Game Engine Architecture

122 3. Fundamentals of Software Engineering for Games

example of such a diagram for the struct Foo listed below is shown in Fig-

ure 3.11.

struct Foo

{

U32 mUnsignedValue;

F32 mFloatValue;

I32 mSignedValue;

};

The sizes of the data members are important and should be represented

in your diagrams. This is easily done by using the width of each data member

to indicate its size in bits—i.e., a 32-bit integer should be roughly four times

the width of an 8-bit integer (see Figure 3.12).

struct Bar

{

U32 mUnsignedValue;

F32 mFloatValue;

bool mBooleanValue; // diagram assumes this is 8 bits

};

3.2.5.1. Alignment and Packing

As we start to think more carefully about the layout of our structs and classes

in memory, we may start to wonder what happens when small data members

are interspersed with larger members. For example:

struct InefficientPacking

{

U32 mU1; // 32 bits

F32 mF2; // 32 bits

U8 mB3; // 8 bits

I32 mI4; // 32 bits

bool mB5; // 8 bits

char* mP6; // 32 bits

};

You might imagine that the compiler simply packs the data members into

memory as tightly as it can. However, this is not usually the case. Instead,

the compiler will typically leave “holes” in the layout, as depicted in Fig-

ure 3.13. (Some compilers can be requested not to leave these holes by us-

ing a preprocessor directive like #pragma pack, or via command-line op-

tions; but the default behavior is to space out the members as shown in Fig-

ure 3.13.)

mU1

mF2

mB3

mI4

mB5

mP6

+0x0
+0x4
+0x8
+0xC
+0x10
+0x14

Figure 3.13. Ineffi cient
struct packing due to
mixed data member
sizes.

mUnsignedValue

mFloatValue

mSignedValue

+0x0
+0x4
+0x8

Figure 3.11. Memory
layout of a simple
struct.

mUnsignedValue

mFloatValue

mBooleanValue

+0x0
+0x4
+0x8

Figure 3.12. A memory
layout using width to
indicate member sizes.