Math Coprocessor Also called Floating Point Unit FPU
- Slides: 20
Math Coprocessor Also called Floating Point Unit FPU 11/23/2020 TUC-N dr. Emil CEBUC
Outline l l l Math Coprocessor Operation Math Coprocessor Structure Data Types Representation error Instructions 11/23/2020 TUC-N dr. Emil CEBUC
Math Coprocessor Operation l l l Shares the same Data, Address and Control BUS as the main processor Two different chips for old processors On the same silicon die starting with 486 DX Instructions preceded by an ESC sequence Operates in parallel with main processor Coprocessor may overtake BUS for longer periods if more data is needed 11/23/2020 TUC-N dr. Emil CEBUC
Math Coprocessor Operation l l l Coprocessor has no access to registers But can use registers for addressing All addressing modes are available except immediate addressing Uses special synchronization signals to cooperate Instructions take tens to hundreds of cycles to complete 11/23/2020 TUC-N dr. Emil CEBUC
Math Coprocessor Structure l l l l Register Stack Control Register Status Register Tag Register Instruction Pointer Data Pointer Execution Unit EU 11/23/2020 Control Register STACK 8 x 80 bit registers TUC-N dr. Emil CEBUC Execution Unit
FPU Register Stack 11/23/2020 TUC-N dr. Emil CEBUC
FPU Control Register 11/23/2020 TUC-N dr. Emil CEBUC
FPU Status Register 11/23/2020 TUC-N dr. Emil CEBUC
Meaning of C 3 C 2 C 1 C 0 bits after compare instructions l l No Conditional Jumps to test FPU compare results Must copy Status Register in to Main Proc registers and examine with x 86 instructions 11/23/2020 TUC-N dr. Emil CEBUC
FPU Tag Register 15 14 13 12 11 10 9 TAG 7 TAG 4 TAG 6 TAG 5 TAG Values 00 01 10 11 11/23/2020 8 7 6 TAG 3 5 4 TAG 2 3 TAG 1 meaning Valid Zero Special, NAN, ∞ Empty TUC-N dr. Emil CEBUC 2 1 0 TAG 0
FPU Pointers l l Instruction Pointer a pointer to the instruction that caused an exception Data Pointer a pointer to the data used by the instruction that caused an exception 11/23/2020 TUC-N dr. Emil CEBUC
Number Representation l l l Integer numbers in C 2 Integer numbers in Packed Decimal Real Numbers in IEEE 754/854 standard format All these representations are ONLY in memory Internally ALL numbers are represented on 80 bits as temporary real’s 11/23/2020 TUC-N dr. Emil CEBUC
Integer Data Types l l Word Integer DW 16 bit C 2 representation Short Integer DD 32 bit C 2 representation Long Integer DQ 64 bit C 2 representation Packed Decimal DT 80 bit Value & Sign 11/23/2020 TUC-N dr. Emil CEBUC
Real Number Representation Normalized Form l (-1)S * 1. mantissa *2 (exponent- Bias) l Bias = 07 FH for short real; 127 d l Bias = 03 FFH for long real; 1023 d l Bias = 03 FFFh for temp real; 16383 d 11/23/2020 TUC-N dr. Emil CEBUC
Short Real 32 bits Double Word DD 11/23/2020 TUC-N dr. Emil CEBUC
Long Real 64 bits Quad Word DQ 11/23/2020 TUC-N dr. Emil CEBUC
Temporary Real 80 bits Ten word DT 11/23/2020 TUC-N dr. Emil CEBUC
Example of real number representation fl 1 dd 0000 1. 0 3 F 800000 0011 1111 1000 0000 0000 Fl 2 dd 0004 2. 0 40000000 0100 0000 0000 fl 1_3 0008 dd -1. 3; actually -1, 299995 BFA 66666 1011 1111 1010 0110 0110 11/23/2020 TUC-N dr. Emil CEBUC
Representation error l l Unavoidable May be very small Must take account when designing numerical algorithm for complex computation Can be controlled with rounding policy from Control Word 11/23/2020 TUC-N dr. Emil CEBUC
FPU Instructions l l l l Data movement Data conversions Arithmetic Instructions Compare Instructions Constant Instructions Transcendental Instructions Miscellaneous Instructions 11/23/2020 TUC-N dr. Emil CEBUC
- Motion coprocessor
- Fixed point vs floating point
- Myfpu moodle
- Fpu & fvu software download
- Fpu registrar
- Floating red markers nuns shape
- Floating red markers are called nuns. what shape are nuns?
- Node cusp and conjugate point
- Fp adder
- Parts of a floating point number
- Floating point addition flowchart
- Mips mul
- Floating point operations per seconds
- Point format example
- Floating point number system
- Ieee denormalized numbers
- Floating point
- Floating point representation
- Floating-point number
- Fp adder hardware
- Eascii