1. 1 Intel The Early Intel Microprocessors




Дата канвертавання22.04.2016
Памер123.98 Kb.
1. Microprocessor Architectures


1.1 Intel

1.2 Motorola

1.3 MIPS

1.1 Intel




The Early Intel Microprocessors


The first microprocessor to appear in the market was the Intel 4004, a 4-bit data bus device. This device was followed by the 8008, which had an 8-bit data bus.

Two more 8-bit microprocessors (reference to the number of bits usually refers to the data bus unless stated otherwise), the 8080 and 8085 were introduced in the mid-1970s. These two devices could address only 216 memory locations.




Fig. 1. Fotografia do microprocessador Intel 4004, de 1971, o primeiro microprocessador. A pastilha tem a dimensão de: 0,3 x 0,4cm2 e possui 2.300 transistores.


The 80X86 Family of Microprocessors


Since its introduction in 1978, the so-called X86 architecture has undergo five major evolutionary stages. The term architecture in relation to microprocessors refers to the internal design and organization of the device.

The first generation of the 80X86 family includes the 8086, the 8088, and the 80186. Next, came the 80286, followed by the 80386, and then the 80486. The Pentium is the fifth generation Intel microprocessor.




Intel 8086/8088 and 80186


  • Introduced in 1978, the 8086 was the first 80X86 family and is the basis for all Intel microprocessors that followed.

  • The 8086 was a 16-bit microprocessor (16-bit data bus) and represented a significant departure from the earlier 8-bit devices.

  • Owned 20 address lines (allowing 220 memory locations to be accessed).

  • The various versions of the 8086 operated at clock frequencies of 5, 8, or 10MHz.

  • The 8088 is essentially an 8086 with the 16-bit internal data bus multiplexed down to an 8-bit external bus. It was intended to meet the demand for applications in simpler 8-bit systems and was used in the original IBM personal computer (PC).

  • The 80186 is an 8086 with several support functions: such as clock generator, system controller, interrupt controller, and direct memory access (DMA) controller integrated on the chip. An increased clock frequency of 12.5MHz was added and the 5MHz available in the 8086/8088 was dropped, resulting in a selection of 8, 10, or 12.5MHz.


Intel 80286


  • Introduced in 1982, the memory addressing capability was increased to 24 address lines.

  • First Intel processor to include an advanced mode of operation (used in the next generations of microprocessors): protected mode. This mode allows access to additional memory locations and advanced programming features (Virtual Memory).

  • It operates at the same clock frequencies as the 80186.


Intel 80386


  • Introduced in 1985, it was the first 32-bit Intel microprocessor (32-bit data bus + 32-bit address bus).

  • The first Intel microprocessor to use instruction pipelining.

  • All 80386 versions can operate in conjunction with math (floating-point) co-processors.

  • More economical versions: 80386SX and 80386SL, in which the data bus is multiplexed down to 16 bits and the address bus down to 24 bits.

  • Versions that operate at clock frequencies of 16, 20, 25 and 33MHz are available.


Intel 80486


  • Introduced in 1989, it incorporates an 8KBytes cache memory (shared for data and instruction).

  • The first Intel processor to present an internal floating point unit (FPU).

  • Different versions operate at clock frequencies from 25 to 100MHz (Intel 80486 DX4).


Pentium


  • Introduced in 1993, retains the 32-bit address bus of the 80486 but doubles the data bus to 64 bits.

  • Presents two 8Kbytes cache memories (one for instruction, one for data).

  • Dual pipeline method, known as superscalar architecture.

  • At present (2001), frequencies up to 1.75GHz, 20-stage pipeline, and 3-level cache memory architectures (Itaniun).




1.2 Motorola




The Early Motorola Microprocessors


The first microprocessors from Motorola were all 8-bit devices (8-bit data bus). The 6800 appeared in 1975 with a clock frequency of 2MHz and capable of addressing 64 KBytes of memory with a 16-bit address bus.

In the 6802, a 128 Bytes RAM was added for use in the place of some registers. The clock frequency was increased in the 6803 to 3.58MHz, and a UART (Universal Asynchronous Receiver/Transmitter) was added for serial communications.

The last of the 8-bit microprocessors was the 6809 which offered an enhanced instruction set including a multiply instruction.

All of the 8-bit devices retained the 16-bit address bus.



The 680X0 Family of microprocessors


The Motorola 680X0 family of microprocessors is split into five generations. It was started in 1978, with the MC68000, and was concluded in 1994 with the MC68060. The PowerPC family is the sixth generation of Motorola microprocessors.


Product

Processor Speed

(MHz)


Performance

(MIPS)


Cache-L1 Inst/Data

Bus Interface

(Bits)


Memory

V

Voltage

MC68060

50, 66, 75

110

8K

32-Bit

On-Chip MMU

3V, 5V tolerant I/O

MC68040

25, 33, 40

44

4K

32-Bit

On-Chip MMU

3.3V, 5V

MC68030

16, 20, 25, 33, 40, 50

18

256 Bytes

32-Bit, Dynamic

On-Chip MMU

-

MC68020

12, 16, 20, 25, 33

10

256 Byte I-Cache

32-Bit, Dynamic

-

-

MC68000

8, 10, 12, 16, 20

2

-

-

-

3.3V, 5V



Motorola 68000





  • Introduced in 1978.

  • The RAM-based register concept used in some of the 8-bit devices was abandoned in favor of 16 general-purpose registers.

  • The industry's lowest cost 32-bit microprocessor.

  • The MC68HC000 is a CMOS version of the original MC68000. The MC68EC000 version provides a lower cost 68000 solution. The MC68SEC000 version provides a static, low power implementation consuming only 15.0mA in normal 3.3V operation and 0.5mA in static standby mode.

  • 32 Bit Data and Address Registers.

  • 24 address lines (16 MBytes Direct Addressing Range).

  • 56 powerful instructions. 14 addressing modes.

  • 2 MIPS at 20MHz.

  • Available in 8, 10, 12, 16 & 20 MHz speeds.



Motorola 68020





  • The MC68020 provides a code-compatible upgrade path from the MC68000.

  • The MC68EC020 offers slightly reduced functionality at a lower cost.

  • Motorola entered the world of 32-bit microprocessors with the 68020. 4 GBytes Direct Linear Address Space.

  • It featured a 256-byte cache memory for instructions, an innovative feature at that time.

  • 10 MIPS @ 33 MHz.

  • Available in 12, 16, 20, 25, and 33 MHz.

  • MC68EC020 available in 16 and 25 MHz.




Motorola 68030





  • Another 256-byte cache was added for data.

  • Internal Harvard Architecture.

  • On-Chip Memory Management Unit (unified for Instructions and Data).

  • Burst Memory Interface.

  • The MC68EC030 offers a lower cost embedded solution by removing the memory management unit.

  • 18 MIPS @ 50MHz.

  • Available in 16, 20, 25, 33, 40, and 50 MHz.

  • MC68EC030 Available in 25 and 40 MHz.



Motorola 68040





  • The data and instruction caches (separated) were increased to 4kbytes each.

  • Internal Harvard Architecture.

  • On-Chip Memory Management Unit (separated).

  • Burst Memory Interface.

  • On-chip math coprocessor (floating point unit) was added in the 68040.

  • 44 MIPS @ 40 MHz.

  • Available in 25, 33, and 40 MHz.

  • MC68040V is a Low Power (3.3V) Version of MC68LC040.



Motorola 68060





  • The MC68060 is the first Motorola microprocessor to offer superscalar integer performance.

  • Comes fully equipped with both a floating-point unit (FPU) and a memory management unit (MMU) for high-performance embedded control applications.

  • Executes Three Instructions per Clock.

  • Dual 8K On-Chip Caches.

  • The top clock frequency is 66MHz.

  • For cost-sensitive embedded control and desktop applications where an MMU is required, but the additional cost of a FPU is not justified, the MC68LC060 offers high performance at a low cost. Specifically designed for low-cost embedded control applications, the MC68EC060 eliminates both the FPU and MMU.

  • Branch Cache Reduces Branches to Zero Cycles.

  • Bus Snooping.

  • Power Management.

  • Available in 50 MHz Speed (MC68060, MC68LC060).

  • Available in 66 and 75 MHz Speeds (MC68EC060).




PowerPC 601/601v




  • The PowerPC 601/601v Microprocessor: The PowerPC 601 microprocessor is the first 32-bit implementation of the PowerPC Reduced Instruction Set Computer (RISC) architecture. The PowerPC 601 microprocessor provides high levels of performance for desktop, workstation, and symmetric multiprocessing computer systems and offers design flexibility through operation at either 2.5 volts (601v) or 3.6 volts (601).

  • Superscalar Microprocessor: The PowerPC 601 is a superscalar design capable of issuing and retiring three instructions per clock. Instructions issue to multiple execution units, execute in parallel, and can complete out of order, while preserving program correctness. The PowerPC 601 integrates three execution units: an integer unit (IU), a branch processing unit (BPU), and a floating-point unit (FPU). It also incorporates a memory management unit (MMU), a unified instruction and data cache, a real-time clock (RTC), and on-chip test capability. The ability to execute multiple instructions in parallel and the use of simple instructions with rapid execution times yield maximum efficiency and throughput for PowerPC systems.

  • Cache and MMU Support: The PowerPC 601 microprocessor includes an on-chip 32-Kbyte, eight-way set-associative, physically addressed, unified (instruction and data) cache. An on-chip MMU contains 256-entry, two-way set-associative unified (instruction and data) translation lookaside buffer (UTLB) and provides support for demand paged virtual memory address translation and variable-sized block translation.




  • Flexible Bus Interface: The PowerPC 601 has a high bandwidth: 64-bit data bus and a separate 32-bit address bus. The interface protocol allows multiple masters to access system resources through a central external arbiter. Additionally, on-chip snooping logic maintains cache coherency in multiprocessor applications.




PowerPC 601 Major Features:



Specifications Summary


  • 32-Kbyte unified cache

  • Superscalar-3 instructions per clock

  • Multiple execution

  • 64-bit data bus

  • L2 cache (on separate dies, same package)

  • Power consumption: Full operation - 10 watts maximum at 80 MHz


Technology


  • 3.6-volt implementation

  • 0.6-micron static CMOS technology

  • 120 mm2 die size

  • 2.8 million transistors Packaging

  • 304 CQFP


PowerPC 601v Major Features:



Specifications Summary


  • 32-Kbyte unified cache

  • Superscalar-3 instructions per clock

  • Multiple execution

  • 64-bit data bus

  • L2 cache (on separate dies, same package)

  • Power consumption: Full operation - 6 watts maximum at 100 MHz


Technology


  • 2.5-volt implementation with 5-volt I/O

  • 0.5-micron static CMOS technology

  • 74 mm2 die size

  • 2.8 million transistors

  • 304 CQFP


PowerPC MPC7450




  • The MPC7450 PowerPC microprocessor is a high-performance, low-power, 32-bit implementation of the PowerPC RISC architecture with a full 128-bit implementation of Motorola's AltiVec technology. This microprocessor is ideal for leading edge computing, embedded network control, and signal processing applications. The MPC7450 has a new, deeper, seven-stage pipeline with two additional execution units. The L2 cache has been integrated onto the die for greater speed, and supports a large backside L3 cache with a 64-bit datapath. The MPC7450 offers increased address space and high-bandwidth MPX bus with minimized signal setup times and reduced idle cycles to increase bus bandwidth to a maximum speed of 133 MHz. MPC7450 processors offer single-cycle throughput double precision floating-point performance and full symmetric multi-processing (SMP) capabilities. Finally, the MPC7450 is software-compatible with existing PowerPC 603e, 750, 7400 and 7410 processors and exploits the full potential of AltiVec technology.



  • The MPC7440 PowerPC is a low-power version of the high performance MPC7450. This microprocessor is a small, 360-pin package that does not include the backside L3 cache. It has a core voltage of 1.5 V and is available at speeds of 600 and 700 MHz.



  • Superscalar Microprocessor: MPC7450 features a high-frequency superscalar PowerPC core, capable of issuing four instructions per clock cycle (three instructions + branch) into eleven independent execution units:

  • Four integer units (3 simple + 1 complex)

  • Double-precision floating-point unit

  • Four AltiVec units (simple, complex, floating, and permute)

  • Load/store unit

  • Branch processing unit



  • Cache and MMU Support: The MPC7450 has separate 32KB, physically addressed instruction and data caches. Both L1 caches feature cache way locking and are eight-way set associative. For greater speed, the L2 cache is integrated on-chip with a 256-bit interface to L1 which operates at processor frequency. This L2 is 256KB eight-way set associative. L2 cache access is fully pipelined. The MPC7450 also supports an L3 cache interface with on-chip tags to support up to 2MB of off-chip cache (same package). The L3 data bus is 64-bits wide, provides multiple SRAM options, and affords critical quad-word forwarding to reduce latency. The off-chip L3 storage can also be configured as a local addressable memory. Finally, in addition to supporting hardware table searching on a TLB miss, the MPC7450 can be configured for software table searching. In this case, TLB entries are loaded by the system software. The MPC7450 microprocessor contains separate memory management units for instructions and data, supporting 4 Petabytes (252) of virtual memory and up to 64 Gigabytes (236) of physical memory. The MPC7450 also has four-instruction block address translation and four data block address translation registers.



  • MPX Bus Interface: MPC7450 microprocessors support the MPX bus protocol with a 64-bit data bus and a 32- or 36-bit address bus. Support is included for burst, split, pipelined and out-of-order transactions. The interface provides snooping for data cache coherency. The MPC7450 implements the cache coherency protocol for multiprocessing support in hardware, allowing access to system memory for additional caching bus masters, such as DMA devices.



  • Power Management: MPC7450 microprocessors feature a low-power 1.8-volt design with three power-saving user-programmable modes (nap, doze (with bus snoop) and sleep) which progressively reduce the power drawn by the processor. The MPC7450 also provides a thermal assist unit and instruction cache throttling for software-controllable thermal management.



  • AltiVec Technology: The AltiVec technology expands the capabilities of Motorola's fourth generation PowerPC microprocessors by providing leading-edge, general purpose processing performance while concurrently addressing high-bandwidth data processing and algorithmic-intensive computations in a single-chip solution. AltiVec technology:

  • Meets the computational demands of networking infrastructure such as echo cancellation equipment, and base station processing.

  • Enables faster, more secure encryption methods optimized for the SIMD processing model.

  • Provides compelling performance for multimedia-oriented desktop computers, desktop publishing, and digital video processing.

  • Enables real-time processing of the most demanding data streams (MPEG-2 encode, continuous speech recognition, real-time high-resolution 3D memory for 3D graphics.)



Processor Speed

MHz


Bus Frequency

MHz


Bus Interface

Bits


Cache-L1 Inst/Data

KBytes


L2 Cache

KBytes


L3 Cache

MBytes


Power Dissipation

(Typ)

Watts


Power Dissipation

(Max)

Watts


Package

Process

Voltage int

V


Voltage i/o

V


Performance

MIPS


533, 667, 733 and 867

133

64

32/32

256

1 or 2

14.0 @ 533 MHz

17.0 @ 533 MHz

483 CGBA

0.18 micron 6LM CMOS

1.8

1.8, 2.5

1324 Drystone 2.1 @ 733 MHz


Data Sheets ID

Name

Format

Size K

Rev #

Date Last Modified

MPC7450EC

MPC7450 RISC Microprocessor Hardware Specifications

pdf

1296

3

6/22/2001


Fact Sheets ID

Name

Format

Size K

Rev #

Date Last Modified



















MPC7450FACT

MPC7450FACT High Performance, Low-Power 32-Bit PowerPC RISC Microprocessor

pdf

2497

1.0

7/18/2001



















ALTIVECFACT

AltiVec Technology Motorola's High-Performance Vector Parallel Processing Expansion to the PowerPC Architecture

pdf

284

0

1/01/1998



















ALTIVECWP

Motorola's AltiVec Technology

pdf

171

0

1/01/1998


Miscellaneous ID

Name

Format

Size K

Rev #

Date Last Modified



















MPC7450TS

MPC7450 Technical Summary

pdf

351

0

5/14/2001



















PRMPC7450

Press Release: MPC7450

pdf

51

-

1/09/2001




















Reference Manual ID

Name

Format

Size K

Rev #

Date Last Modified



















ALTIVECPIM

AltiVec Technology Programming Interface Manual

pdf

2195

0

6/01/1999



















MPC7450UM

MPC7450 RISC Microprocessor Users Manual

pdf

9308

0

3/29/2001



















MPC7450UM_CH

Additional options for downloading the MPC7450 RISC Microprocessor Users Manual

html

6

0

3/29/2001



















MPC7450UM_ZIP

MPC7450 Users Manual (Compressed)

zip

3064

0

3/29/2001



















ALTIVECPEM

AltiVec Technology Programming Environments Manual

pdf

3675

1.1

5/04/2001


Reports or Presentations ID

Name

Format

Size K

Rev #

Date Last Modified

PPCCPUSUMM

PowerPC CPU Summary

pdf

9

-

3/21/2001


White Paper ID

Name

Format

Size K

Rev #

Date Last Modified

G4WP

PowerPC G4 Architecture White Paper

pdf

53

0

1/23/2001

 




1.3 MIPS





Fig. 1. Fotografia mostrando o interior do microprocessador MIPS R3000. No lado direito do chip, pode-se ver a parte operativa. O lado superior esquerdo contém a unidade de controle. O canto inferior esquerdo contém a parte do sistema de memória chamada "Translation Lookaside Buffer", ou seja, a memória cache do microprocessador.

Fig. 2. Fotografia mostrando o interior do microprocessador MIPS R4000. A pastilha tem a dimensão de 1,5cm por 1,1cm, contendo cerca de 1,3 milhão de transistores. No lado direito do chip pode-se ver a parte operativa para cálculo de operações com números inteiros, enquanto que o lado esquerdo contem a parte operativa para cálculos em ponto flutuante. A unidade de controle está no centro do chip e os dois grandes blocos no alto são as memórias cache do microprocessador.


Fig. 3. Fotografia de um wafer de 6 polegadas contendo microprocessadores MIPS R4000 (esquerda) e um wafer de 6 polegadas contendo microprocessadores MIPS R3000 (direita). O número de pastilhas do R3000 no wafer da direita, para um rendimento de 100%, é 210. Cada pastilha mede 0,8 x 0,9cm2 e contem cerca de 125.000 transistores. O número de pastilhas do R4000 no wafer da esquerda, para um rendimento de 100%, é de 59. O tamanho da pastilha é 1,5 x 1,1cm2 e cada pastilha contem cerca de 1,3 milhão de transistores. Alguns dos chips do R3000 colocados nas bordas do wafer não serão utilizados; eles foram incluídos porque é mais fácil para se criar as máscaras para imprimir o circuito no silício. Os retângulos "vazios" em ambos wafers contém circuitos de teste usados para rapidamente se testar o wafer. O wafer do MIPS R4000 tem quatro circuitos de teste adicionais nas bordas.


База данных защищена авторским правом ©shkola.of.by 2016
звярнуцца да адміністрацыі

    Галоўная старонка