Version 0.9.0 prerelease (2009.07.16):
+    * added:   support for wide characters (wchar_t)
+               wide characters are used when macro TTMATH_USE_WCHAR is defined
+               this macro is defined automatically when there is macro UNICODE or _UNICODE defined
+               some types have been changed
+                 char               -> tt_char
+                 std::string        -> tt_string
+                 std::ostringstream -> tt_ostringstream
+                 std::ostream       -> tt_ostream
+                 std::istream       -> tt_istream
+               normally tt_char is equal char but when you are using wide characters then tt_char will be wchar_t (and so on)
+               (all typedef's are in ttmathtypes.h)
+    * added:   Big::IsInteger()
+               returns true if the value is integer (without fraction)
+               (NaN flag is not checked)
+    * added:   global Gamma() function
+    * added:   gamma() function to the parser
+    * added:   CGamma<ValueType> class
+               is used with Gamma() and Factorial() in multithreaded environment
+    * changed: Factorial() is using the Gamma() function now
+    * removed: Parser<>::SetFactorialMax() method
+               the factorial() is such a fast now that we don't need the method longer
+    * removed: ErrorCode::err_too_big_factorial
+
+
+Version 0.8.5 (2009.06.16):
+    * fixed:   Big::Mod(x) didn't correctly return a carry
+               and the result was sometimes very big (even greater than x)
+    * fixed:   global function Mod(x) didn't set an ErrorCode object
+    * fixed:   global function Round() didn't test a carry
+               now it sets ErrorCode object
+    * changed: function Sin(x) to Sin(x, ErrorCode * err=0)
+               when x was very big the function returns zero
+               now it sets ErrorCode object to err_overflow
+               and the result has a NaN flag set
+               the same is to Cos() function
+    * changed: PrepareSin(x) is using Big::Mod() now when reducing 2PI period
+               should be a little accurate especially on a very big 'x'
+    * changed: uint Mul(const UInt<value_size> & ss2, uint algorithm = 100)
+               void MulBig(const UInt<value_size> & ss2, UInt<value_size*2> & result, uint algorithm = 100)
+               those methods by default use MulFastest() and MulFastestBig()
+    * changed: changed a little Mul2Big() to cooperate with Mul3Big()
+    * added:   uint UInt::Mul3(const UInt<value_size> & ss2)
+               void UInt::Mul3Big(const UInt<value_size> & ss2, UInt<value_size*2> & result)
+               a new multiplication algorithm: Karatsuba multiplication,
+               on a vector UInt<100> with all items different from zero this algorithm is faster
+               about 3 times than Mul2Big(), and on a vector UInt<1000> with all items different from
+               zero this algorithm is faster more than 5 times than Mul2Big()
+               (measured on 32bit platform with GCC 4.3.3 with -O3 and -DTTMATH_RELEASE)
+    * added:   uint MulFastest(const UInt<value_size> & ss2)
+               void MulFastestBig(const UInt<value_size> & ss2, UInt<value_size*2> & result)
+               those methods are trying to select the fastest multiplication algorithm
+    * added:   uint AddVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result)
+               uint SubVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result)
+               three forms: asm x86, asm x86_64, no-asm
+               those methods are used by the Karatsuba multiplication algorithm
+    * added:   to Big<> class: support for NaN flag (Not a Number)
+               bool Big::IsNan() - returns true if the NaN flag is set
+               void Big::SetNan() - sets the NaN flag
+               The NaN flag is set by default after creating an object:
+                 Big<1, 2> a;    // NaN is set (it means the object has not a valid number)
+                 std::cout << a; // cout gives "NaN"
+                 a = 123;        // now NaN is not set
+                 std::cout << a; // cout gives "123"
+               The NaN is set if there was a carry during calculations
+                 a.Mul(very_big_value); // a will have a NaN set
+               The NaN is set if an argument is NaN too
+                 b.SetNan();
+                 a.Add(b);  // a will have NaN because b has NaN too
+               If you try to do something on a NaN object, the result is a NaN too
+                 a.SetNan();
+                 a.Add(2);  // a is still a NaN 
+               The NaN is set if you use incorrect arguments
+                 a.Ln(-10); // a will have the NaN flag
+               The only way to clear the NaN flag is to assign a correct value or other correct object,
+               supposing 'a' has NaN flag, to remove the flag you can either:
+                 a = 10;
+                 a.FromInt(30);
+                 a.SetOne(); 
+                 a.FromBig(other_object_without_nan);
+                 etc.
+               
+               
 Version 0.8.4 (2009.05.08):
     * fixed:   UInt::DivInt() didn't check whether the divisor is zero
                there was a hardware interruption when the divisor was zero

constgen/Makefile

@@ -1,6 +1,6 @@
 o      = main.o
 CC     = g++
-CFLAGS = -s -O2 -DCONSTANTSGENERATOR
+CFLAGS = -s -O2 -DTTMATH_CONSTANTSGENERATOR
 name   = gen
 
 

constgen/main.cpp

@@ -91,7 +91,7 @@ void CalcE()
 	ttmath::Big<1,400> e;
 	ttmath::uint steps;
 
-	// macro CONSTANTSGENERATOR has to be defined	
+	// macro TTMATH_CONSTANTSGENERATOR has to be defined	
 	e.ExpSurrounding0(1, &steps);
 	std::cout << "---------------- e ----------------" << std::endl;
 	e.mantissa.PrintTable(std::cout);
@@ -105,7 +105,7 @@ void CalcLn(int x)
 	ttmath::Big<1,400> ln;
 	ttmath::uint steps;
 
-	// macro CONSTANTSGENERATOR has to be defined	
+	// macro TTMATH_CONSTANTSGENERATOR has to be defined	
 	ln.LnSurrounding1(x, &steps);
 	std::cout << "---------------- ln(" << x << ") ----------------" << std::endl;
 	ln.mantissa.PrintTable(std::cout);

ttmath/ttmathconfig.h

@@ -0,0 +1,110 @@
+/*
+ * This file is a part of TTMath Bignum Library
+ * and is distributed under the PNG licence.
+ * Author: Christian Kaiser <chk@online.de>
+ */
+
+/* 
+	Copyright (c) 2009 Christian Kaiser
+
+	This software is provided 'as-is', without any express or implied
+	warranty. In no event will the authors be held liable for any damages
+	arising from the use of this software.
+
+	Permission is granted to anyone to use this software for any purpose,
+	including commercial applications, and to alter it and redistribute it
+	freely, subject to the following restrictions:
+
+		1. The origin of this software must not be misrepresented; you must not
+		claim that you wrote the original software. If you use this software
+		in a product, an acknowledgment in the product documentation would be
+		appreciated but is not required.
+
+		2. Altered source versions must be plainly marked as such, and must not be
+		misrepresented as being the original software.
+
+		3. This notice may not be removed or altered from any source
+		distribution.
+ */
+
+#ifndef headerfilettmathconfig
+#define headerfilettmathconfig
+#pragma once
+
+#include <sstream>
+
+namespace ttmath
+{
+#if defined(WIN32)
+	#include <windows.h>
+	
+	#if defined(_MT)
+		class 	clsCrit
+			{
+			private:
+				mutable CRITICAL_SECTION 		_Crit;
+
+												clsCrit(const clsCrit&) // inhibit copy (easy mistake to do; use clsCritObj instead!!!)
+													{
+													}
+				clsCrit&						operator=(const clsCrit& rhs); // inhibit assignment
+			public:
+												clsCrit(void)
+													{
+													InitializeCriticalSection(&_Crit);
+													}
+				virtual							~clsCrit(void)
+													{
+													DeleteCriticalSection(&_Crit);
+													}
+
+				void							Enter(void) const
+													{
+													EnterCriticalSection(&_Crit);
+													}
+				void							Leave(void) const
+													{
+													LeaveCriticalSection(&_Crit);
+													}
+			};
+
+		class 	clsCritObj
+			{
+			private:
+				const clsCrit&					_Crit;
+				
+				clsCritObj&						operator=(const clsCritObj& rhs); // not applicable
+			public:
+												clsCritObj(const clsCrit& Sync)
+													: _Crit(Sync)
+													{
+													_Crit.Enter();
+													}
+												~clsCritObj(void)
+													{
+													_Crit.Leave();
+													}
+			};
+		#define TTMATH_IMPLEMENT_THREADSAFE_OBJ					\
+			private:											\
+				clsCrit CritSect;								\
+			public:												\
+				operator clsCrit&()								\
+				{												\
+					return(CritSect);							\
+				}
+		#define TTMATH_USE_THREADSAFE_OBJ(c)	clsCritObj	lock(c)
+	#endif
+#else // defined(WIN32) 
+	// not Windows world: no threading synchronization for now
+#endif
+
+#if !defined(TTMATH_IMPLEMENT_THREADSAFE_OBJ)
+	// if we don't know about serialization, make it a no-op
+	#define TTMATH_IMPLEMENT_THREADSAFE_OBJ		/* */
+	#define TTMATH_USE_THREADSAFE_OBJ(c)		/* */
+#endif
+
+} // namespace
+
+#endif // headerfilettmathconfig

ttmath/ttmathint.h

@@ -1,7 +1,7 @@
 /*
  * This file is a part of TTMath Bignum Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /* 
@@ -47,6 +47,10 @@
 
 #include "ttmathuint.h"
 
+#if defined(_MSC_VER)
+	#pragma warning(disable:4127) // conditional expression is constant
+#endif
+
 namespace ttmath
 {
 
@@ -641,8 +645,14 @@ public:
 
 		// there can be a carry here when the size of this value is equal one word
 		// and the 'value' has the highest bit set
+		#if defined(_MSC_VER)
+			#pragma warning(disable:4127) // conditional expression is constant
+		#endif
 		if( value_size==1 && (value & TTMATH_UINT_HIGHEST_BIT)!=0 )
 			return 1;
+		#if defined(_MSC_VER)
+			#pragma warning(default:4127) // conditional expression is constant
+		#endif
 
 	return 0;
 	}
@@ -821,7 +831,7 @@ public:
 	/*!
 		a constructor for converting string to this class (with the base=10)
 	*/
-	Int(const char * s)
+	Int(const tt_char * s)
 	{
 		FromString(s);
 	}
@@ -830,7 +840,7 @@ public:
 	/*!
 		a constructor for converting a string to this class (with the base=10)
 	*/
-	Int(const std::string & s)
+	Int(const tt_string & s)
 	{
 		FromString( s.c_str() );
 	}
@@ -869,7 +879,7 @@ public:
 	/*!	
 		this method converts the value to a string with a base equal 'b'
 	*/
-	void ToString(std::string & result, uint b = 10) const
+	void ToString(tt_string & result, uint b = 10) const
 	{
 		if( IsSign() )
 		{
@@ -890,12 +900,14 @@ public:
 
 	/*!
 		this method converts a string into its value
+		string is given either as 'const char *' or 'const wchar_t *'
+
 		it returns carry=1 if the value will be too big or an incorrect base 'b' is given
 
 		string is ended with a non-digit value, for example:
 			"-12" will be translated to -12
 			as well as:
-			"- 12foo" will be translated to 12 too
+			"- 12foo" will be translated to -12 too
 
 		existing first white characters will be ommited
 		(between '-' and a first digit can be white characters too)
@@ -904,7 +916,7 @@ public:
 
 		value_read (if exists) tells whether something has actually been read (at least one digit)
 	*/
-	uint FromString(const char * s, uint b = 10, const char ** after_source = 0, bool * value_read = 0)
+	uint FromString(const tt_char * s, uint b = 10, const tt_char ** after_source = 0, bool * value_read = 0)
 	{
 	bool is_sign = false;
 	
@@ -961,16 +973,16 @@ public:
 		this method converts a string into its value
 		it returns carry=1 if the value will be too big or an incorrect base 'b' is given
 	*/
-	uint FromString(const std::string & s, uint b = 10)
+	uint FromString(const tt_string & s, uint b = 10)
 	{
-		return FromString( s.c_str() );
+		return FromString( s.c_str(), b );
 	}
 
 
 	/*!
 		this operator converts a string into its value (with base = 10)
 	*/
-	Int<value_size> & operator=(const char * s)
+	Int<value_size> & operator=(const tt_char * s)
 	{
 		FromString(s);
 
@@ -981,7 +993,7 @@ public:
 	/*!
 		this operator converts a string into its value (with base = 10)
 	*/
-	Int<value_size> & operator=(const std::string & s)
+	Int<value_size> & operator=(const tt_string & s)
 	{
 		FromString( s.c_str() );
 
@@ -1268,9 +1280,14 @@ public:
 	*
 	*/
 
-	friend std::ostream & operator<<(std::ostream & s, const Int<value_size> & l)
+	/*!
+		output for standard streams
+
+		tt_ostream is either std::ostream or std::wostream
+	*/
+	friend tt_ostream & operator<<(tt_ostream & s, const Int<value_size> & l)
 	{
-	std::string ss;
+	tt_string ss;
 
 		l.ToString(ss);
 		s << ss;
@@ -1279,13 +1296,17 @@ public:
 	}
 
 
+	/*!
+		input from standard streams
 
-	friend std::istream & operator>>(std::istream & s, Int<value_size> & l)
+		tt_istream is either std::istream or std::wistream
+	*/
+	friend tt_istream & operator>>(tt_istream & s, Int<value_size> & l)
 	{
-	std::string ss;
+	tt_string ss;
 	
-	// char for operator>>
-	unsigned char z;
+	// tt_char for operator>>
+	tt_char z;
 	
 		// operator>> omits white characters if they're set for ommiting
 		s >> z;
@@ -1300,7 +1321,7 @@ public:
 		while( s.good() && UInt<value_size>::CharToDigit(z, 10)>=0 )
 		{
 			ss += z;
-			z = s.get();
+			z = static_cast<tt_char>(s.get());
 		}
 
 		// we're leaving the last readed character
@@ -1316,5 +1337,9 @@ public:
 
 } // namespace
 
+#if defined(_MSC_VER)
+	#pragma warning(default:4127) // conditional expression is constant
+#endif
+
 
 #endif

ttmath/ttmathobjects.h

@@ -1,7 +1,7 @@
 /*
  * This file is a part of TTMath Mathematical Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /* 
@@ -47,6 +47,7 @@
 #include "ttmathtypes.h"
 
 #include <string>
+#include <vector>
 #include <list>
 #include <map>
 
@@ -73,18 +74,19 @@ public:
 	struct Item
 	{
 		// name of a variable of a function
-		std::string value;
+		// (either std::string or std::wstring)
+		tt_string value;
 
 		// number of parameters required by the function
 		// (if there's a variable this 'param' is ignored)
 		int param;
 
 		Item() {}
-		Item(const std::string & v, int p) : value(v), param(p) {}
+		Item(const tt_string & v, int p) : value(v), param(p) {}
 	};
 
 	// 'Table' is the type of our table
-	typedef std::map<std::string, Item> Table;
+	typedef std::map<tt_string, Item> Table;
 	typedef	Table::iterator Iterator;
 	typedef	Table::const_iterator CIterator;
 
@@ -112,7 +114,7 @@ public:
 	/*!
 		this method returns true if the name can be as a name of an object
 	*/
-	static bool IsNameCorrect(const std::string & name)
+	static bool IsNameCorrect(const tt_string & name)
 	{
 		if( name.empty() )
 			return false;
@@ -120,7 +122,7 @@ public:
 		if( !CorrectCharacter(name[0], false) )
 			return false;
 
-		std::string::const_iterator i=name.begin();
+		tt_string::const_iterator i=name.begin();
 
 		for(++i ; i!=name.end() ; ++i)
 			if( !CorrectCharacter(*i, true) )
@@ -133,7 +135,7 @@ public:
 	/*!
 		this method returns true if such an object is defined (name exists)
 	*/
-	bool IsDefined(const std::string & name)
+	bool IsDefined(const tt_string & name)
 	{
 		Iterator i = table.find(name);
 
@@ -148,7 +150,7 @@ public:
 	/*!
 		this method adds one object (variable of function) into the table
 	*/
-	ErrorCode Add(const std::string & name, const std::string & value, int param = 0)
+	ErrorCode Add(const tt_string & name, const tt_string & value, int param = 0)
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -205,7 +207,7 @@ public:
 	/*!
 		this method changes the value and the number of parameters for a specific object
 	*/
-	ErrorCode EditValue(const std::string & name, const std::string & value, int param = 0)
+	ErrorCode EditValue(const tt_string & name, const tt_string & value, int param = 0)
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -225,7 +227,7 @@ public:
 	/*!
 		this method changes the name of a specific object
 	*/
-	ErrorCode EditName(const std::string & old_name, const std::string & new_name)
+	ErrorCode EditName(const tt_string & old_name, const tt_string & new_name)
 	{
 		if( !IsNameCorrect(old_name) || !IsNameCorrect(new_name) )
 			return err_incorrect_name;
@@ -256,7 +258,7 @@ public:
 	/*!
 		this method deletes an object
 	*/
-	ErrorCode Delete(const std::string & name)
+	ErrorCode Delete(const tt_string & name)
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -275,7 +277,7 @@ public:
 	/*!
 		this method gets the value of a specific object
 	*/
-	ErrorCode GetValue(const std::string & name, std::string & value) const
+	ErrorCode GetValue(const tt_string & name, tt_string & value) const
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -298,7 +300,7 @@ public:
 		this method gets the value of a specific object
 		(this version is used for not copying the whole string)
 	*/
-	ErrorCode GetValue(const std::string & name, const char ** value) const
+	ErrorCode GetValue(const tt_string & name, const tt_char ** value) const
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -321,7 +323,7 @@ public:
 		this method gets the value and the number of parameters
 		of a specific object
 	*/
-	ErrorCode GetValueAndParam(const std::string & name, std::string & value, int * param) const
+	ErrorCode GetValueAndParam(const tt_string & name, tt_string & value, int * param) const
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -347,7 +349,7 @@ public:
 		of a specific object
 		(this version is used for not copying the whole string)
 	*/
-	ErrorCode GetValueAndParam(const std::string & name, const char ** value, int * param) const
+	ErrorCode GetValueAndParam(const tt_string & name, const tt_char ** value, int * param) const
 	{
 		if( !IsNameCorrect(name) )
 			return err_incorrect_name;
@@ -430,7 +432,7 @@ public:
 	*/
 	History()
 	{
-		buffer_max_size = 10;
+		buffer_max_size = 15;
 	}
 
 
@@ -487,10 +489,118 @@ public:
 	return false;
 	}
 
+
+	/*!
+		this methods deletes an item
+
+		we assume that there is only one item with the 'key'
+		(this methods removes the first one)
+	*/
+	bool Remove(const ValueType & key)
+	{
+		typename buffer_type::iterator i = buffer.begin();
+
+		for( ; i != buffer.end() ; ++i )
+		{
+			if( i->key == key )
+			{
+				buffer.erase(i);
+				return true;
+			}
+		}
+
+	return false;
+	}
+
+
 }; // end of class History
 
 
 
+/*!
+	this is an auxiliary class used when calculating Gamma() or Factorial()
+
+	in multithreaded environment you can provide an object of this class to
+	the Gamma() or Factorial() function, e.g;
+		typedef Big<1, 3> MyBig;
+		MyBig x = 123456;
+		CGamma<MyBig> cgamma;
+		std::cout << Gamma(x, cgamma);
+	each thread should have its own CGamma<> object
+
+	in a single-thread environment a CGamma<> object is a static variable
+	in a second version of Gamma() and you don't have to explicitly use it, e.g.
+		typedef Big<1, 3> MyBig;
+		MyBig x = 123456;
+		std::cout << Gamma(x);
+*/
+template<class ValueType>
+struct CGamma
+{
+	/*!
+		this table holds factorials
+			1
+			1
+			2
+			6
+			24
+			120
+			720
+			.......
+	*/
+	std::vector<ValueType> fact;
+
+
+	/*!
+		this table holds Bernoulli numbers
+			1
+			-0.5
+			0.166666666666666666666666667
+			0
+			-0.0333333333333333333333333333
+			0
+			0.0238095238095238095238095238
+			0
+			-0.0333333333333333333333333333
+			0
+			0.075757575757575757575757576
+			.....
+	*/
+	std::vector<ValueType> bern;
+
+
+	/*!
+		here we store some calculated values
+		(this is for speeding up, if the next argument of Gamma() or Factorial()
+		is in the 'history' then the result we are not calculating but simply
+		return from the 'history' object)
+	*/
+	History<ValueType> history;
+
+
+	/*!
+		this method prepares some coefficients: factorials and Bernoulli numbers
+		stored in 'fact' and 'bern' objects
+		
+		how many values should be depends on the size of the mantissa - if
+		the mantissa is larger then we must calculate more values
+		    for a mantissa which consists of 256 bits (8 words on a 32bit platform)
+			we have to calculate about 30 values (the size of fact and bern will be 30),
+			and for a 2048 bits mantissa we have to calculate 306 coefficients
+
+		you don't have to call this method, these coefficients will be automatically calculated
+		when they are needed
+
+		you must note that calculating of the coefficients is a little time-consuming operation,
+		(especially when the mantissa is large) and first called to Gamma() or Factorial()
+		can take more time than next calls, and in the end this is the point when InitAll()
+		comes in handy: you can call this method somewhere at the beginning of your program
+	*/
+	void InitAll();
+	// definition is in ttmath.h
+};
+
+
 
 
 } // namespace

ttmath/ttmathparser.h

@@ -1,7 +1,7 @@
 /*
- * This file is a part of TTMath Mathematical Library
+ * This file is a part of TTMath Bignum Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /* 
@@ -45,7 +45,7 @@
     \brief A mathematical parser
 */
 
-#include <fstream>
+#include <cstdio>
 #include <vector>
 #include <map>
 #include <set>
@@ -137,7 +137,6 @@ namespace ttmath
 template<class ValueType>
 class Parser
 {
-
 private:
 
 /*!
@@ -256,7 +255,7 @@ public:
 		bool function;
 
 		// if function is true
-		std::string function_name;
+		tt_string function_name;
 
 		/*
 			the sign of value
@@ -311,10 +310,11 @@ ErrorCode error;
 
 /*!
 	pointer to the currently reading char
+	it's either char* or wchar_t*
 
 	when an error has occured it may be used to count the index of the wrong character
 */
-const char * pstring;
+const tt_char * pstring;
 
 
 /*!
@@ -351,7 +351,7 @@ const Objects * puser_variables;
 const Objects * puser_functions;
 
 
-typedef std::map<std::string, ValueType> FunctionLocalVariables;
+typedef std::map<tt_string, ValueType> FunctionLocalVariables;
 
 /*!
 	a pointer to the local variables of a function
@@ -362,13 +362,13 @@ const FunctionLocalVariables * pfunction_local_variables;
 /*!
 	a temporary set using during parsing user defined variables
 */
-std::set<std::string> visited_variables;
+std::set<tt_string> visited_variables;
 
 
 /*!
 	a temporary set using during parsing user defined functions
 */
-std::set<std::string> visited_functions;
+std::set<tt_string> visited_functions;
 
 
 
@@ -396,10 +396,10 @@ typedef void (ValueType::*pfunction_var)();
 	table of mathematic functions
 
 	this map consists of:
-		std::string - function's name
+		tt_string - function's name
 		pfunction - pointer to specific function
 */
-typedef std::map<std::string, pfunction> FunctionsTable;
+typedef std::map<tt_string, pfunction> FunctionsTable;
 FunctionsTable functions_table;
 
 
@@ -407,10 +407,10 @@ FunctionsTable functions_table;
 	table of mathematic operators
 
 	this map consists of:
-		std::string - operators's name
+		tt_string - operators's name
 		MatOperator::Type - type of the operator
 */
-typedef std::map<std::string, typename MatOperator::Type> OperatorsTable;
+typedef std::map<tt_string, typename MatOperator::Type> OperatorsTable;
 OperatorsTable operators_table;
 
 
@@ -418,18 +418,17 @@ OperatorsTable operators_table;
 	table of mathematic variables
 
 	this map consists of:
-		std::string   - variable's name
+		tt_string     - variable's name
 		pfunction_var - pointer to specific function which returns value of variable
 */
-typedef std::map<std::string, pfunction_var> VariablesTable;
+typedef std::map<tt_string, pfunction_var> VariablesTable;
 VariablesTable variables_table;
 
 
 /*!
-	you can't calculate the factorial if the argument is greater than 'factorial_max'
-	default value is zero which means there are not any limitations
+	some coefficients used when calculating the gamma (or factorial) function
 */
-ValueType factorial_max;
+CGamma<ValueType> cgamma;
 
 
 /*!
@@ -456,7 +455,7 @@ void SkipWhiteCharacters()
 /*!
 	an auxiliary method for RecurrenceParsingVariablesOrFunction(...)
 */
-void RecurrenceParsingVariablesOrFunction_CheckStopCondition(bool variable, const std::string & name)
+void RecurrenceParsingVariablesOrFunction_CheckStopCondition(bool variable, const tt_string & name)
 {
 	if( variable )
 	{
@@ -474,7 +473,7 @@ void RecurrenceParsingVariablesOrFunction_CheckStopCondition(bool variable, cons
 /*!
 	an auxiliary method for RecurrenceParsingVariablesOrFunction(...)
 */
-void RecurrenceParsingVariablesOrFunction_AddName(bool variable, const std::string & name)
+void RecurrenceParsingVariablesOrFunction_AddName(bool variable, const tt_string & name)
 {
 	if( variable )
 		visited_variables.insert( name );
@@ -486,7 +485,7 @@ void RecurrenceParsingVariablesOrFunction_AddName(bool variable, const std::stri
 /*!
 	an auxiliary method for RecurrenceParsingVariablesOrFunction(...)
 */
-void RecurrenceParsingVariablesOrFunction_DeleteName(bool variable, const std::string & name)
+void RecurrenceParsingVariablesOrFunction_DeleteName(bool variable, const tt_string & name)
 {
 	if( variable )
 		visited_variables.erase( name );
@@ -505,7 +504,8 @@ void RecurrenceParsingVariablesOrFunction_DeleteName(bool variable, const std::s
 	(there can be a recurrence here therefore we're using 'visited_variables'
 	and 'visited_functions' sets to make a stop condition)
 */
-ValueType RecurrenceParsingVariablesOrFunction(bool variable, const std::string & name, const char * new_string, FunctionLocalVariables * local_variables = 0)
+ValueType RecurrenceParsingVariablesOrFunction(bool variable, const tt_string & name, const tt_char * new_string,
+											   FunctionLocalVariables * local_variables = 0)
 {
 	RecurrenceParsingVariablesOrFunction_CheckStopCondition(variable, name);
 	RecurrenceParsingVariablesOrFunction_AddName(variable, name);
@@ -548,12 +548,12 @@ public:
 /*!
 	this method returns the user-defined value of a variable
 */
-bool GetValueOfUserDefinedVariable(const std::string & variable_name,ValueType & result)
+bool GetValueOfUserDefinedVariable(const tt_string & variable_name,ValueType & result)
 {
 	if( !puser_variables )
 		return false;
 
-	const char * string_value;
+	const tt_char * string_value;
 
 	if( puser_variables->GetValue(variable_name, &string_value) != err_ok )
 		return false;
@@ -567,7 +567,7 @@ return true;
 /*!
 	this method returns the value of a local variable of a function
 */
-bool GetValueOfFunctionLocalVariable(const std::string & variable_name, ValueType & result)
+bool GetValueOfFunctionLocalVariable(const tt_string & variable_name, ValueType & result)
 {
 	if( !pfunction_local_variables )
 		return false;
@@ -589,7 +589,7 @@ return true;
 	we make an object of type ValueType then call a method which 
 	sets the correct value in it and finally we'll return the object
 */
-ValueType GetValueOfVariable(const std::string & variable_name)
+ValueType GetValueOfVariable(const tt_string & variable_name)
 {
 ValueType result;
 
@@ -600,7 +600,7 @@ ValueType result;
 		return result;
 
 
-	typename std::map<std::string, pfunction_var>::iterator i =
+	typename std::map<tt_string, pfunction_var>::iterator i =
 													variables_table.find(variable_name);
 
 	if( i == variables_table.end() )
@@ -674,6 +674,20 @@ return result;
 }
 
 
+void Gamma(int sindex, int amount_of_args, ValueType & result)
+{
+	if( amount_of_args != 1 )
+		Error( err_improper_amount_of_arguments );
+
+	ErrorCode err;
+	
+	result = ttmath::Gamma(stack[sindex].value, cgamma, &err, pstop_calculating);
+
+	if(err != err_ok)
+		Error( err );
+}
+
+
 /*!
 	factorial
 	result = 1 * 2 * 3 * 4 * .... * x
@@ -685,10 +699,7 @@ void Factorial(int sindex, int amount_of_args, ValueType & result)
 
 	ErrorCode err;
 
-	if( !factorial_max.IsZero() && stack[sindex].value > factorial_max )
-		Error( err_too_big_factorial );
-
-	result = ttmath::Factorial(stack[sindex].value, &err, pstop_calculating);
+	result = ttmath::Factorial(stack[sindex].value, cgamma, &err, pstop_calculating);
 
 	if(err != err_ok)
 		Error( err );
@@ -708,7 +719,11 @@ void Sin(int sindex, int amount_of_args, ValueType & result)
 	if( amount_of_args != 1 )
 		Error( err_improper_amount_of_arguments );
 
-	result = ttmath::Sin( ConvertAngleToRad(stack[sindex].value) );
+	ErrorCode err;
+	result = ttmath::Sin( ConvertAngleToRad(stack[sindex].value), &err );
+
+	if(err != err_ok)
+		Error( err );
 }
 
 void Cos(int sindex, int amount_of_args, ValueType & result)
@@ -716,7 +731,11 @@ void Cos(int sindex, int amount_of_args, ValueType & result)
 	if( amount_of_args != 1 )
 		Error( err_improper_amount_of_arguments );
 
-	result = ttmath::Cos( ConvertAngleToRad(stack[sindex].value) );
+	ErrorCode err;
+	result = ttmath::Cos( ConvertAngleToRad(stack[sindex].value), &err );
+
+	if(err != err_ok)
+		Error( err );
 }
 
 void Tan(int sindex, int amount_of_args, ValueType & result)
@@ -757,7 +776,10 @@ void Round(int sindex, int amount_of_args, ValueType & result)
 	if( amount_of_args != 1 )
 		Error( err_improper_amount_of_arguments );
 
-	result = ttmath::Round(stack[sindex].value);
+	result = stack[sindex].value;
+
+	if( result.Round() )
+		Error( err_overflow );
 }
 
 
@@ -973,7 +995,7 @@ void Not(int sindex, int amount_of_args, ValueType & result)
 
 void DegToRad(int sindex, int amount_of_args, ValueType & result)
 {
-	ErrorCode err;
+	ErrorCode err = err_ok;
 
 	if( amount_of_args == 1 )
 	{
@@ -1052,7 +1074,7 @@ void RadToGrad(int sindex, int amount_of_args, ValueType & result)
 
 void DegToGrad(int sindex, int amount_of_args, ValueType & result)
 {
-	ErrorCode err;
+	ErrorCode err = err_ok;
 
 	if( amount_of_args == 1 )
 	{
@@ -1333,17 +1355,34 @@ void Avg(int sindex, int amount_of_args, ValueType & result)
 }	
 
 
+
+/*!
+	we use such a method because 'wvsprintf' is not everywhere defined
+*/
+void Sprintf(tt_char * buffer, int par)
+{
+char buf[30]; // char, not tt_char
+int i;
+
+	sprintf(buf, "%d", par);
+	for(i=0 ; buf[i] != 0 ; ++i)
+		buffer[i] = buf[i];
+
+	buffer[i] = 0;
+}
+
+
 /*!
 	this method returns the value from a user-defined function
 
 	(look at the description in 'CallFunction(...)')
 */
-bool GetValueOfUserDefinedFunction(const std::string & function_name, int amount_of_args, int sindex)
+bool GetValueOfUserDefinedFunction(const tt_string & function_name, int amount_of_args, int sindex)
 {
 	if( !puser_functions )
 		return false;
 
-	const char * string_value;
+	const tt_char * string_value;
 	int param;
 
 	if( puser_functions->GetValueAndParam(function_name, &string_value, &param) != err_ok )
@@ -1357,15 +1396,17 @@ bool GetValueOfUserDefinedFunction(const std::string & function_name, int amount
 
 	if( amount_of_args > 0 )
 	{
-		char buffer[20];
+		tt_char buffer[30];
 
 		// x = x1
-		sprintf(buffer,"x");
+		buffer[0] = 'x';
+		buffer[1] = 0;
 		local_variables.insert( std::make_pair(buffer, stack[sindex].value) );
 
 		for(int i=0 ; i<amount_of_args ; ++i)
 		{
-			sprintf(buffer,"x%d",i+1);
+			buffer[0] = 'x';
+			Sprintf(buffer+1, i+1);
 			local_variables.insert( std::make_pair(buffer, stack[sindex + i*2].value) );
 		}
 	}
@@ -1389,7 +1430,7 @@ return true;
 	result will be stored in 'stack[sindex-1].value'
 	(we don't have to set the correct type of this element, it'll be set later)
 */
-void CallFunction(const std::string & function_name, int amount_of_args, int sindex)
+void CallFunction(const tt_string & function_name, int amount_of_args, int sindex)
 {
 	if( GetValueOfUserDefinedFunction(function_name, amount_of_args, sindex) )
 		return;
@@ -1415,9 +1456,9 @@ void CallFunction(const std::string & function_name, int amount_of_args, int sin
 	function_name - name of the function
 	pf - pointer to the function (to the wrapper)
 */
-void InsertFunctionToTable(const char * function_name, pfunction pf)
+void InsertFunctionToTable(const tt_char * function_name, pfunction pf)
 {
-	functions_table.insert( std::make_pair(std::string(function_name), pf));
+	functions_table.insert( std::make_pair(tt_string(function_name), pf));
 }
 
 
@@ -1428,9 +1469,9 @@ void InsertFunctionToTable(const char * function_name, pfunction pf)
 	variable_name - name of the function
 	pf - pointer to the function
 */
-void InsertVariableToTable(const char * variable_name, pfunction_var pf)
+void InsertVariableToTable(const tt_char * variable_name, pfunction_var pf)
 {
-	variables_table.insert( std::make_pair(std::string(variable_name), pf));
+	variables_table.insert( std::make_pair(tt_string(variable_name), pf));
 }
 
 
@@ -1439,67 +1480,65 @@ void InsertVariableToTable(const char * variable_name, pfunction_var pf)
 */
 void CreateFunctionsTable()
 {
-	/*
-		names of functions should consist of small letters
-	*/
-	InsertFunctionToTable("factorial",	&Parser<ValueType>::Factorial);
-	InsertFunctionToTable("abs",   		&Parser<ValueType>::Abs);
-	InsertFunctionToTable("sin",   		&Parser<ValueType>::Sin);
-	InsertFunctionToTable("cos",   		&Parser<ValueType>::Cos);
-	InsertFunctionToTable("tan",   		&Parser<ValueType>::Tan);
-	InsertFunctionToTable("tg",		  	&Parser<ValueType>::Tan);
-	InsertFunctionToTable("cot",  		&Parser<ValueType>::Cot);
-	InsertFunctionToTable("ctg",  		&Parser<ValueType>::Cot);
-	InsertFunctionToTable("int",	   	&Parser<ValueType>::Int);
-	InsertFunctionToTable("round",	 	&Parser<ValueType>::Round);
-	InsertFunctionToTable("ln",	    	&Parser<ValueType>::Ln);
-	InsertFunctionToTable("log",	   	&Parser<ValueType>::Log);
-	InsertFunctionToTable("exp",	   	&Parser<ValueType>::Exp);
-	InsertFunctionToTable("max",	   	&Parser<ValueType>::Max);
-	InsertFunctionToTable("min",	   	&Parser<ValueType>::Min);
-	InsertFunctionToTable("asin",   	&Parser<ValueType>::ASin);
-	InsertFunctionToTable("acos",   	&Parser<ValueType>::ACos);
-	InsertFunctionToTable("atan",   	&Parser<ValueType>::ATan);
-	InsertFunctionToTable("atg",	   	&Parser<ValueType>::ATan);
-	InsertFunctionToTable("acot",   	&Parser<ValueType>::ACot);
-	InsertFunctionToTable("actg",   	&Parser<ValueType>::ACot);
-	InsertFunctionToTable("sgn",   		&Parser<ValueType>::Sgn);
-	InsertFunctionToTable("mod",   		&Parser<ValueType>::Mod);
-	InsertFunctionToTable("if",   		&Parser<ValueType>::If);
-	InsertFunctionToTable("or",   		&Parser<ValueType>::Or);
-	InsertFunctionToTable("and",  		&Parser<ValueType>::And);
-	InsertFunctionToTable("not",  		&Parser<ValueType>::Not);
-	InsertFunctionToTable("degtorad",	&Parser<ValueType>::DegToRad);
-	InsertFunctionToTable("radtodeg",	&Parser<ValueType>::RadToDeg);
-	InsertFunctionToTable("degtodeg",	&Parser<ValueType>::DegToDeg);
-	InsertFunctionToTable("gradtorad",	&Parser<ValueType>::GradToRad);
-	InsertFunctionToTable("radtograd",	&Parser<ValueType>::RadToGrad);
-	InsertFunctionToTable("degtograd",	&Parser<ValueType>::DegToGrad);
-	InsertFunctionToTable("gradtodeg",	&Parser<ValueType>::GradToDeg);
-	InsertFunctionToTable("ceil",		&Parser<ValueType>::Ceil);
-	InsertFunctionToTable("floor",		&Parser<ValueType>::Floor);
-	InsertFunctionToTable("sqrt",		&Parser<ValueType>::Sqrt);
-	InsertFunctionToTable("sinh",		&Parser<ValueType>::Sinh);
-	InsertFunctionToTable("cosh",		&Parser<ValueType>::Cosh);
-	InsertFunctionToTable("tanh",		&Parser<ValueType>::Tanh);
-	InsertFunctionToTable("tgh",		&Parser<ValueType>::Tanh);
-	InsertFunctionToTable("coth",		&Parser<ValueType>::Coth);
-	InsertFunctionToTable("ctgh",		&Parser<ValueType>::Coth);
-	InsertFunctionToTable("root",		&Parser<ValueType>::Root);
-	InsertFunctionToTable("asinh",		&Parser<ValueType>::ASinh);
-	InsertFunctionToTable("acosh",		&Parser<ValueType>::ACosh);
-	InsertFunctionToTable("atanh",		&Parser<ValueType>::ATanh);
-	InsertFunctionToTable("atgh",		&Parser<ValueType>::ATanh);
-	InsertFunctionToTable("acoth",		&Parser<ValueType>::ACoth);
-	InsertFunctionToTable("actgh",		&Parser<ValueType>::ACoth);
-	InsertFunctionToTable("bitand",		&Parser<ValueType>::BitAnd);
-	InsertFunctionToTable("bitor",		&Parser<ValueType>::BitOr);
-	InsertFunctionToTable("bitxor",		&Parser<ValueType>::BitXor);
-	InsertFunctionToTable("band",		&Parser<ValueType>::BitAnd);
-	InsertFunctionToTable("bor",		&Parser<ValueType>::BitOr);
-	InsertFunctionToTable("bxor",		&Parser<ValueType>::BitXor);
-	InsertFunctionToTable("sum",		&Parser<ValueType>::Sum);
-	InsertFunctionToTable("avg",		&Parser<ValueType>::Avg);
+	InsertFunctionToTable(TTMATH_TEXT("gamma"),		&Parser<ValueType>::Gamma);
+	InsertFunctionToTable(TTMATH_TEXT("factorial"),	&Parser<ValueType>::Factorial);
+	InsertFunctionToTable(TTMATH_TEXT("abs"),   	&Parser<ValueType>::Abs);
+	InsertFunctionToTable(TTMATH_TEXT("sin"),   	&Parser<ValueType>::Sin);
+	InsertFunctionToTable(TTMATH_TEXT("cos"),   	&Parser<ValueType>::Cos);
+	InsertFunctionToTable(TTMATH_TEXT("tan"),   	&Parser<ValueType>::Tan);
+	InsertFunctionToTable(TTMATH_TEXT("tg"),		&Parser<ValueType>::Tan);
+	InsertFunctionToTable(TTMATH_TEXT("cot"),  		&Parser<ValueType>::Cot);
+	InsertFunctionToTable(TTMATH_TEXT("ctg"),  		&Parser<ValueType>::Cot);
+	InsertFunctionToTable(TTMATH_TEXT("int"),	   	&Parser<ValueType>::Int);
+	InsertFunctionToTable(TTMATH_TEXT("round"),	 	&Parser<ValueType>::Round);
+	InsertFunctionToTable(TTMATH_TEXT("ln"),	    &Parser<ValueType>::Ln);
+	InsertFunctionToTable(TTMATH_TEXT("log"),	   	&Parser<ValueType>::Log);
+	InsertFunctionToTable(TTMATH_TEXT("exp"),	   	&Parser<ValueType>::Exp);
+	InsertFunctionToTable(TTMATH_TEXT("max"),	   	&Parser<ValueType>::Max);
+	InsertFunctionToTable(TTMATH_TEXT("min"),	   	&Parser<ValueType>::Min);
+	InsertFunctionToTable(TTMATH_TEXT("asin"),   	&Parser<ValueType>::ASin);
+	InsertFunctionToTable(TTMATH_TEXT("acos"),   	&Parser<ValueType>::ACos);
+	InsertFunctionToTable(TTMATH_TEXT("atan"),   	&Parser<ValueType>::ATan);
+	InsertFunctionToTable(TTMATH_TEXT("atg"),	   	&Parser<ValueType>::ATan);
+	InsertFunctionToTable(TTMATH_TEXT("acot"),   	&Parser<ValueType>::ACot);
+	InsertFunctionToTable(TTMATH_TEXT("actg"),   	&Parser<ValueType>::ACot);
+	InsertFunctionToTable(TTMATH_TEXT("sgn"),   	&Parser<ValueType>::Sgn);
+	InsertFunctionToTable(TTMATH_TEXT("mod"),   	&Parser<ValueType>::Mod);
+	InsertFunctionToTable(TTMATH_TEXT("if"),   		&Parser<ValueType>::If);
+	InsertFunctionToTable(TTMATH_TEXT("or"),   		&Parser<ValueType>::Or);
+	InsertFunctionToTable(TTMATH_TEXT("and"),  		&Parser<ValueType>::And);
+	InsertFunctionToTable(TTMATH_TEXT("not"),  		&Parser<ValueType>::Not);
+	InsertFunctionToTable(TTMATH_TEXT("degtorad"),	&Parser<ValueType>::DegToRad);
+	InsertFunctionToTable(TTMATH_TEXT("radtodeg"),	&Parser<ValueType>::RadToDeg);
+	InsertFunctionToTable(TTMATH_TEXT("degtodeg"),	&Parser<ValueType>::DegToDeg);
+	InsertFunctionToTable(TTMATH_TEXT("gradtorad"),	&Parser<ValueType>::GradToRad);
+	InsertFunctionToTable(TTMATH_TEXT("radtograd"),	&Parser<ValueType>::RadToGrad);
+	InsertFunctionToTable(TTMATH_TEXT("degtograd"),	&Parser<ValueType>::DegToGrad);
+	InsertFunctionToTable(TTMATH_TEXT("gradtodeg"),	&Parser<ValueType>::GradToDeg);
+	InsertFunctionToTable(TTMATH_TEXT("ceil"),		&Parser<ValueType>::Ceil);
+	InsertFunctionToTable(TTMATH_TEXT("floor"),		&Parser<ValueType>::Floor);
+	InsertFunctionToTable(TTMATH_TEXT("sqrt"),		&Parser<ValueType>::Sqrt);
+	InsertFunctionToTable(TTMATH_TEXT("sinh"),		&Parser<ValueType>::Sinh);
+	InsertFunctionToTable(TTMATH_TEXT("cosh"),		&Parser<ValueType>::Cosh);
+	InsertFunctionToTable(TTMATH_TEXT("tanh"),		&Parser<ValueType>::Tanh);
+	InsertFunctionToTable(TTMATH_TEXT("tgh"),		&Parser<ValueType>::Tanh);
+	InsertFunctionToTable(TTMATH_TEXT("coth"),		&Parser<ValueType>::Coth);
+	InsertFunctionToTable(TTMATH_TEXT("ctgh"),		&Parser<ValueType>::Coth);
+	InsertFunctionToTable(TTMATH_TEXT("root"),		&Parser<ValueType>::Root);
+	InsertFunctionToTable(TTMATH_TEXT("asinh"),		&Parser<ValueType>::ASinh);
+	InsertFunctionToTable(TTMATH_TEXT("acosh"),		&Parser<ValueType>::ACosh);
+	InsertFunctionToTable(TTMATH_TEXT("atanh"),		&Parser<ValueType>::ATanh);
+	InsertFunctionToTable(TTMATH_TEXT("atgh"),		&Parser<ValueType>::ATanh);
+	InsertFunctionToTable(TTMATH_TEXT("acoth"),		&Parser<ValueType>::ACoth);
+	InsertFunctionToTable(TTMATH_TEXT("actgh"),		&Parser<ValueType>::ACoth);
+	InsertFunctionToTable(TTMATH_TEXT("bitand"),	&Parser<ValueType>::BitAnd);
+	InsertFunctionToTable(TTMATH_TEXT("bitor"),		&Parser<ValueType>::BitOr);
+	InsertFunctionToTable(TTMATH_TEXT("bitxor"),	&Parser<ValueType>::BitXor);
+	InsertFunctionToTable(TTMATH_TEXT("band"),		&Parser<ValueType>::BitAnd);
+	InsertFunctionToTable(TTMATH_TEXT("bor"),		&Parser<ValueType>::BitOr);
+	InsertFunctionToTable(TTMATH_TEXT("bxor"),		&Parser<ValueType>::BitXor);
+	InsertFunctionToTable(TTMATH_TEXT("sum"),		&Parser<ValueType>::Sum);
+	InsertFunctionToTable(TTMATH_TEXT("avg"),		&Parser<ValueType>::Avg);
 }
 
 
@@ -1508,11 +1547,8 @@ void CreateFunctionsTable()
 */
 void CreateVariablesTable()
 {
-	/*
-		names of variables should consist of small letters
-	*/
-	InsertVariableToTable("pi", &ValueType::SetPi);
-	InsertVariableToTable("e",  &ValueType::SetE);
+	InsertVariableToTable(TTMATH_TEXT("pi"), &ValueType::SetPi);
+	InsertVariableToTable(TTMATH_TEXT("e"),  &ValueType::SetE);
 }
 
 
@@ -1538,7 +1574,7 @@ return c;
 	what should be returned is tested just by a '(' character that means if there's
 	a '(' character after a name that function returns 'true'
 */
-bool ReadName(std::string & result)
+bool ReadName(tt_string & result)
 {
 int character;
 
@@ -1555,7 +1591,7 @@ int character;
 
 	do
 	{
-		result   += character;
+		result   += static_cast<tt_char>( character );
 		character = * ++pstring;
 	}
 	while(	(character>='a' && character<='z') ||
@@ -1610,7 +1646,7 @@ return false;
 */
 bool ReadVariableOrFunction(Item & result)
 {
-std::string name;
+tt_string name;
 bool is_it_name_of_function = ReadName(name);
 
 	if( is_it_name_of_function )
@@ -1639,7 +1675,7 @@ return is_it_name_of_function;
 */
 void ReadValue(Item & result, int reading_base)
 {
-const char * new_stack_pointer;
+const tt_char * new_stack_pointer;
 bool value_read;
 
 	int carry = result.value.FromString(pstring, reading_base, &new_stack_pointer, &value_read);
@@ -1812,9 +1848,9 @@ return 0;
 }
 
 
-void InsertOperatorToTable(const std::string & name, typename MatOperator::Type type)
+void InsertOperatorToTable(const tt_char * name, typename MatOperator::Type type)
 {
-	operators_table.insert( std::make_pair(name, type) );
+	operators_table.insert( std::make_pair(tt_string(name), type) );
 }
 
 
@@ -1823,19 +1859,19 @@ void InsertOperatorToTable(const std::string & name, typename MatOperator::Type
 */
 void CreateMathematicalOperatorsTable()
 {
-	InsertOperatorToTable(std::string("||"), MatOperator::lor);
-	InsertOperatorToTable(std::string("&&"), MatOperator::land);
-	InsertOperatorToTable(std::string("!="), MatOperator::neq);
-	InsertOperatorToTable(std::string("=="), MatOperator::eq);
-	InsertOperatorToTable(std::string(">="), MatOperator::get);
-	InsertOperatorToTable(std::string("<="), MatOperator::let);
-	InsertOperatorToTable(std::string(">"),  MatOperator::gt);
-	InsertOperatorToTable(std::string("<"),  MatOperator::lt);
-	InsertOperatorToTable(std::string("-"),  MatOperator::sub);
-	InsertOperatorToTable(std::string("+"),  MatOperator::add);
-	InsertOperatorToTable(std::string("/"),  MatOperator::div);
-	InsertOperatorToTable(std::string("*"),  MatOperator::mul);
-	InsertOperatorToTable(std::string("^"),  MatOperator::pow);
+	InsertOperatorToTable(TTMATH_TEXT("||"), MatOperator::lor);
+	InsertOperatorToTable(TTMATH_TEXT("&&"), MatOperator::land);
+	InsertOperatorToTable(TTMATH_TEXT("!="), MatOperator::neq);
+	InsertOperatorToTable(TTMATH_TEXT("=="), MatOperator::eq);
+	InsertOperatorToTable(TTMATH_TEXT(">="), MatOperator::get);
+	InsertOperatorToTable(TTMATH_TEXT("<="), MatOperator::let);
+	InsertOperatorToTable(TTMATH_TEXT(">"),  MatOperator::gt);
+	InsertOperatorToTable(TTMATH_TEXT("<"),  MatOperator::lt);
+	InsertOperatorToTable(TTMATH_TEXT("-"),  MatOperator::sub);
+	InsertOperatorToTable(TTMATH_TEXT("+"),  MatOperator::add);
+	InsertOperatorToTable(TTMATH_TEXT("/"),  MatOperator::div);
+	InsertOperatorToTable(TTMATH_TEXT("*"),  MatOperator::mul);
+	InsertOperatorToTable(TTMATH_TEXT("^"),  MatOperator::pow);
 }
 
 
@@ -1845,12 +1881,12 @@ void CreateMathematicalOperatorsTable()
 	e.g.
 	true when str1="test" and str2="te"
 */
-bool IsSubstring(const std::string & str1, const std::string & str2)
+bool IsSubstring(const tt_string & str1, const tt_string & str2)
 {
 	if( str2.length() > str1.length() )
 		return false;
 
-	for(std::string::size_type i=0 ; i<str2.length() ; ++i)
+	for(tt_string::size_type i=0 ; i<str2.length() ; ++i)
 		if( str1[i] != str2[i] )
 			return false;
 
@@ -1863,7 +1899,7 @@ return true;
 */
 void ReadMathematicalOperator(Item & result)
 {
-std::string oper;
+tt_string oper;
 typename OperatorsTable::iterator iter_old, iter_new;
 
 	iter_old = operators_table.end();
@@ -2393,7 +2429,6 @@ Parser(): default_stack_size(100)
 	base = 10;
 	deg_rad_grad = 1;
 	error = err_ok;
-	factorial_max.SetZero();
 
 	CreateFunctionsTable();
 	CreateVariablesTable();
@@ -2413,7 +2448,6 @@ Parser<ValueType> & operator=(const Parser<ValueType> & p)
 	base              = p.base;
 	deg_rad_grad      = p.deg_rad_grad;
 	error             = err_ok;
-	factorial_max = p.factorial_max;
 
 	/*
 		we don't have to call 'CreateFunctionsTable()' etc.
@@ -2495,22 +2529,11 @@ void SetFunctions(const Objects * pf)
 }
 
 
-/*!
-	you will not be allowed to calculate the factorial 
-	if its argument is greater than 'm'
-	there'll be: ErrorCode::err_too_big_factorial
-	default 'factorial_max' is zero which means you can calculate what you want to
-*/
-void SetFactorialMax(const ValueType & m)
-{
-	factorial_max = m;
-}
-
 
 /*!
 	the main method using for parsing string
 */
-ErrorCode Parse(const char * str)
+ErrorCode Parse(const tt_char * str)
 {
 	stack_index  = 0;
 	pstring      = str;
@@ -2533,11 +2556,11 @@ return error;
 }
 
 
-
-
-
 };
 
+
+
+
 } // namespace
 
 

ttmath/ttmathtypes.h

@@ -1,7 +1,7 @@
 /*
  * This file is a part of TTMath Bignum Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /* 
@@ -54,7 +54,7 @@
 
 #include <stdexcept>
 #include <sstream>
-
+#include <vector>
 
 /*!
 	the version of the library
@@ -63,9 +63,9 @@
 	if zero that means this is the release version of the library
 */
 #define TTMATH_MAJOR_VER		0
-#define TTMATH_MINOR_VER		8
-#define TTMATH_REVISION_VER		4
-#define TTMATH_PRERELEASE_VER	0
+#define TTMATH_MINOR_VER		9
+#define TTMATH_REVISION_VER		0
+#define TTMATH_PRERELEASE_VER	1
 
 
 /*!
@@ -120,6 +120,7 @@ namespace ttmath
 	typedef unsigned int uint;
 	typedef signed   int sint;
 
+
 	/*!
 		this type is twice bigger than uint
 		(64bit on a 32bit platforms)
@@ -128,43 +129,39 @@ namespace ttmath
 		but it is defined in C99 and in upcoming C++0x /3.9.1 (2)/ and many compilers support it
 
 		this type is used in UInt::MulTwoWords and UInt::DivTwoWords when macro TTMATH_NOASM is defined
-		but only on a 32bit platform
 	*/
-	#ifdef TTMATH_NOASM
 	typedef unsigned long long int ulint;
-	#endif
-
-	/*!
-		how many bits there are in the uint type
-	*/
-	#define TTMATH_BITS_PER_UINT 32u
 
 	/*!
 		the mask for the highest bit in the unsigned 32bit word (2^31)
 	*/
-	#define TTMATH_UINT_HIGHEST_BIT 2147483648u
+	const uint TTMATH_UINT_HIGHEST_BIT = 0x80000000ul;
 
 	/*!
 		the max value of the unsigned 32bit word (2^32 - 1)
 		(all bits equal one)
 	*/
-	#define TTMATH_UINT_MAX_VALUE 4294967295u
+	const uint TTMATH_UINT_MAX_VALUE = 0xfffffffful;
 
 	/*!
 		the number of words (32bit words on 32bit platform)
 		which are kept in built-in variables for a Big<> type
 		(these variables are defined in ttmathbig.h)
 	*/
-	#define TTMATH_BUILTIN_VARIABLES_SIZE 256u
+	const uint TTMATH_BUILTIN_VARIABLES_SIZE  = 256u;
 
 #else
 
 	/*!
 		on 64bit platforms one word (uint, sint) will be equal 64bits
 	*/
-	typedef unsigned long uint;
-	typedef signed   long sint;
-
+	#if defined(_MSC_VER)
+		typedef unsigned __int64 uint;
+		typedef signed __int64 sint;
+	#else
+		typedef unsigned long long uint;
+		typedef signed long long sint;
+	#endif
 	/*!
 		on 64bit platform we do not define ulint
 		sizeof(long long) is 8 (64bit) but we need 128bit
@@ -174,33 +171,61 @@ namespace ttmath
 	*/
 	//typedef unsigned long long int ulint;
 
-	/*!
-		how many bits there are in the uint type
-	*/
-	#define TTMATH_BITS_PER_UINT 64ul
-
 	/*!
 		the mask for the highest bit in the unsigned 64bit word (2^63)
 	*/
-	#define TTMATH_UINT_HIGHEST_BIT 9223372036854775808ul
+	const uint TTMATH_UINT_HIGHEST_BIT = 0x8000000000000000ul;
 
 	/*!
 		the max value of the unsigned 64bit word (2^64 - 1)
 		(all bits equal one)
 	*/
-	#define TTMATH_UINT_MAX_VALUE 18446744073709551615ul
+	const uint TTMATH_UINT_MAX_VALUE = 0xfffffffffffffffful;
 
 	/*!
 		the number of words (64bit words on 64bit platforms)
 		which are kept in built-in variables for a Big<> type
 		(these variables are defined in ttmathbig.h)
 	*/
-	#define TTMATH_BUILTIN_VARIABLES_SIZE 128ul
+	const uint TTMATH_BUILTIN_VARIABLES_SIZE = 128ul;
 
 #endif
+
+	const uint TTMATH_BITS_PER_UINT = (sizeof(uint)*8);
+
 }
 
 
+#if defined(UNICODE) || defined(_UNICODE)
+#define TTMATH_USE_WCHAR
+#endif
+
+
+#ifdef TTMATH_USE_WCHAR
+
+	typedef	wchar_t					tt_char;
+	typedef	std::wstring			tt_string;
+	typedef std::wostringstream		tt_ostringstream;
+	typedef std::wostream			tt_ostream;
+	typedef std::wistream			tt_istream;
+	#define TTMATH_TEXT_HELPER(txt)	L##txt
+
+#else
+
+	typedef	char					tt_char;
+	typedef	std::string				tt_string;
+	typedef std::ostringstream		tt_ostringstream;
+	typedef std::ostream			tt_ostream;
+	typedef std::istream			tt_istream;
+	#define TTMATH_TEXT_HELPER(txt)	txt
+
+#endif
+
+#define TTMATH_TEXT(txt) 	TTMATH_TEXT_HELPER(txt)
+
+
+
+
 
 /*!
 	characters which represent the comma operator
@@ -237,6 +262,29 @@ namespace ttmath
 
 
 
+/*!
+	this is a limit when calculating Karatsuba multiplication
+	if the size of a vector is smaller than TTMATH_USE_KARATSUBA_MULTIPLICATION_FROM_SIZE
+	the Karatsuba algorithm will use standard schoolbook multiplication
+*/
+#ifdef __GNUC__
+#define TTMATH_USE_KARATSUBA_MULTIPLICATION_FROM_SIZE 3
+#else
+#define TTMATH_USE_KARATSUBA_MULTIPLICATION_FROM_SIZE 5
+#endif
+
+
+/*!
+	this is a special value used when calculating the Gamma(x) function
+	if x is greater than this value then the Gamma(x) will be calculated using
+	some kind of series
+
+	don't use smaller values than about 100
+*/
+#define TTMATH_GAMMA_BOUNDARY 2000
+
+
+
 
 namespace ttmath
 {
@@ -271,7 +319,6 @@ namespace ttmath
 		err_object_exists,
 		err_unknown_object,
 		err_still_calculating,
-		err_too_big_factorial,
 		err_in_short_form_used_function
 	};
 
@@ -302,20 +349,20 @@ namespace ttmath
 	*/
 	class ExceptionInfo
 	{
-	const char * file;
+	const tt_char * file;
 	int line;
 
 	public:
 		ExceptionInfo() : file(0), line(0) {}
-		ExceptionInfo(const char * f, int l) : file(f), line(l) {}
+		ExceptionInfo(const tt_char * f, int l) : file(f), line(l) {}
 
-		std::string Where() const
+		tt_string Where() const
 		{
 			if( !file )
-				return "unknown";
+				return TTMATH_TEXT("unknown");
 
-			std::ostringstream result;
-			result << file << ":" << line;
+			tt_ostringstream result;
+			result << file << TTMATH_TEXT(":") << line;
 
 		return result.str();
 		}
@@ -329,7 +376,7 @@ namespace ttmath
 		can throw an exception of this type
 
 		If you compile with gcc you can get a small benefit 
-		from using method Where() (it returns std::string with
+		from using method Where() (it returns std::string (or std::wstring) with
 		the name and the line of a file where the macro TTMATH_REFERENCE_ASSERT
 		was used)
 
@@ -349,7 +396,7 @@ namespace ttmath
 			foo.Add(foo);
 		but there are only few methods which can do that
 	*/
-	class ReferenceError : public std::logic_error, ExceptionInfo
+	class ReferenceError : public std::logic_error, public ExceptionInfo
 	{
 	public:
 
@@ -357,12 +404,12 @@ namespace ttmath
 		{
 		}
 
-		ReferenceError(const char * f, int l) :
+		ReferenceError(const tt_char * f, int l) :
 							std::logic_error ("reference error"), ExceptionInfo(f,l)
 		{
 		}
 
-		std::string Where() const
+		tt_string Where() const
 		{
 			return ExceptionInfo::Where();
 		}
@@ -377,11 +424,11 @@ namespace ttmath
 		of this type
 
 		if you compile with gcc you can get a small benefit 
-		from using method Where() (it returns std::string with
+		from using method Where() (it returns std::string (or std::wstring) with
 		the name and the line of a file where the macro TTMATH_ASSERT
 		was used)
 	*/
-	class RuntimeError : public std::runtime_error, ExceptionInfo
+	class RuntimeError : public std::runtime_error, public ExceptionInfo
 	{
 	public:
 
@@ -389,12 +436,12 @@ namespace ttmath
 		{
 		}
 
-		RuntimeError(const char * f, int l) :
+		RuntimeError(const tt_char * f, int l) :
 						std::runtime_error ("internal error"), ExceptionInfo(f,l)
 		{
 		}
 
-		std::string Where() const
+		tt_string Where() const
 		{
 			return ExceptionInfo::Where();
 		}
@@ -409,11 +456,22 @@ namespace ttmath
 
 		#if defined(__FILE__) && defined(__LINE__)
 
+			#ifdef TTMATH_USE_WCHAR
+				#define TTMATH_FILE_HELPER2(arg)  L##arg
+				#define TTMATH_FILE_HELPER(x)     TTMATH_FILE_HELPER2(x)
+				#define TTMATH_FILE               TTMATH_FILE_HELPER(__FILE__)
+			#else
+				#define TTMATH_FILE               __FILE__
+			#endif
+
 			#define TTMATH_REFERENCE_ASSERT(expression) \
-				if( &(expression) == this ) throw ttmath::ReferenceError(__FILE__, __LINE__);
+				if( &(expression) == this ) throw ttmath::ReferenceError(TTMATH_FILE, __LINE__);
 
 			#define TTMATH_ASSERT(expression) \
-				if( !(expression) ) throw ttmath::RuntimeError(__FILE__, __LINE__);
+				if( !(expression) ) throw ttmath::RuntimeError(TTMATH_FILE, __LINE__);
+
+			#define TTMATH_VERIFY(expression) \
+				if( !(expression) ) throw ttmath::RuntimeError(TTMATH_TEXT(__FILE__), __LINE__);
 
 		#else
 
@@ -422,19 +480,32 @@ namespace ttmath
 
 			#define TTMATH_ASSERT(expression) \
 				if( !(expression) ) throw RuntimeError();
+
+			#define TTMATH_VERIFY(expression) \
+				if( !(expression) ) throw RuntimeError();
 		#endif
 
 	#else
 		#define TTMATH_REFERENCE_ASSERT(expression)
 		#define TTMATH_ASSERT(expression)
+		#define TTMATH_VERIFY(expression)	(void)(expression);
 	#endif
 
-
+	#if !defined(LOG_PRINTF)
+		#define LOG_PRINTF printf
+	#endif
 
 	#ifdef TTMATH_DEBUG_LOG
 	
-		#define TTMATH_LOG(msg) \
+		#ifdef TTMATH_USE_WCHAR
+			#define TTMATH_LOG_HELPER(msg) \
+				PrintLog(L##msg, std::wcout);
+		#else
+			#define TTMATH_LOG_HELPER(msg) \
 				PrintLog(msg, std::cout);
+		#endif
+
+		#define TTMATH_LOG(msg) TTMATH_LOG_HELPER(msg)
 
 	#else
 
@@ -443,7 +514,9 @@ namespace ttmath
 	#endif
 
 
+
 } // namespace
 
 
 #endif
+

ttmath/ttmathuint.h

@@ -1,7 +1,7 @@
 /*
  * This file is a part of TTMath Bignum Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /*
@@ -40,6 +40,7 @@
 #ifndef headerfilettmathuint
 #define headerfilettmathuint
 
+
 /*!
 	\file ttmathuint.h
     \brief template class UInt<uint>
@@ -48,9 +49,12 @@
 #include <iostream>
 #include <iomanip>
 
+
 #include "ttmathtypes.h"
 
-
+#if defined(_MSC_VER)
+	#pragma warning(disable:4127) // conditional expression is constant
+#endif
 
 /*!
     \brief a namespace for the TTMath library
@@ -84,7 +88,10 @@ public:
 
 		it prints the table in a nice form of several columns
 	*/
-	void PrintTable(std::ostream & output) const
+	#ifndef TTMATH_USE_WCHAR
+	// gcc has a problem with std::setfill when wchar_t is used
+
+	void PrintTable(tt_ostream & output) const
 	{
 		// how many columns there'll be
 		const int columns = 8;
@@ -92,7 +99,7 @@ public:
 		int c = 1;
 		for(int i=value_size-1 ; i>=0 ; --i)
 		{
-			output	<< "0x" << std::setfill('0');
+			output << TTMATH_TEXT("0x") << std::setfill('0');
 
 			#ifdef TTMATH_PLATFORM32
 				output << std::setw(8);
@@ -104,7 +111,7 @@ public:
 
 			if( i>0 )
 			{
-				output << ", ";		
+				output << TTMATH_TEXT(", ");		
 
 				if( ++c > columns )
 				{
@@ -116,14 +123,14 @@ public:
 
 		output << std::dec << std::endl;
 	}
+	#endif
 
-
-	void PrintLog(const char * msg, std::ostream & output) const
+	void PrintLog(const tt_char * msg, tt_ostream & output) const
 	{
 		output << msg << std::endl;
 
 		for(uint i=0 ; i<value_size ; ++i)
-			output << " table[" << i << "]: " << table[i] << std::endl;
+			output << TTMATH_TEXT(" table[") << i << TTMATH_TEXT("]: ") << table[i] << std::endl;
 	}
 
 
@@ -840,8 +847,10 @@ public:
 
 	/*!
 		the multiplication 'this' = 'this' * ss2
+
+		algorithm: 100 - means automatically choose the fastest algorithm
 	*/
-	uint Mul(const UInt<value_size> & ss2, uint algorithm = 2)
+	uint Mul(const UInt<value_size> & ss2, uint algorithm = 100)
 	{
 		switch( algorithm )
 		{
@@ -849,8 +858,14 @@ public:
 			return Mul1(ss2);
 
 		case 2:
-		default:
 			return Mul2(ss2);
+
+		case 3:
+			return Mul3(ss2);
+
+		case 100:
+		default:
+			return MulFastest(ss2);
 		}
 	}
 
@@ -860,10 +875,12 @@ public:
 
 		since the 'result' is twice bigger than 'this' and 'ss2'
 		this method never returns a carry
+
+		algorithm: 100 - means automatically choose the fastest algorithm
 	*/
 	void MulBig(const UInt<value_size> & ss2,
 				UInt<value_size*2> & result,
-				uint algorithm = 2)
+				uint algorithm = 100)
 	{
 		switch( algorithm )
 		{
@@ -871,8 +888,14 @@ public:
 			return Mul1Big(ss2, result);
 
 		case 2:
-		default:
 			return Mul2Big(ss2, result);
+
+		case 3:
+			return Mul3Big(ss2, result);
+
+		case 100:
+		default:
+			return MulFastestBig(ss2, result);
 		}
 	}
 
@@ -964,7 +987,7 @@ public:
 	uint Mul2(const UInt<value_size> & ss2)
 	{
 	UInt<value_size*2> result;
-	uint i;
+	uint i, c = 0;
 
 		Mul2Big(ss2, result);
 
@@ -975,11 +998,14 @@ public:
 		// testing carry
 		for( ; i<value_size*2 ; ++i)
 			if( result.table[i] != 0 )
-				return 1;
+			{
+				c = 1;
+				break;
+			}
 
 		TTMATH_LOG("UInt::Mul2")
 
-	return 0;
+	return c;
 	}
 
 
@@ -991,44 +1017,385 @@ public:
 	*/
 	void Mul2Big(const UInt<value_size> & ss2, UInt<value_size*2> & result)
 	{
-	uint r2,r1;
-	uint x1size=value_size, x2size=value_size;
+		Mul2Big2<value_size>(table, ss2.table, result);
+
+		TTMATH_LOG("UInt::Mul2Big")
+	}
+
+
+private:
+
+	/*!
+		an auxiliary method for calculating the multiplication
+
+		arguments we're taking as pointers (this is to improve the Mul3Big2()- avoiding
+		unnecessary copying objects), the result should be taken as a pointer too,
+		but at the moment there is no method AddTwoInts() which can operate on pointers
+	*/
+	template<uint ss_size>
+	void Mul2Big2(const uint * ss1, const uint * ss2, UInt<ss_size*2> & result)
+	{
+	uint x1size  = ss_size, x2size  = ss_size;
 	uint x1start = 0,       x2start = 0;
 
+		if( ss_size > 2 )
+		{
+			// if the ss_size is smaller than or equal to 2
+			// there is no sense to set x1size (and others) to another values
+
+			for(x1size=ss_size ; x1size>0 && ss1[x1size-1]==0 ; --x1size);
+			for(x2size=ss_size ; x2size>0 && ss2[x2size-1]==0 ; --x2size);
+
+			for(x1start=0 ; x1start<x1size && ss1[x1start]==0 ; ++x1start);
+			for(x2start=0 ; x2start<x2size && ss2[x2start]==0 ; ++x2start);
+		}
+
+		Mul2Big3<ss_size>(ss1, ss2, result, x1start, x1size, x2start, x2size);
+	}
+
+
+
+	/*!
+		an auxiliary method for calculating the multiplication
+	*/
+	template<uint ss_size>
+	void Mul2Big3(const uint * ss1, const uint * ss2, UInt<ss_size*2> & result, uint x1start, uint x1size, uint x2start, uint x2size)
+	{
+	uint r2, r1;
+
 		result.SetZero();
 
-		if( value_size > 2 )
+		if( x1size==0 || x2size==0 )
+			return;
+
+		for(uint x1=x1start ; x1<x1size ; ++x1)
 		{
-			// if the value_size is smaller than or equal to 2
-			// there is no sense to set x1size (and others) to another values
+			for(uint x2=x2start ; x2<x2size ; ++x2)
+			{
+				MulTwoWords(ss1[x1], ss2[x2], &r2, &r1);
+				result.AddTwoInts(r2, r1, x2+x1);
+				// here will never be a carry
+			}
+		}
+	}
+
+
+public:
+
+
+	/*!
+		multiplication: this = this * ss2
+
+		This is Karatsuba Multiplication algorithm, we're using it when value_size is greater than
+		or equal to TTMATH_USE_KARATSUBA_MULTIPLICATION_FROM_SIZE macro (defined in ttmathuint.h).
+		If value_size is smaller then we're using Mul2Big() instead.
+
+		Karatsuba multiplication:
+		Assume we have:
+			this = x = x1*B^m + x0
+			ss2  = y = y1*B^m + y0
+		where x0 and y0 are less than B^m
+		the product from multiplication we can show as:
+	    x*y = (x1*B^m + x0)(y1*B^m + y0) = z2*B^(2m) + z1*B^m + z0
+		where
+		    z2 = x1*y1
+			z1 = x1*y0 + x0*y1
+			z0 = x0*y0
+		this is standard schoolbook algorithm with O(n^2), Karatsuba observed that z1 can be given in other form:
+			z1 = (x1 + x0)*(y1 + y0) - z2 - z0    / z1 = (x1*y1 + x1*y0 + x0*y1 + x0*y0) - x1*y1 - x0*y0 = x1*y0 + x0*y1 /
+		and to calculate the multiplication we need only three multiplications (with some additions and subtractions)
+
+		Our objects 'this' and 'ss2' we divide into two parts and by using recurrence we calculate the multiplication.
+		Karatsuba multiplication has O( n^(ln(3)/ln(2)) )
+	*/
+	uint Mul3(const UInt<value_size> & ss2)
+	{
+	UInt<value_size*2> result;
+	uint i, c = 0;
+
+		Mul3Big(ss2, result);
+
+		// copying result
+		for(i=0 ; i<value_size ; ++i)
+			table[i] = result.table[i];
+
+		// testing carry
+		for( ; i<value_size*2 ; ++i)
+			if( result.table[i] != 0 )
+			{
+				c = 1;
+				break;
+			}
+
+		TTMATH_LOG("UInt::Mul3")
+
+	return c;
+	}
+
+
+
+	/*!
+		multiplication: result = this * ss2
+
+		result is twice bigger than this and ss2,
+		this method never returns carry,
+		(Karatsuba multiplication)
+	*/
+	void Mul3Big(const UInt<value_size> & ss2, UInt<value_size*2> & result)
+	{
+		Mul3Big2<value_size>(table, ss2.table, result.table);
+
+		TTMATH_LOG("UInt::Mul3Big")
+	}
+
+
+
+private:
+
+	/*!
+		an auxiliary method for calculating the Karatsuba multiplication
+
+		result_size is equal ss_size*2
+	*/
+	template<uint ss_size>
+	void Mul3Big2(const uint * ss1, const uint * ss2, uint * result)
+	{
+	const uint * x1, * x0, * y1, * y0;
+
+
+		if( ss_size>1 && ss_size<TTMATH_USE_KARATSUBA_MULTIPLICATION_FROM_SIZE )
+		{
+			UInt<ss_size*2> res;
+			Mul2Big2<ss_size>(ss1, ss2, res);
+
+			for(uint i=0 ; i<ss_size*2 ; ++i)
+				result[i] = res.table[i];
+
+		return;
+		}
+		else
+		if( ss_size == 1 )
+		{
+			return MulTwoWords(*ss1, *ss2, &result[1], &result[0]);
+		}
+
+
+		if( (ss_size & 1) == 1 )
+		{
+			// ss_size is odd
+			x0 = ss1;
+			y0 = ss2;
+			x1 = ss1 + ss_size / 2 + 1;
+			y1 = ss2 + ss_size / 2 + 1;
+
+			// the second vectors (x1 and y1) are smaller about one from the first ones (x0 and y0)
+			Mul3Big3<ss_size/2 + 1, ss_size/2, ss_size*2>(x1, x0, y1, y0, result);
+		}
+		else
+		{
+			// ss_size is even
+			x0 = ss1;
+			y0 = ss2;
+			x1 = ss1 + ss_size / 2;
+			y1 = ss2 + ss_size / 2;
+
+			// all four vectors (x0 x1 y0 y1) are equal in size
+			Mul3Big3<ss_size/2, ss_size/2, ss_size*2>(x1, x0, y1, y0, result);
+		}
+	}
+
+
+
+#ifdef _MSC_VER
+#pragma warning (disable : 4717)
+//warning C4717: recursive on all control paths, function will cause runtime stack overflow
+//we have the stop point in Mul3Big2() method
+#endif
+
+
+	/*!
+		an auxiliary method for calculating the Karatsuba multiplication
+
+			x = x1*B^m + x0
+			y = y1*B^m + y0
+
+			first_size  - is the size of vectors: x0 and y0
+			second_size - is the size of vectors: x1 and y1 (can be either equal first_size or smaller about one from first_size)
+
+			x*y = (x1*B^m + x0)(y1*B^m + y0) = z2*B^(2m) + z1*B^m + z0
+		      where
+			   z0 = x0*y0
+			   z2 = x1*y1
+			   z1 = (x1 + x0)*(y1 + y0) - z2 - z0
+	*/
+	template<uint first_size, uint second_size, uint result_size>
+	void Mul3Big3(const uint * x1, const uint * x0, const uint * y1, const uint * y0, uint * result)
+	{
+	uint i, c, xc, yc;
+
+		UInt<first_size>   temp, temp2;
+		UInt<first_size*3> z1;
+
+		// z0 and z2 we store directly in the result (we don't use any temporary variables)
+		Mul3Big2<first_size>(x0, y0, result);                  // z0
+		Mul3Big2<second_size>(x1, y1, result+first_size*2);    // z2
+
+		// now we calculate z1
+		// temp  = (x0 + x1)
+		// temp2 = (y0 + y1)
+		// we're using temp and temp2 with UInt<first_size>, although there can be a carry but
+		// we simple remember it in xc and yc (xc and yc can be either 0 or 1),
+		// and (x0 + x1)*(y0 + y1) we calculate in this way (schoolbook algorithm):
+		//
+		//                 xc     |     temp
+		//                 yc     |     temp2
+		//               --------------------
+		//               (temp    *   temp2)
+		//               xc*temp2 |
+		//               yc*temp  |
+		//       xc*yc |
+		//       ----------     z1     --------
+		//
+		// and the result is never larger in size than 3*first_size
+
+		xc = AddVector(x0, x1, first_size, second_size, temp.table);
+		yc = AddVector(y0, y1, first_size, second_size, temp2.table);
+
+		Mul3Big2<first_size>(temp.table, temp2.table, z1.table);
+
+		// clearing the rest of z1
+		for(i=first_size*2 ; i<first_size*3 ; ++i)
+			z1.table[i] = 0;
+
+
+		if( xc )
+		{
+			c = AddVector(z1.table+first_size, temp2.table, first_size*3-first_size, first_size, z1.table+first_size);
+			TTMATH_ASSERT( c==0 )
+		}
+
+		if( yc )
+		{
+			c = AddVector(z1.table+first_size, temp.table, first_size*3-first_size, first_size, z1.table+first_size);
+			TTMATH_ASSERT( c==0 )
+		}
+
+
+		if( xc && yc )
+		{
+			for( i=first_size*2 ; i<first_size*3 ; ++i )
+				if( ++z1.table[i] != 0 )
+					break;  // break if there was no carry
+		}
+
+		// z1 = z1 - z2
+		c = SubVector(z1.table, result+first_size*2, first_size*3, second_size*2, z1.table);
+		TTMATH_ASSERT(c==0)
+
+		// z1 = z1 - z0
+		c = SubVector(z1.table, result, first_size*3, first_size*2, z1.table);
+		TTMATH_ASSERT(c==0)
+
+		// here we've calculated the z1
+		// now we're adding it to the result
+
+		if( first_size > second_size )
+		{
+			uint z1_size = result_size - first_size;
+			TTMATH_ASSERT( z1_size <= first_size*3 )
+
+			for(i=z1_size ; i<first_size*3 ; ++i)
+				TTMATH_ASSERT( z1.table[i] == 0 )
+				;
+
+			c = AddVector(result+first_size, z1.table, result_size-first_size, z1_size, result+first_size);
+			TTMATH_ASSERT(c==0)
+		}
+		else
+		{
+			c = AddVector(result+first_size, z1.table, result_size-first_size, first_size*3, result+first_size);
+			TTMATH_ASSERT(c==0)
+		}
+	}
+
+
+#ifdef _MSC_VER
+#pragma warning (default : 4717)
+#endif
+
+
+public:
+
+
+	/*!
+		multiplication this = this * ss2
+	*/
+	uint MulFastest(const UInt<value_size> & ss2)
+	{
+	UInt<value_size*2> result;
+	uint i, c = 0;
+
+		MulFastestBig(ss2, result);
+
+		// copying result
+		for(i=0 ; i<value_size ; ++i)
+			table[i] = result.table[i];
+
+		// testing carry
+		for( ; i<value_size*2 ; ++i)
+			if( result.table[i] != 0 )
+			{
+				c = 1;
+				break;
+			}
+
+		TTMATH_LOG("UInt::MulFastest")
+
+	return c;
+	}
+
+
+	/*!
+		multiplication result = this * ss2
+
+		this method is trying to select the fastest algorithm
+		(in the future this method can be improved)
+	*/
+	void MulFastestBig(const UInt<value_size> & ss2, UInt<value_size*2> & result)
+	{
+		if( value_size < TTMATH_USE_KARATSUBA_MULTIPLICATION_FROM_SIZE )
+			return Mul2Big(ss2, result);
+
+		uint x1size  = value_size, x2size  = value_size;
+		uint x1start = 0,          x2start = 0;
 
 		for(x1size=value_size ; x1size>0 && table[x1size-1]==0 ; --x1size);
 		for(x2size=value_size ; x2size>0 && ss2.table[x2size-1]==0 ; --x2size);
 
 		if( x1size==0 || x2size==0 )
 		{
-				TTMATH_LOG("UInt::Mul2Big")
+			// either 'this' or 'ss2' is equal zero - the result is zero too
+			result.SetZero();
 			return;
 		}
 
 		for(x1start=0 ; x1start<x1size && table[x1start]==0 ; ++x1start);
 		for(x2start=0 ; x2start<x2size && ss2.table[x2start]==0 ; ++x2start);
-		}
 
-		for(uint x1=x1start ; x1<x1size ; ++x1)
-		{
-			for(uint x2=x2start ; x2<x2size ; ++x2)
-			{
-				MulTwoWords(table[x1], ss2.table[x2], &r2, &r1);
-				result.AddTwoInts(r2,r1,x2+x1);
-				// here will never be a carry
-			}
-		}
+		uint distancex1 = x1size - x1start;
+		uint distancex2 = x2size - x2start;
 
-		TTMATH_LOG("UInt::Mul2Big")
-	}
+		if( distancex1 < 3 || distancex2 < 3 )
+			// either 'this' or 'ss2' have only 2 (or 1) item different from zero (side by side)
+			// (this condition in the future can be improved)
+			return Mul2Big3<value_size>(table, ss2.table, result, x1start, x1size, x2start, x2size);
 
 
+		// Karatsuba multiplication
+		Mul3Big(ss2, result);
+
+		TTMATH_LOG("UInt::MulFastestBig")
+	}
 
 
 	/*!
@@ -1453,7 +1820,7 @@ private:
 
 		if( Div2_DivisorGreaterOrEqual(	divisor, remainder,
 										table_id, index,
-										divisor_table_id, divisor_index) )
+										divisor_index) )
 		{
 			TTMATH_LOG("UInt::Div2_FindLeadingBitsAndCheck")
 			return 0;
@@ -1473,7 +1840,7 @@ private:
 	bool Div2_DivisorGreaterOrEqual(	const UInt<value_size> & divisor,
 										UInt<value_size> * remainder,
 										uint table_id, uint index,
-										uint divisor_table_id, uint divisor_index  )
+										uint divisor_index  )
 	{
 		if( divisor_index > index )
 		{
@@ -1604,7 +1971,7 @@ private:
 		for(uint i = j+1 ; i<value_size ; ++i)
 			q.table[i] = 0;
 
-		while( true )
+		for (;;)
 		{
 			u1 = table[j+n-1];
 			u0 = table[j+n-2];
@@ -1921,7 +2288,7 @@ public:
 	*/
 	bool IsTheLowestBitSet() const
 	{
-		return (*table & 1) != 0;
+		return (table[0] & 1) != 0;
 	}
 
 
@@ -2005,7 +2372,7 @@ public:
 
 
 	/*!
-		this method converts a digit into a char
+		this method converts a digit into a tt_char
 		digit should be from <0,F>
 		(we don't have to get a base)
 
@@ -2015,12 +2382,12 @@ public:
 			10 -> A
 			15 -> F
 	*/
-	static uint DigitToChar(uint digit)
+	static tt_char DigitToChar(uint digit)
 	{
 		if( digit < 10 )
-			return digit + '0';
+			return (tt_char)(digit + '0');
 
-	return digit - 10 + 'A';
+	return((tt_char)(digit - 10 + 'A'));
 	}
 
 
@@ -2141,7 +2508,7 @@ public:
 		this constant 10 has the int type (signed int), if we don't give such
 		operators and constructors the compiler will not compile the program,
 		because it has to make a conversion and doesn't know into which type
-		(the UInt class has operator=(const char*), operator=(uint) etc.)
+		(the UInt class has operator=(const tt_char*), operator=(uint) etc.)
 	*/
 	UInt<value_size> & operator=(sint i)
 	{
@@ -2256,18 +2623,18 @@ public:
 	/*!
 		a constructor for converting a string to this class (with the base=10)
 	*/
-	UInt(const char * s)
+	UInt(const tt_char * s)
 	{
 		FromString(s);
 
-		TTMATH_LOG("UInt::UInt(const char *)")
+		TTMATH_LOG("UInt::UInt(const tt_char *)")
 	}
 
 
 	/*!
 		a constructor for converting a string to this class (with the base=10)
 	*/
-	UInt(const std::string & s)
+	UInt(const tt_string & s)
 	{
 		FromString( s.c_str() );
 	}
@@ -2282,6 +2649,7 @@ public:
 	{
 	}
 
+
 	/*!
 		a copy constructor
 	*/
@@ -2332,10 +2700,10 @@ public:
 	/*!
 		this method converts the value to a string with a base equal 'b'
 	*/
-	void ToString(std::string & result, uint b = 10) const
+	void ToString(tt_string & result, uint b = 10) const
 	{
 	UInt<value_size> temp( *this );
-	char character;
+	tt_char character;
 	uint rem;
 
 		result.clear();
@@ -2346,7 +2714,7 @@ public:
 		do
 		{
 			temp.DivInt(b, &rem);
-			character = DigitToChar( rem );
+			character = static_cast<tt_char>( DigitToChar(rem) );
 			result.insert(result.begin(), character);
 		}
 		while( !temp.IsZero() );
@@ -2360,7 +2728,7 @@ public:
 	/*
 		this method's ommiting any white characters from the string
 	*/
-	static void SkipWhiteCharacters(const char * & c)
+	static void SkipWhiteCharacters(const tt_char * & c)
 	{
 		while( (*c==' ' ) || (*c=='\t') || (*c==13 ) || (*c=='\n') )
 			++c;
@@ -2384,7 +2752,7 @@ public:
 
 		value_read (if exists) tells whether something has actually been read (at least one digit)
 	*/
-	uint FromString(const char * s, uint b = 10, const char ** after_source = 0, bool * value_read = 0)
+	uint FromString(const tt_char * s, uint b = 10, const tt_char ** after_source = 0, bool * value_read = 0)
 	{
 	UInt<value_size> base( b );
 	UInt<value_size> temp;
@@ -2434,7 +2802,7 @@ public:
 
 		(it returns carry=1 if the value will be too big or an incorrect base 'b' is given)
 	*/
-	uint FromString(const std::string & s, uint b = 10)
+	uint FromString(const tt_string & s, uint b = 10)
 	{
 		return FromString( s.c_str(), b );
 	}
@@ -2444,11 +2812,11 @@ public:
 	/*!
 		this operator converts a string into its value (with base = 10)
 	*/
-	UInt<value_size> & operator=(const char * s)
+	UInt<value_size> & operator=(const tt_char * s)
 	{
 		FromString(s);
 
-		TTMATH_LOG("UInt::operator=(const char *)")
+		TTMATH_LOG("UInt::operator=(const tt_char *)")
 
 	return *this;
 	}
@@ -2457,7 +2825,7 @@ public:
 	/*!
 		this operator converts a string into its value (with base = 10)
 	*/
-	UInt<value_size> & operator=(const std::string & s)
+	UInt<value_size> & operator=(const tt_string & s)
 	{
 		FromString( s.c_str() );
 
@@ -2825,9 +3193,15 @@ public:
 	*
 	*/
 
-	friend std::ostream & operator<<(std::ostream & s, const UInt<value_size> & l)
+
+	/*!
+		output for standard streams
+
+		tt_ostream is either std::ostream or std::wostream
+	*/
+	friend tt_ostream & operator<<(tt_ostream & s, const UInt<value_size> & l)
 	{
-	std::string ss;
+	tt_string ss;
 
 		l.ToString(ss);
 		s << ss;
@@ -2837,12 +3211,17 @@ public:
 
 
 
-	friend std::istream & operator>>(std::istream & s, UInt<value_size> & l)
-	{
-	std::string ss;
+	/*!
+		input from standard streams
 
-	// char for operator>>
-	unsigned char z;
+		tt_istream is either std::istream or std::wistream
+	*/
+	friend tt_istream & operator>>(tt_istream & s, UInt<value_size> & l)
+	{
+	tt_string ss;
+
+	// tt_char for operator>>
+	tt_char z;
 
 		// operator>> omits white characters if they're set for ommiting
 		s >> z;
@@ -2851,10 +3230,10 @@ public:
 		while( s.good() && CharToDigit(z, 10)>=0 )
 		{
 			ss += z;
-			z = s.get();
+			z = static_cast<tt_char>(s.get());
 		}
 
-		// we're leaving the last readed character
+		// we're leaving the last read character
 		// (it's not belonging to the value)
 		s.unget();
 
@@ -2874,7 +3253,6 @@ public:
 			ttmathuint_noasm.h
 	*/
 
-#ifdef TTMATH_NOASM
 	static uint AddTwoWords(uint a, uint b, uint carry, uint * result);
 	static uint SubTwoWords(uint a, uint b, uint carry, uint * result);
 
@@ -2899,8 +3277,6 @@ public:
 	static void MultiplySubtract(uint_ & u_, unsigned int & u3, unsigned int & q, uint_ v_);
 
 #endif // TTMATH_PLATFORM64
-#endif // TTMATH_NOASM
-
 
 private:
 	uint Rcl2_one(uint c);
@@ -2909,12 +3285,13 @@ private:
 	uint Rcr2(uint bits, uint c);
 
 public:
-	
 	uint Add(const UInt<value_size> & ss2, uint c=0);
 	uint AddInt(uint value, uint index = 0);
 	uint AddTwoInts(uint x2, uint x1, uint index);
+	static uint AddVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result);
 	uint Sub(const UInt<value_size> & ss2, uint c=0);
 	uint SubInt(uint value, uint index = 0);
+	static uint SubVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result);
 	static sint FindLeadingBitInWord(uint x);
 	static uint SetBitInWord(uint & value, uint bit);
 	static void MulTwoWords(uint a, uint b, uint * result_high, uint * result_low);
@@ -2922,8 +3299,29 @@ public:
 };
 
 
+
+/*!
+	this specialization is needed in order to not confused the compiler "error: ISO C++ forbids zero-size array"
+	when compiling Mul3Big2() method
+*/
+template<>
+class UInt<0>
+{
+public:
+	uint table[1];
+
+	void Mul2Big(const UInt<0> &, UInt<0> &) { TTMATH_ASSERT(false) };
+	void SetZero() { TTMATH_ASSERT(false) };
+	uint AddTwoInts(uint, uint, uint) { TTMATH_ASSERT(false) return 0; };
+};
+
+
 } //namespace
 
+#if defined(_MSC_VER)
+	#pragma warning(default:4127) // conditional expression is constant
+#endif
+
 
 #include "ttmathuint_x86.h"
 #include "ttmathuint_x86_64.h"

ttmath/ttmathuint_noasm.h

@@ -1,7 +1,7 @@
 /*
  * This file is a part of TTMath Bignum Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /* 
@@ -41,6 +41,8 @@
 
 #ifdef TTMATH_NOASM
 
+#pragma message("TTMATH_NOASM")
+
 /*!
 	\file ttmathuint_noasm.h
     \brief template class UInt<uint> with methods without any assembler code
@@ -95,7 +97,7 @@ namespace ttmath
 		for(i=0 ; i<value_size ; ++i)
 			c = AddTwoWords(table[i], ss2.table[i], c, &table[i]);
 
-		TTMATH_LOG("UInt_noasm::Add")
+		TTMATH_LOG("UInt::Add")
 	
 	return c;
 	}
@@ -131,7 +133,7 @@ namespace ttmath
 		for(i=index+1 ; i<value_size && c ; ++i)
 			c = AddTwoWords(table[i], 0, c, &table[i]);
 
-		TTMATH_LOG("UInt_noasm::AddInt")
+		TTMATH_LOG("UInt::AddInt")
 	
 	return c;
 	}
@@ -184,13 +186,54 @@ namespace ttmath
 		for(i=index+2 ; i<value_size && c ; ++i)
 			c = AddTwoWords(table[i], 0, c, &table[i]);
 
-		TTMATH_LOG("UInt64::AddTwoInts")
+		TTMATH_LOG("UInt::AddTwoInts")
 	
 	return c;
 	}
 
 
 
+	/*!
+		this static method addes one vector to the other
+		'ss1' is larger in size or equal to 'ss2'
+
+		ss1 points to the first (larger) vector
+		ss2 points to the second vector
+		ss1_size - size of the ss1 (and size of the result too)
+		ss2_size - size of the ss2
+		result - is the result vector (which has size the same as ss1: ss1_size)
+
+		Example:  ss1_size is 5, ss2_size is 3
+		ss1:      ss2:   result (output):
+		  5        1         5+1
+		  4        3         4+3
+		  2        7         2+7
+		  6                  6
+		  9                  9
+	  of course the carry is propagated and will be returned from the last item
+	  (this method is used by the Karatsuba multiplication algorithm)
+	*/
+	template<uint value_size>
+	uint UInt<value_size>::AddVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result)
+	{
+	uint i, c = 0;
+
+		TTMATH_ASSERT( ss1_size >= ss2_size )
+		
+		for(i=0 ; i<ss2_size ; ++i)
+			c = AddTwoWords(ss1[i], ss2[i], c, &result[i]);
+
+		for( ; i<ss1_size ; ++i)
+			c = AddTwoWords(ss1[i], 0, c, &result[i]);
+
+		TTMATH_LOG("UInt::AddVector")
+
+	return c;
+	}
+
+
+
+
 	template<uint value_size>
 	uint UInt<value_size>::SubTwoWords(uint a, uint b, uint carry, uint * result)
 	{
@@ -232,7 +275,7 @@ namespace ttmath
 		for(i=0 ; i<value_size ; ++i)
 			c = SubTwoWords(table[i], ss2.table[i], c, &table[i]);
 
-		TTMATH_LOG("UInt_noasm::Sub")
+		TTMATH_LOG("UInt::Sub")
 
 	return c;
 	}
@@ -270,12 +313,51 @@ namespace ttmath
 		for(i=index+1 ; i<value_size && c ; ++i)
 			c = SubTwoWords(table[i], 0, c, &table[i]);
 
-		TTMATH_LOG("UInt_noasm::SubInt")
+		TTMATH_LOG("UInt::SubInt")
 	
 	return c;
 	}
 
 
+	/*!
+		this static method subtractes one vector from the other
+		'ss1' is larger in size or equal to 'ss2'
+
+		ss1 points to the first (larger) vector
+		ss2 points to the second vector
+		ss1_size - size of the ss1 (and size of the result too)
+		ss2_size - size of the ss2
+		result - is the result vector (which has size the same as ss1: ss1_size)
+
+		Example:  ss1_size is 5, ss2_size is 3
+		ss1:      ss2:   result (output):
+		  5        1         5-1
+		  4        3         4-3
+		  2        7         2-7
+		  6                  6-1  (the borrow from previous item)
+		  9                  9
+		                 return (carry): 0
+	  of course the carry (borrow) is propagated and will be returned from the last item
+	  (this method is used by the Karatsuba multiplication algorithm)
+	*/
+	template<uint value_size>
+	uint UInt<value_size>::SubVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result)
+	{
+	uint i, c = 0;
+
+		TTMATH_ASSERT( ss1_size >= ss2_size )
+		
+		for(i=0 ; i<ss2_size ; ++i)
+			c = SubTwoWords(ss1[i], ss2[i], c, &result[i]);
+
+		for( ; i<ss1_size ; ++i)
+			c = SubTwoWords(ss1[i], 0, c, &result[i]);
+
+		TTMATH_LOG("UInt::SubVector")
+
+	return c;
+	}
+
 
 
 	/*!
@@ -305,7 +387,7 @@ namespace ttmath
 			c        = new_c;
 		}
 
-		TTMATH_LOG("UInt64::Rcl2_one")
+		TTMATH_LOG("UInt::Rcl2_one")
 
 	return c;
 	}
@@ -344,7 +426,7 @@ namespace ttmath
 			c        = new_c;
 		}
 
-		TTMATH_LOG("UInt64::Rcr2_one")
+		TTMATH_LOG("UInt::Rcr2_one")
 
 	return c;
 	}
@@ -421,7 +503,7 @@ namespace ttmath
 			c        = new_c;
 		}
 
-		TTMATH_LOG("UInt64::Rcr2")
+		TTMATH_LOG("UInt::Rcr2")
 
 	return (c & TTMATH_UINT_HIGHEST_BIT) ? 1 : 0;
 	}

ttmath/ttmathuint_x86_64.h

@@ -1,7 +1,7 @@
 /*
  * This file is a part of TTMath Bignum Library
  * and is distributed under the (new) BSD licence.
- * Author: Tomasz Sowa <t.sowa@slimaczek.pl>
+ * Author: Tomasz Sowa <t.sowa@ttmath.org>
  */
 
 /*
@@ -39,11 +39,10 @@
 #ifndef headerfilettmathuint_x86_64
 #define headerfilettmathuint_x86_64
 
-
 #ifndef TTMATH_NOASM
 #ifdef TTMATH_PLATFORM64
 
-
+#pragma message("TTMATH_ASM64")
 /*!
 	\file ttmathuint_x86_64.h
     \brief template class UInt<uint> with assembler code for 64bit x86_64 processors
@@ -51,10 +50,31 @@
 	this file is included at the end of ttmathuint.h
 */
 
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+
 
 namespace ttmath
 {
 
+	#if defined(_M_X64)
+		#include <intrin.h>
+
+		extern "C"
+			{
+			uint	__fastcall	adc_x64(uint* p1, const uint* p2, uint nSize, uint c);
+			uint	__fastcall	addindexed_x64(uint* p1, uint nSize, uint nPos, uint nValue);
+			uint	__fastcall	addindexed2_x64(uint* p1, uint nSize, uint nPos, uint nValue1, uint nValue2);
+			uint	__fastcall	sbb_x64(uint* p1, const uint* p2, uint nSize, uint c);
+			uint	__fastcall	subindexed_x64(uint* p1, uint nSize, uint nPos, uint nValue);
+			uint	__fastcall	rcl_x64(uint* p1, uint nSize, uint nLowestBit);
+			uint	__fastcall	rcr_x64(uint* p1, uint nSize, uint nLowestBit);
+			uint	__fastcall	div_x64(uint* pnValHi, uint* pnValLo, uint nDiv);
+			uint	__fastcall	rcl2_x64(uint* p1, uint nSize, uint nBits, uint c);
+			uint	__fastcall	rcr2_x64(uint* p1, uint nSize, uint nBits, uint c);
+			};
+	#endif
+
 	/*!
 	*
 	*	basic mathematic functions
@@ -78,23 +98,27 @@ namespace ttmath
 	uint b = value_size;
 	uint * p1 = table;
 	const uint * p2 = ss2.table;
-	uint dummy, dummy2;
 
 		// we don't have to use TTMATH_REFERENCE_ASSERT here
 		// this algorithm doesn't require it
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = adc_x64(p1,p2,b,c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+			uint dummy, dummy2;
 			/*
 				this part should be compiled with gcc
 			*/
 			__asm__ __volatile__(
 
 				"xorq %%rdx, %%rdx				\n"
-				"neg %%rax						\n"     // CF=1 if rax!=0 , CF=0 if rax==0
+				"negq %%rax						\n"     // CF=1 if rax!=0 , CF=0 if rax==0
 
 			"1:									\n"
 				"movq (%%rsi,%%rdx,8), %%rax	\n"
@@ -112,7 +136,7 @@ namespace ttmath
 
 		#endif
 
-		TTMATH_LOG("UInt64::Add")
+		TTMATH_LOG("UInt::Add")
 
 	return c;
 	}
@@ -145,15 +169,19 @@ namespace ttmath
 	uint b = value_size;
 	uint * p1 = table;
 	uint c;
-	uint dummy, dummy2;
 
 		TTMATH_ASSERT( index < value_size )
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = addindexed_x64(p1,b,index,value);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+			uint dummy, dummy2;
 
 			__asm__ __volatile__(
 
@@ -173,12 +201,12 @@ namespace ttmath
 				"movzx %%al, %%rdx				\n"
 
 				: "=d" (c), "=a" (dummy), "=c" (dummy2)
-				: "a" (value), "c" (b), "0" (index), "b" (p1)
+				: "0" (index), "1" (value),  "2" (b), "b" (p1)
 				: "cc", "memory" );
 
 		#endif
 
-		TTMATH_LOG("UInt64::AddInt")
+		TTMATH_LOG("UInt::AddInt")
 
 	return c;
 	}
@@ -219,19 +247,55 @@ namespace ttmath
 	*/
 	template<uint value_size>
 	uint UInt<value_size>::AddTwoInts(uint x2, uint x1, uint index)
+	#if 0
+	{
+	uint i, c;
+
+		TTMATH_ASSERT( index < value_size )
+
+		printf("add %Id + %Id\n",x1,x2);
+		for(int i=index ; i<value_size ; ++i)
+			printf("%d: %Id\n",i,table[i]);
+
+		c = AddTwoWords(table[index],   x1, 0, &table[index]);
+		c = AddTwoWords(table[index+1], x2, c, &table[index+1]);
+
+		for(i=index+2 ; i<value_size && c ; ++i)
+			c = AddTwoWords(table[i], 0, c, &table[i]);
+		for(i=index ; i<value_size ; ++i)
+			printf("%d: %Id\n",i,table[i]);
+		printf(" -> %d\n",c);
+
+		TTMATH_LOG("UInt::AddTwoInts")
+
+	return c;
+	}
+	#else
 	{
 	uint b = value_size;
 	uint * p1 = table;
 	uint c;
-	uint dummy, dummy2;
 
 		TTMATH_ASSERT( index < value_size - 1 )
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				//printf("add %Id + %Id\n",x1,x2);
+				//for(int i=index ; i<value_size ; ++i)
+				//	printf("%d: %Id\n",i,table[i]);
+				//if (table[0] == 1265784741359897913) DebugBreak();
+				c = addindexed2_x64(p1,b,index,x1,x2);
+				//for(int i=index ; i<value_size ; ++i)
+				//	printf("%d: %Id\n",i,table[i]);
+				//printf(" -> %d\n",c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+			uint dummy, dummy2;
+
 			__asm__ __volatile__(
 
 				"subq %%rdx, %%rcx 				\n"
@@ -254,15 +318,16 @@ namespace ttmath
 				"movzx %%al, %%rax				\n"
 
 				: "=a" (c), "=c" (dummy), "=d" (dummy2)
-				: "1" (b), "2" (index), "b" (p1), "S" (x1), "0" (x2)
+				: "0" (x2), "1" (b),      "2" (index), "b" (p1), "S" (x1)
 				: "cc", "memory" );
 
 		#endif
 
-		TTMATH_LOG("UInt64::AddTwoInts")
+		TTMATH_LOG("UInt::AddTwoInts")
 
 	return c;
 	}
+	#endif
 
 
 
@@ -284,20 +349,25 @@ namespace ttmath
 	uint b = value_size;
 	uint * p1 = table;
 	const uint * p2 = ss2.table;
-	uint dummy, dummy2;
 
 		// we don't have to use TTMATH_REFERENCE_ASSERT here
 		// this algorithm doesn't require it
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = sbb_x64(p1,p2,b,c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+			uint dummy, dummy2;
+
 			__asm__  __volatile__(
 
 				"xorq %%rdx, %%rdx				\n"
-				"neg %%rax						\n"     // CF=1 if rax!=0 , CF=0 if rax==0
+				"negq %%rax						\n"     // CF=1 if rax!=0 , CF=0 if rax==0
 
 			"1:									\n"
 				"movq (%%rsi,%%rdx,8), %%rax	\n"
@@ -316,7 +386,7 @@ namespace ttmath
 
 		#endif
 
-		TTMATH_LOG("UInt64::Sub")
+		TTMATH_LOG("UInt::Sub")
 
 	return c;
 	}
@@ -347,15 +417,20 @@ namespace ttmath
 	uint b = value_size;
 	uint * p1 = table;
 	uint c;
-	uint dummy, dummy2;
 
 		TTMATH_ASSERT( index < value_size )
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = subindexed_x64(p1,b,index,value);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+			uint dummy, dummy2;
+
 			__asm__ __volatile__(
 
 				"subq %%rdx, %%rcx 				\n"
@@ -374,7 +449,7 @@ namespace ttmath
 				"movzx %%al, %%rdx				\n"
 
 				: "=d" (c),    "=a" (dummy), "=c" (dummy2)
-				: "1" (value), "2" (b), "0" (index), "b" (p1)
+				: "0" (index), "1" (value),  "2" (b), "b" (p1)
 				: "cc", "memory" );
 
 		#endif
@@ -404,17 +479,22 @@ namespace ttmath
 	{
 	sint b = value_size;
 	uint * p1 = table;
-	uint dummy, dummy2;
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = rcl_x64(p1,b,c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+		uint dummy, dummy2;
+
 		__asm__  __volatile__(
 
 			"xorq %%rdx, %%rdx			\n"   // rdx=0
-			"neg %%rax					\n"   // CF=1 if rax!=0 , CF=0 if rax==0
+			"negq %%rax					\n"   // CF=1 if rax!=0 , CF=0 if rax==0
 
 		"1:								\n"
 			"rclq $1, (%%rbx, %%rdx, 8)	\n"
@@ -426,12 +506,12 @@ namespace ttmath
 			"adcq %%rcx, %%rcx			\n"
 
 			: "=c" (c), "=a" (dummy), "=d" (dummy2)
-			: "1" (c), "0" (b), "b" (p1)
+			: "0" (b),  "1" (c), "b" (p1)
 			: "cc", "memory" );
 
 		#endif
 
-		TTMATH_LOG("UInt64::Rcl2_one")
+		TTMATH_LOG("UInt::Rcl2_one")
 
 	return c;
 	}
@@ -456,16 +536,21 @@ namespace ttmath
 	{
 	sint b = value_size;
 	uint * p1 = table;
-	uint dummy;
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = rcr_x64(p1,b,c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+		uint dummy;
+
 		__asm__  __volatile__(
 
-			"neg %%rax						\n"   // CF=1 if rax!=0 , CF=0 if rax==0
+			"negq %%rax						\n"   // CF=1 if rax!=0 , CF=0 if rax==0
 
 		"1:									\n"
 			"rcrq $1, -8(%%rbx, %%rcx, 8)	\n"
@@ -476,12 +561,12 @@ namespace ttmath
 			"adcq %%rcx, %%rcx				\n"
 
 			: "=c" (c), "=a" (dummy)
-			: "1" (c), "0" (b), "b" (p1)
+			: "0" (b),  "1" (c), "b" (p1)
 			: "cc", "memory" );
 
 		#endif
 
-		TTMATH_LOG("UInt64::Rcr2_one")
+		TTMATH_LOG("UInt::Rcr2_one")
 
 	return c;
 	}
@@ -509,13 +594,18 @@ namespace ttmath
 
 	uint b = value_size;
 	uint * p1 = table;
-	uint dummy, dummy2, dummy3;
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = rcl2_x64(p1,b,bits,c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
+		uint dummy, dummy2, dummy3;
+
 		__asm__  __volatile__(
 
 			"movq %%rcx, %%rsi				\n"
@@ -553,7 +643,7 @@ namespace ttmath
 
 		#endif
 
-		TTMATH_LOG("UInt64::Rcl2")
+		TTMATH_LOG("UInt::Rcl2")
 
 	return c;
 	}
@@ -580,14 +670,19 @@ namespace ttmath
 
 	sint b = value_size;
 	uint * p1 = table;
-	uint dummy, dummy2, dummy3;
 
 		#ifndef __GNUC__
+			#if defined(_M_X64)
+				c = rcr2_x64(p1,b,bits,c);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 
 		#ifdef __GNUC__
+			uint dummy, dummy2, dummy3;
+
 			__asm__  __volatile__(
 
 			"movq %%rcx, %%rsi				\n"
@@ -628,7 +723,7 @@ namespace ttmath
 
 		#endif
 
-		TTMATH_LOG("UInt64::Rcr2")
+		TTMATH_LOG("UInt::Rcr2")
 
 	return c;
 	}
@@ -646,19 +741,29 @@ namespace ttmath
 	register sint result;
 
 		#ifndef __GNUC__
+			#if defined(_MSC_VER)
+				unsigned long	nIndex(0);
+
+				if (_BitScanReverse64(&nIndex,x) == 0)
+					result = -1;
+				  else
+					result = nIndex;
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
-			__asm__  __volatile__(
+		uint dummy;
 
-			"bsrq %1, %0		\n"
-			"jnz 1f				\n"
-			"movq $-1, %0		\n"
-			"1:					\n"
+				__asm__ (
 
-			: "=R" (result)
-			: "R" (x)
+				"movq $-1, %1          \n"
+				"bsrq %2, %0           \n"
+				"cmovz %1, %0          \n"
+
+				: "=r" (result), "=&r" (dummy)
+				: "r" (x)
 				: "cc" );
 
 		#endif
@@ -691,14 +796,22 @@ namespace ttmath
 
 
 		#ifndef __GNUC__
+			#if defined(_MSC_VER)
+				#if defined(TTMATH_PLATFORM64)
+					old_bit = _bittestandset64((__int64*)&value,bit) != 0;
+				  #else
+					old_bit = _bittestandset((long*)&value,bit) != 0;
+				#endif
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
-			__asm__  __volatile__(
+
+			__asm__ (
 
 			"btsq %%rbx, %%rax		\n"
-
 			"setc %%bl				\n"
 			"movzx %%bl, %%rbx		\n"
 			
@@ -742,16 +855,20 @@ namespace ttmath
 		this has no effect in visual studio but it's usefull when
 		using gcc and options like -O
 	*/
-	register uint result1_;
-	register uint result2_;
+	uint result1_;
+	uint result2_;
 
 		#ifndef __GNUC__
+			#if defined(_MSC_VER)
+				result1_ = _umul128(a,b,&result2_);
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
 
-		__asm__ __volatile__(
+		__asm__ (
 		
 			"mulq %%rdx			\n"
 
@@ -776,6 +893,7 @@ namespace ttmath
 	 *
 	*/
 
+	#ifndef __GNUC__
 
 	/*!
 		this method calculates 64bits word a:b / 32bits c (a higher, b lower word)
@@ -793,8 +911,8 @@ namespace ttmath
 	template<uint value_size>
 	void UInt<value_size>::DivTwoWords(uint a,uint b, uint c, uint * r, uint * rest)
 	{
-		register uint r_;
-		register uint rest_;
+		uint r_;
+		uint rest_;
 		/*
 			these variables have similar meaning like those in
 			the multiplication algorithm MulTwoWords
@@ -803,12 +921,18 @@ namespace ttmath
 		TTMATH_ASSERT( c != 0 )
 
 		#ifndef __GNUC__
+			#if defined(_MSC_VER)
+				div_x64(&a,&b,c);
+				r_ = a;
+				rest_ = b;
+			#else
 				#error "another compiler than GCC is currently not supported in 64bit mode"
 			#endif
+		#endif
 
 		#ifdef __GNUC__
 		
-			__asm__ __volatile__(
+			__asm__ (
 
 			"divq %%rcx				\n"
 
@@ -823,6 +947,132 @@ namespace ttmath
 		*rest = rest_;
 	}
 
+	template<uint value_size>
+	uint UInt<value_size>::AddTwoWords(uint a, uint b, uint carry, uint * result)
+	{
+	uint temp;
+
+		if( carry == 0 )
+		{
+			temp = a + b;
+
+			if( temp < a )
+				carry = 1;
+		}
+		else
+		{
+			carry = 1;
+			temp  = a + b + carry;
+
+			if( temp > a ) // !(temp<=a)
+				carry = 0;
+		}
+
+		*result = temp;
+
+	return carry;
+	}
+
+	template<uint value_size>
+	uint UInt<value_size>::SubTwoWords(uint a, uint b, uint carry, uint * result)
+	{
+		if( carry == 0 )
+		{
+			*result = a - b;
+
+			if( a < b )
+				carry = 1;
+		}
+		else
+		{
+			carry   = 1;
+			*result = a - b - carry;
+
+			if( a > b ) // !(a <= b )
+				carry = 0;
+		}
+
+	return carry;
+	}
+
+	/*!
+		this static method addes one vector to the other
+		'ss1' is larger in size or equal to 'ss2'
+
+		ss1 points to the first (larger) vector
+		ss2 points to the second vector
+		ss1_size - size of the ss1 (and size of the result too)
+		ss2_size - size of the ss2
+		result - is the result vector (which has size the same as ss1: ss1_size)
+
+		Example:  ss1_size is 5, ss2_size is 3
+		ss1:      ss2:   result (output):
+		  5        1         5+1
+		  4        3         4+3
+		  2        7         2+7
+		  6                  6
+		  9                  9
+	  of course the carry is propagated and will be returned from the last item
+	  (this method is used by the Karatsuba multiplication algorithm)
+	*/
+	template<uint value_size>
+	uint UInt<value_size>::AddVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result)
+	{
+	uint i, c = 0;
+
+		TTMATH_ASSERT( ss1_size >= ss2_size )
+
+		for(i=0 ; i<ss2_size ; ++i)
+			c = AddTwoWords(ss1[i], ss2[i], c, &result[i]);
+
+		for( ; i<ss1_size ; ++i)
+			c = AddTwoWords(ss1[i], 0, c, &result[i]);
+
+		TTMATH_LOG("UInt::AddVector")
+
+	return c;
+	}
+
+	/*!
+		this static method subtractes one vector from the other
+		'ss1' is larger in size or equal to 'ss2'
+
+		ss1 points to the first (larger) vector
+		ss2 points to the second vector
+		ss1_size - size of the ss1 (and size of the result too)
+		ss2_size - size of the ss2
+		result - is the result vector (which has size the same as ss1: ss1_size)
+
+		Example:  ss1_size is 5, ss2_size is 3
+		ss1:      ss2:   result (output):
+		  5        1         5-1
+		  4        3         4-3
+		  2        7         2-7
+		  6                  6-1  (the borrow from previous item)
+		  9                  9
+		                 return (carry): 0
+	  of course the carry (borrow) is propagated and will be returned from the last item
+	  (this method is used by the Karatsuba multiplication algorithm)
+	*/
+	template<uint value_size>
+	uint UInt<value_size>::SubVector(const uint * ss1, const uint * ss2, uint ss1_size, uint ss2_size, uint * result)
+	{
+	uint i, c = 0;
+
+		TTMATH_ASSERT( ss1_size >= ss2_size )
+
+		for(i=0 ; i<ss2_size ; ++i)
+			c = SubTwoWords(ss1[i], ss2[i], c, &result[i]);
+
+		for( ; i<ss1_size ; ++i)
+			c = SubTwoWords(ss1[i], 0, c, &result[i]);
+
+		TTMATH_LOG("UInt::SubVector")
+
+	return c;
+	}
+
+	#endif // #ifndef __GNUC__
 
 } //namespace
 

ttmath/ttmathuint_x86_amd64_msvc.asm

@@ -0,0 +1,386 @@
+PUBLIC	adc_x64
+PUBLIC	addindexed_x64
+PUBLIC	addindexed2_x64
+
+PUBLIC	sbb_x64
+PUBLIC	subindexed_x64
+
+PUBLIC	rcl_x64
+PUBLIC	rcr_x64
+
+PUBLIC	rcl2_x64
+PUBLIC	rcr2_x64
+
+PUBLIC	div_x64
+
+;
+;	"rax, rcx, rdx, r8-r11 are volatile."
+;	"rbx, rbp, rdi, rsi, r12-r15 are nonvolatile."
+;
+
+.CODE
+
+        ALIGN       8
+
+;----------------------------------------
+
+adc_x64				PROC
+        ; rcx = p1
+        ; rdx = p2
+        ; r8 = nSize
+        ; r9 = nCarry
+
+        xor		rax, rax
+        xor		r11, r11
+        sub		rax, r9		; sets CARRY if r9 != 0
+
+		ALIGN 16
+ loop1:
+		mov		rax,qword ptr [rdx + r11 * 8]
+		adc		qword ptr [rcx + r11 * 8], rax
+		lea		r11, [r11+1]
+		dec		r8
+		jnz		loop1
+
+		setc	al
+		movzx	rax, al
+
+		ret
+
+adc_x64				ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+addindexed_x64	PROC
+
+        ; rcx = p1
+        ; rdx = nSize
+        ; r8 = nPos
+        ; r9 = nValue
+
+		xor		rax, rax			; rax = result
+		sub		rdx, r8				; rdx = remaining count of uints
+
+		add		qword ptr [rcx + r8 * 8], r9
+		jc		next1
+
+		ret
+
+next1:
+		mov		r9, 1
+
+		ALIGN 16
+loop1:
+		dec		rdx
+		jz		done_with_cy
+		lea		r8, [r8+1]
+		add		qword ptr [rcx + r8 * 8], r9
+		jc		loop1
+
+		ret
+
+done_with_cy:
+		lea		rax, [rax+1]		; rax = 1
+
+		ret
+
+addindexed_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+addindexed2_x64	PROC
+
+        ; rcx = p1 (pointer)
+        ; rdx = b  (value size)
+        ; r8 = nPos
+        ; r9 = nValue1
+        ; [esp+0x28] = nValue2
+
+		xor		rax, rax			; return value
+		mov		r11, rcx			; table
+		sub		rdx, r8				; rdx = remaining count of uints
+		mov		r10, [esp+028h]		; r10 = nValue2
+
+		add		qword ptr [r11 + r8 * 8], r9
+		lea		r8, [r8+1]
+		lea		rdx, [rdx-1]
+		adc		qword ptr [r11 + r8 * 8], r10
+		jc		next
+		ret
+
+		ALIGN 16
+loop1:
+		lea		r8, [r8+1]
+		add		qword ptr [r11 + r8 * 8], 1
+		jc		next
+		ret
+
+next:
+		dec		rdx					; does not modify CY too...
+		jnz		loop1
+		lea		rax, [rax+1]
+		ret
+
+addindexed2_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+sbb_x64				PROC
+
+        ; rcx = p1
+        ; rdx = p2
+        ; r8 = nCount
+        ; r9 = nCarry
+
+        xor		rax, rax
+        xor		r11, r11
+        sub		rax, r9				; sets CARRY if r9 != 0
+
+		ALIGN 16
+ loop1:
+		mov		rax,qword ptr [rdx + r11 * 8]
+		sbb		qword ptr [rcx + r11 * 8], rax
+		lea		r11, [r11+1]
+		dec		r8
+		jnz		loop1
+
+		setc	al
+		movzx	rax, al
+
+		ret
+
+sbb_x64				ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+subindexed_x64	PROC
+        ; rcx = p1
+        ; rdx = nSize
+        ; r8 = nPos
+        ; r9 = nValue
+
+		sub		rdx, r8				; rdx = remaining count of uints
+
+		ALIGN 16
+loop1:
+		sub		qword ptr [rcx + r8 * 8], r9
+		jnc		done
+
+		lea		r8, [r8+1]
+		mov		r9, 1
+		dec		rdx
+		jnz		loop1
+		jc		return_1	; most of the times, there will be NO carry (I hope)
+
+done:
+		xor		rax, rax
+		ret
+
+  return_1:
+		mov		rax, 1
+		ret
+
+subindexed_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+rcl_x64	PROC
+        ; rcx = p1
+        ; rdx = b
+        ; r8 = nLowestBit
+
+		mov		r11, rcx			; table
+		xor		r10, r10
+		neg		r8					; CY set if r8 <> 0
+
+		ALIGN 16
+loop1:
+		rcl		qword ptr [r11 + r10 * 8], 1
+		lea		r10, [r10+1]
+		dec		rdx
+		jnz		loop1
+
+		setc	al
+		movzx	rax, al
+
+        ret
+
+rcl_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+rcr_x64	PROC
+        ; rcx = p1
+        ; rdx = nSize
+        ; r8 = nLowestBit
+
+		xor		r10, r10
+		neg		r8					; CY set if r8 <> 0
+
+		ALIGN 16
+loop1:
+		rcr		qword ptr -8[rcx + rdx * 8], 1
+		dec		rdx
+		jnz		loop1
+
+		setc	al
+		movzx	rax, al
+
+        ret
+
+rcr_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+div_x64	PROC
+
+        ; rcx = &Hi
+        ; rdx = &Lo
+        ; r8 = nDiv
+
+        mov		r11, rcx
+        mov		r10, rdx
+
+        mov		rdx, qword ptr [r11]
+        mov		rax, qword ptr [r10]
+        div		r8
+        mov		qword ptr [r10], rdx ; remainder
+        mov		qword ptr [r11], rax ; value
+
+        ret
+
+div_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+rcl2_x64	PROC
+        ; rcx = p1
+        ; rdx = nSize
+        ; r8 = bits
+        ; r9 = c
+
+        push	rbx
+
+        mov		r10, rcx	; r10 = p1
+        xor		rax, rax
+
+        mov		rcx, 64
+        sub		rcx, r8
+
+        mov		r11, -1
+        shr		r11, cl		; r11 = mask
+
+		mov		rcx, r8		; rcx = count of bits
+
+		mov		rbx, rax	; rbx = old value = 0
+		or		r9, r9
+		cmovnz	rbx, r11	; if (c) then old value = mask
+
+        mov		r9, rax		; r9 = index (0..nSize-1)
+
+		ALIGN 16
+loop1:
+		rol		qword ptr [r10+r9*8], cl
+		mov		rax, qword ptr [r10+r9*8]
+		and		rax, r11
+		xor		qword ptr [r10+r9*8], rax
+		or		qword ptr [r10+r9*8], rbx
+		mov		rbx, rax
+
+		lea		r9, [r9+1]
+		dec		rdx
+
+		jnz		loop1
+
+		and		rax, 1
+		pop		rbx
+        ret
+
+rcl2_x64	ENDP
+
+;----------------------------------------
+
+        ALIGN       8
+
+;----------------------------------------
+
+rcr2_x64	PROC
+        ; rcx = p1
+        ; rdx = nSize
+        ; r8 = bits
+        ; r9 = c
+
+        push	rbx
+        mov		r10, rcx	; r10 = p1
+        xor		rax, rax
+
+        mov		rcx, 64
+        sub		rcx, r8
+
+        mov		r11, -1
+        shl		r11, cl		; r11 = mask
+
+		mov		rcx, r8		; rcx = count of bits
+
+		mov		rbx, rax	; rbx = old value = 0
+		or		r9, r9
+		cmovnz	rbx, r11	; if (c) then old value = mask
+
+        mov		r9, rdx		; r9 = index (0..nSize-1)
+		lea		r9, [r9-1]
+
+		ALIGN 16
+loop1:
+		ror		qword ptr [r10+r9*8], cl
+		mov		rax, qword ptr [r10+r9*8]
+		and		rax, r11
+		xor		qword ptr [r10+r9*8], rax
+		or		qword ptr [r10+r9*8], rbx
+		mov		rbx, rax
+
+		lea		r9, [r9-1]
+		dec		rdx
+
+		jnz		loop1
+
+		rol		rax, 1
+		and		rax, 1
+		pop		rbx
+
+        ret
+
+rcr2_x64	ENDP
+
+END