TweetFollow Us on Twitter

June 96 - 64-Bit Integer Math on 680x0 Machines

64-Bit Integer Math on 680x0 Machines


When an application has to perform integer arithmetic with numbers larger than 32 bits on both the PowerPC and 680x0 platforms, you could use the floating-point types of the SANE and PowerPC Numerics libraries. But if all you really need is a larger integer, a better choice is to use the existing 64-bit math routines available on the PowerPC platform and write an equivalent library for the 680x0 Macintosh. This article presents just such a library.

Developers of PowerPC applications that need 64-bit math can simply call the various "wide" Toolbox routines. These routines perform addition, subtraction, multiplication, division, square root, and a few other operations. On the 680x0-based Macintosh, some of these same routines are available in QuickDraw GX. But if you can't assume your customers have QuickDraw GX installed, you need a library that supports 64-bit math.

The Wide library presented in this article works on both platforms and has exactly the same interface and types as the wide routines in the Toolbox on PowerPC machines. The library also provides some new routines such as 32-bit to 64-bit add and subtract and a 64-bit-to-string conversion function. The library is included on this issue's CD, along with its source code.

All the routines use the 64-bit data type defined in the header file Types.h, which is the standard type used for signed 64-bit integers on both the PowerPC and 680x0 Macintosh:

struct wide {
   Sint32   hi;   /* upper 32 bits (signed) */
   Uint32   lo;   /* lower 32 bits (unsigned) */
typedef struct wide wide, *WidePtr;


Before plunging into the Wide library, let's see what 64-bit math routines I'll be talking about. First, I'll introduce those that are available on PowerPC machines, then those you'll find on a 680x0 Macintosh with QuickDraw GX, and finally the routines in the Wide library.


In the header file FixMath.h, the routines listed in Table 1 are defined for 64-bit math on the PowerPC platform.


On 680x0 machines that have QuickDraw GX installed, all the wide routines for the PowerPC platform listed in Table 1 are available, with the exception of WideBitShift. The QuickDraw GX header file GXTypes.h defines the wide routine types and function prototypes in exactly the same way that the header file FixMath.h does for PowerPC machines.

In addition, QuickDraw GX on 680x0 machines has a routine that the PowerPC platform doesn't have: WideScale. This function returns the bit number of the highest-order nonzero bit in a 64-bit number. The Wide library implements this function on the PowerPC platform.


The Wide 64-bit integer math library on this issue's CD provides all the wide routines that are available on PowerPC machines and on 680x0 machines with QuickDraw GX, plus a few extras. The extra routines, which are available on both the PowerPC and 680x0 platforms, are listed in Table 2.

WideAssign32, WideAdd32, WideSubtract32. These routines are self-explanatory.

WideToDecStr. This routine converts a signed 64-bit integer to the SANE string type decimal, which is also defined by the PowerPC Numerics library. This string structure is a good intermediate format for final conversion to a string format of your choosing.

Since WideToDecStr calls the SANE library to generate the string, SANE must be linked with your 680x0 application. The SANE library is included with all the major development systems.

To convert the string returned by WideToDecStr to a Pascal string, call the SANE routine dec2str.

    If you want to generate a localized number, take a look at the article "International Number Formatting" in develop Issue 16. You could call the LocalizeNumberString function from that article after converting the output of WideToDecStr to a Pascal string, or you could modify LocalizeNumberString to accept the output of WideToDecStr.*
p> WideInit. The library is self-initializing; the first time you call any wide routine, WideInit is also called. If the execution speed of your first runtime call to a wide routine is important, you have the option of calling WideInit during your application's startup to avoid that overhead.

The purpose of WideInit is to determine what processor is being used, or emulated; it calls Gestalt to make this determination. If your Macintosh has a 68020-68040 CPU (68020, 68030, or 68040), the library will use the 64-bit multiply and divide instructions available on that processor; otherwise, the library will have to call software subroutines for those operations. On 68000 machines, such as the Macintosh Plus and SE, the processor's multiply instruction is limited to 32 bits and the library has no choice but to use the slower algorithmic approach for multiplication and division.


The library can be compiled on the 680x0 and PowerPC platforms using either the Metrowerks CodeWarrior or Symantec C development system. The library tests which development system is compiling it and, if it's not CodeWarrior or Symantec, the preprocessor displays an error message saying the library needs to be ported to your environment. This is necessary because there's some inline assembly language in the source file, as discussed later in this section, and different C compilers handle assembly language differently.

While the interface routines to our 64-bit library are the same on the PowerPC and 680x0 machines, when you compile the library a different subset of routines is linked in, depending on your environment:

  • If you build the library for a 680x0 machine without QuickDraw GX headers, all the Wide library routines are defined.

  • If you build the library for a 680x0 machine and include the QuickDraw GX header file GXTypes.h or GXMath.h before the Wide library's Wide.h header file, the extra routines and the WideBitShift routine are defined. The other wide routines are already available via the QuickDraw GX traps.

  • When you compile for the PowerPC platform, only the five extra routines (WideAssign32, WideAdd32, WideSubtract32, WideToDecStr, and WideInit) are defined in the library. All the other wide routines already exist in the PowerPC Toolbox. Additionally, if GXTypes.h or GXMath.h isn't included, WideScale is defined.
Table 3 summarizes where the wide routines can be found on the different platforms.

Note that the Wide library decides at compile time which routines to use. When QuickDraw GX header files are not included, the Wide library routines are called. If your application needs to make a runtime decision about whether to use QuickDraw GX, you'll need to make some changes to the library. One solution is to rename the Wide library routines and remove the conditional compilation tests for QuickDraw GX from the source. Then at run time you can decide which version to call -- the QuickDraw GX routines if they're available, or the internal Wide library routinesif not.


The Wide library was compiled with version 2.1 of Apple's universal headers. The latest headers are available on this issue's CD. You should make sure you have a recent version of these headers, because the library uses the constant GENERATING68K. If the header file ConditionalMacros.h doesn't contain this constant, your version of the universal headers is too old.


Some of the routines in the library are written in assembly language to take advantage of the 64-bit multiply and divide instructions on 68020-68040 machines, because on these machines the C language will use only 32-bit multiply and divide instructions. On PowerPC machines, the Wide library doesn't need assembly language because the 64-bit multiply and divide routines are provided by the Toolbox.

The library's source file Wide.c contains both C and assembly language. It has been successfully compiled by Symantec C 7.0.4 and CodeWarrior 7. If you want to compile the library on any other development system, you may have to do a little work porting it. Most of the changes will be confined to the conditional compilation statements at the beginning of Wide.c where the differences in SANE types and inline assembly language are handled.


Now let's look at a couple of the more interesting routines in the Wide library to see how they work. See the source code on the CD for full implementations of all the routines.


WideMultiply (Listing 1) performs a 32-by-32-bit multiply and produces a 64-bit result. The first and second parameters are the two signed 32-bit integers to be multiplied together. The return value is a pointer to the 64-bit result that's also returned via the third parameter.

Listing 1. The multiply routine

wide *WideMultiply (
   long multiplicand,   /* in: first value to multiply */
   long multiplier,      /* in: second value to multiply */
   wide *target_ptr)      /* out: 64 bits to be assigned */
   /* Initialize Wide library if not already done. */
   if (!gWide_Initialized) WideInit();

   /* If the 64-bit multiply instruction is available... */
   if (gWide_64instr) {
      /* Execute the assembly-language instruction MULS.L */
      Wide_MulS64(multiplicand, multiplier, target_ptr);
   else {
      /* Call the Toolbox to perform the multiply. */
      LongMul(multiplicand, multiplier, (Int64Bit *) target_ptr);

   return (target_ptr);
WideMultiply first tests whether the library has been initialized yet; if not, it calls WideInit. Next the routine tests whether the 64-bit multiply instruction is available on the current CPU by examining the global variable gWide_64instr (which was set by the initialization routine WideInit). If the instruction is available, WideMultiply calls the assembly-language function Wide_MulS64 to take advantage of it (as described later); otherwise, WideMultiply calls the Toolbox routine LongMul to perform the multiplication, as would be the case on 68000 machines.


The WideSquareRoot function (Listing 2) takes a 64-bit unsigned number as input and returns a 32-bit unsigned result. All possible results can be expressed in 32 bits, so overflow isn't possible.

Listing 2. The square root routine

unsigned long WideSquareRoot (
   const wide *source_ptr) /* in: value to take the square root of */
   wide            work_integer;
   Extended_80      extended_80_number;

   /* Initialize Wide library if not already done. */
   if (!gWide_Initialized) WideInit();

   /* Convert "wide" number to "extended" format. */
   Wide_ToExtended(&extended_80_number, source_ptr);

   /* If compiling with CodeWarrior, the parameter to sqrt is a
      pointer instead of a value, as defined in PowerPC Numerics. */
#ifdef __MWERKS__
   extended_80_number = sqrt(extended_80_number);

   /* Convert "extended" format to "wide" number. */
   Wide_FromExtended(&work_integer, &extended_80_number);
   /* OK to ignore work_integer.hi as it's always 0. */
   return (work_integer.lo);
For this routine I decided to let the SANE library do the work of generating the square root. The routine converts the 64-bit input number to an 80-bit floating-point number and then calls the SANE library function sqrt to calculate the square root. Finally, WideSquareRoot converts the resulting 80-bit floating-point number back to a 64-bit integer and returns the low-order half of the result.

When a 64-bit integer is converted to an 80-bit floating-point number, no loss in precision occurs. An 80-bit floating-point number is made up of three parts -- the sign (1 bit), the exponent (15 bits), and the fractional part (64 bits). As you can see, a 64-bit integer exactly fits in the fractional part.

Two differences between the CodeWarrior and Symantec development systems that show up in the Wide library's WideSquareRoot function are the 80-bit floating-point types and the parameters of the SANE library's square root function. Under CodeWarrior, the Wide library internal type Extended_80 is defined as the type extended80, and Sqrt returns the result to the same location as the input number. Under Symantec C, Extended_80 is defined as the type extended, and sqrt returns the result as a function return value.


The Wide library uses internal assembly-language routines to execute 64-bit multiply and divide instructions on machines that support those instructions. In case you're interested, here are the details.

Symantec and CodeWarrior handle the asm keyword differently, so I used some preprocessor commands (#defines) to handle the differences between the two development systems. Near the beginning of the Wide.c source file there are four #defines that differ depending on which development system you're using, as shown in Table 4.


Wide_MulS64 (Listing 3) is an internal assembly-language routine that WideMultiply calls to execute the 64-bit multiply instruction on the 68020-68040 CPUs. It starts with ASM_FUNC_HEAD, as mentioned in Table 4. The three definitions at the start of the function (MULTIPLICAND, MULTIPLIER, and OUT_PTR) are the byte offsets to the parameters. Although in Symantec C it's possible to refer to function parameters by name via A6, this isn't possible in CodeWarrior. I had to give up accessing the parameters by name and use #defines instead.

Listing 3. 64-bit multiply instruction

ASM_FUNC_HEAD static void Wide_MulS64 (
   long multiplicand,   /* in: first value to multiply */
   long multiplier,     /* in: second value to multiply */
   wide *out_ptr)       /* out: 64 bits to be assigned */
#define MULTIPLICAND     8
#define MULTIPLIER      12
#define OUT_PTR         16

      MOVE.L   MULTIPLICAND(A6),D0     //
      DC.W      0x4C2E,0x0C01,0x000C   // MULS.L multiplier(A6),D1-D0
      MOVE.L   OUT_PTR(A6),A0          //
      MOVE.L    D0,WIDE_LO(A0)         //
      MOVE.L   D1,WIDE_HI(A0)          //
To execute the 64-bit multiply instruction I had to define it with a DC.W directive that generates the desired object code. This was necessary because the Symantec C inline assembler supports only the 32-bit multiply instruction and won't recognize the 64-bit assembly opcode.


If the 64-bit divide instruction isn't available, the library calls the internal assembly-language routine Wide_DivideU (Listing 4) to perform the division using an algorithm. The algorithm is basically a binary version of the paper and pencil method of doing long division that all of us learned in school. It's a loop that executes once for each bit in the size of the divisor, which is 32 in our case. The Wide_DivideU subroutine actually handles only unsigned division, but the library function that calls it will take care of converting the input parameters to positive values and, if required, converting the result to a negative value.

Listing 4. 64-bit unsigned division algorithm

ASM_FUNC_HEAD static void Wide_DivideU (
   wide *dividend_ptr,      /* in/out: 64 bits to be divided */
   long divisor,            /* in: value to divide by */
   long *remainder_ptr)     /* out: the remainder of the division */
#define DIVIDEND_PTR      8
#define DIVISOR         12
#define REMAINDER_PTR   16

      MOVEM.L  D2-D7,-(SP)           // save work registers
      CLR.L    D0                    //
      CLR.L    D1                // D0-D1 is the quotient accumulator
      MOVE.L   DIVIDEND_PTR(A6),A0   //
      MOVE.L   WIDE_HI(A0),D2        //
      MOVE.L   WIDE_LO(A0),D3        // D2-D3 = remainder accumulator
      CLR.L    D4                    //
      MOVE.L   D2,D5                 // D5 = copy of dividend.hi
      MOVE.L   DIVISOR(A6),D6        // D6 = copy of divisor

      MOVEQ.L  #31,D7                // FOR number of bits in divisor
      LSL.L    #1,D0             // shift quotient.hi accum left once
      LSL.L    #1,D1             // shift quotient.lo accum left once
      LSL.L    #1,D4                 //
      LSL.L    #1,D3                 //
      ROXL.L   #1,D2             // shift remainder accum left once
      SUB.L    D6,D2                 // remainder -= divisor
      BCS      @div50                // If CS, remainder is negative
      BSET     #0,D1                 // quotient.lo |= 1
      BRA.S    @div77                //
      ADD.L    D6,D2                 // remainder += divisor
      BTST     D7,D5                 //
      BEQ      @div90            // If EQ, bit not set in dividend.hi
      BSET     #0,D4                  //
      CMP.L    D6,D4                  //
      BCS      @div99                 // If CS, divisor < D4
      SUB.L    D6,D4                  // D4 -= divisor
      BSET     #0,D0                  // quotient.hi |= 1
      DBF      D7,@divloop            // loop until D7 == -1
      MOVE.L   DIVIDEND_PTR(A6),A0    // output the remainder
      MOVE.L   D0,WIDE_HI(A0)         //
      MOVE.L   D1,WIDE_LO(A0)         //
      MOVE.L   REMAINDER_PTR(A6),A0   // output the remainder
      MOVE.L   D2,(A0)                //
      MOVEM.L  (SP)+,D2-D7            // restore work registers
The top of the assembly-language loop starts at the @divloop label. For each loop, the algorithm shifts the quotient and the remainder left one bit position before trying to subtract the divisor from the remainder. If the subtraction can be done, the least-significant bit in quotient.lo is set; otherwise, the subtraction is undone by the add instruction near the @div50 label. Then, if the divisor is greater than the loop bits that are accumulating in register D4, the least-significant bit in quotient.hi is set.

Notice that the first assembly-language statement in Wide_DivideU is a MOVEM.L instruction that saves on the stack all the registers that the division loop uses; the last instruction is a MOVEM.L instruction that restores these registers. Fortunately, this subroutine can place all its working variables in registers and avoid the stack for its loop, thus improving performance.


There you have it. Now 64-bit integer math can be handled with the same API on both the 680x0 and PowerPC platforms. Having the same function-level interface on these two very different processors makes life a lot easier for application programmers. Don't you wish all libraries had the same interface regardless of the CPU or system software version?

DALE SEMCHISHEN ( lives in Vancouver, British Columbia, with his wife Josephine. He works for Glenayre Technologies as a paging software developer (they make the control systems that send messages to your belt beeper). Recently, he had to accept the fact that the world is changing when his retired father started talking about his Internet provider.*

Thanks to our technical reviewers Dave Evans, Quinn "The Eskimo!", and Dave Radcliffe. Special thanks to Dave Johnson for software testing.*


Community Search:
MacTech Search:

Software Updates via MacUpdate

Shadow Blade: Reload guide - How to hack...
Shadow Blade: Reload is the kind of action-platformer that would have happily sucked up hours of your time on a console a few years back.Now, you can take it with you wherever you go, and its mobile conversion is not too shabby at all. To help you... | Read more »
Tomb of the Mask guide - How to increase...
Tomb of the Mask is a great endless arcade game from Happymagenta in which quick reflexes and a persistent attitude can go a long way toward earning a top score. Check out these tips to see if you can give yourself an edge on the leaderboards. [... | Read more »
Smooth Operator! (Games)
Smooth Operator! 1.0 Device: iOS Universal Category: Games Price: $2.99, Version: 1.0 (iTunes) Description: Smooth Operator is a weird, weird two-player kissing game. Squeeze in for 2 player fun on a single iPad, creating awkward... | Read more »
Sinless: Remastered (Games)
Sinless: Remastered 1.0 Device: iOS Universal Category: Games Price: $1.99, Version: 1.0 (iTunes) Description: | Read more »
_PRISM Guide - How to solve those puzzle...
_PRISM is a rather delightful puzzle game that’s been tailor made for touch screens. While part of the fun is figuring things out as you go along, we thought we’d offer you a helping hand at getting in the right mindset. Don’t worry about messing... | Read more »
Fractal Space (Games)
Fractal Space 1.3.1 Device: iOS Universal Category: Games Price: $.99, Version: 1.3.1 (iTunes) Description: Live the memorable experience of Fractal Space, a unique first person adventure & puzzle game by Haze Games! Will you... | Read more »
Set off on an adventure through the Cand...
Like match three puzzlers? If so, Jelly Blast, the innovative iOS and Android game which launched last year, is worth a look. Jelly Blast sees you head off on an epic adventure through the Candy Kingdom with your friends Lily, Mr. Hare, and Mr.... | Read more »
Ellipsis - Touch. Explore. Survive. (...
Ellipsis - Touch. Explore. Survive. 1.0 Device: iOS Universal Category: Games Price: $2.99, Version: 1.0 (iTunes) Description: | Read more »
Abzorb (Games)
Abzorb 1 Device: iOS Universal Category: Games Price: $2.99, Version: 1 (iTunes) Description: Abzorb is a tilt game about consuming orbs by getting as close as you can to them without touching. By tilting your mobile device, collect... | Read more »
A Short Tale (Games)
A Short Tale 1.0 Device: iOS Universal Category: Games Price: $3.99, Version: 1.0 (iTunes) Description: It’s been years since I lost Ben, so long yet it doesn’t seem long enough. I never thought I’d return here, to where everything... | Read more »

Price Scanner via

cb Hardcase – Handmade and Premium Protective...
Baden-Baden, Germany based company cb innovations has introduced the new cb Hardcase for iPhone. Featuring fine Italian Premium leather that makes for a unique look and feel, the cb Hardcase... Read more
Sale! B&H Photo offers 12-inch Retina Mac...
B&H Photo has 12″ Retina MacBooks on sale for $300 off MSRP for a limited time. Shipping is free, and B&H charges NY tax only: - 12″ 1.1GHz Gray Retina MacBook: $999 $300 off MSRP - 12″ 1.... Read more
App Annie Reveals Future of the App Economy:...
App Annie, a San Francisco based mobile app data and insights platform, has launched its first comprehensive app economy forecast. This new offering will provide brands, agencies, investors and app... Read more
Apple restocks Certified Refurbished Mac mini...
Apple has restocked Certified Refurbished 2014 Mac minis, with models available starting at $419. Apple’s one-year warranty is included with each mini, and shipping is free: - 1.4GHz Mac mini: $419 $... Read more
What iPad Pro Still Needs To Make It Truly Pr...
I love my iPad Air 2. So much that I’m grudgingly willing to put up with its compromises and limitations as a production tool in order to take advantage of its virtues. However, since a computer for... Read more
21-inch 3.1GHz 4K on sale for $1399, $100 off...
B&H Photo has the 21″ 3.1GHz 4K iMac on sale $1399 for a limited time. Shipping is free, and B&H charges NY sales tax only. Their price is $100 off MSRP: - 21″ 3.1GHz 4K iMac (MK452LL/A): $... Read more
Apple price trackers, updated continuously
Scan our Apple Price Trackers for the latest information on sales, bundles, and availability on systems from Apple’s authorized internet/catalog resellers. We update the trackers continuously: - 15″... Read more
Save up to $240 with Apple Certified Refurbis...
Apple is now offering Certified Refurbished 12″ Retina MacBooks for up to $240 off the cost of new models. Apple will include a standard one-year warranty with each MacBook, and shipping is free. The... Read more
Apple refurbished 13-inch Retina MacBook Pros...
Apple has Certified Refurbished 13″ Retina MacBook Pros available for up to $270 off the cost of new models. An Apple one-year warranty is included with each model, and shipping is free: - 13″ 2.7GHz... Read more
Apple refurbished Time Capsules available for...
Apple has certified refurbished Time Capsules available for $120 off MSRP. Apple’s one-year warranty is included with each Time Capsule, and shipping is free: - 2TB Time Capsule: $179, $120 off - 3TB... Read more

Jobs Board

Lead Engineer *Apple* OSX & Hardware -...
Lead Engineer Apple OSX & Hardware **Job ID:** 3125919 **Full/Part\-Time:** Full\-time **Regular/Temporary:** Regular **Listed:** 2016\-02\-10 **Location:** Cary, Read more
*Apple* Mobile Master - Best Buy (United Sta...
Job Title Apple Mobile Master **Brand** Best Buy **Job Description** **What does a Best Buy Apple Mobile Master do?** At Best Buy, our mission is to leverage the Read more
Infrastructure Engineer - *Apple* /Mac - Rem...
…part of a team Requires proven problem solving skills Preferred Additional: Apple Certified System Administrator (ACSA) Apple Certified Technical Coordinator (ACTC) Read more
Lead Engineer - *Apple* OSX & Hardware...
Lead Engineer - Apple OSX & Hardware **Job ID:** 3125919 **Full/Part\-Time:** Full\-time **Regular/Temporary:** Regular **Listed:** 2016\-02\-10 **Location:** Cary, Read more
Simply Mac *Apple* Specialist- Service Repa...
Simply Mac is the largest premier retailer of Apple products in the nation. In order to support our growing customer base, we are currently looking for a driven Read more
All contents are Copyright 1984-2011 by Xplain Corporation. All rights reserved. Theme designed by Icreon.