CPU: Difference between revisions

From Braindump
Jump to navigation Jump to search
Line 140: Line 140:
|University of Berkeley
|University of Berkeley
| colspan="2" |Intended for research
| colspan="2" |Intended for research
|-
|
|
|
|-
|-
|1981
|1981
Line 167: Line 171:
|
|
|DARPA VLSI Project University funding
|DARPA VLSI Project University funding
|-
|
|RISC
|University of Berkley (David Patterson)
|-
|-
|
|

Revision as of 09:20, 10 August 2024

Super Computers

1943 ENIAC

  EDVAC

   BINAC

   1951 UNIVAC

  ORDVAC

  UNIVAC

  UNISYS

ILLIAC

EDSAC

Manchaster Baby

CTC Computer Terminal Corporation

Datapoint 3300

ERA Remington Rand UNIVAC

CDC

CRAY

1955: Shockley Semiconductor Laboratory (Caltech / MIT / Bell labs)

1957: Fairchild Semiconductor (traitorous eight)

1968: Intel (Noyce and Moore)

1974: Zilog (Federico Faggin)

Z80 is a 5V spinoff from an Intel 8080

Now Littlefuse > IXYS > Zilog

CPU-Z Z270 i5-7600K
CPU-Z Z270 i5-7600K
Fairchild, Intel, Motorola, TI, MOS, Zilog, Sun, Dec, IBM, ARM
Fairchild IC's (traitorous eight)
Intel 4004, 8008, 8088
Zilog Z80
Motorola 6800
MOS 6502
Acorn ARM
CPU-History.png
RISC Berkley / IBM
MIPS Stanford
IBM PowerPC
SUN Sparc
DEC Alpha
MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, EM64T, VT-x, AES, AVX, AVX2, FMA3, TSX

Moore's Law / Transistors

Integration level Year Logic gates
SSI Small Scale Integration 1964
MSI Medium Scale Integration 1968 20-200
LSI Large‐scale integration 1970 200-2000 10 µm
VLSI Very large‐scale integration 1980 2.000-20.000 1.5 µm
ULSI Ultra large‐scale Integration 1990 20.000-200.000 600 nm
SLSI Super large‐scale integration 2000 200.000- 2 million 130 nm
2010 2 million - 20 million 22 nm
2020 20 million - 200 million 5 nm

CISC v.s. RISC

RISC is a philosophy to use simpler instructions to achieve the same as complex instruction. For example for AES rowshifting an Intel can use XMM instructions to shift multiple rows in one operation, while RISC processors need to loop through the bytes and execute simpler instructions. It takes longer to execute, but without SIMD the processor is smaller, cheaper and wins in power efficiency.

Intel CISC

PCLMULHQHQDQ xmmreg,xmmrm

ARM RISC

result = operand1 EOR operand2;

for s = 0 to segments-1

Elem[result, s, 128] = AESSubBytes(AESShiftRows(Elem[result, s, 128]));

RISC

University of Berkeley Intended for research
1981 RISC I 44,500 transistors, 31 instructions, 78 32-bit registers
RISC Foundation Open ISA free to implement
2015 RISC V

MIPS

1976 IBM 801 first conceptual RISC processor
1980 DARPA VLSI Project University funding
MIPS Stanford University (John Hennessy)
1984 MIPS Computer Systems, Inc.
1992 Acquired by SGI
1998 SGI spun off MIPS
2008 lost money on buying and selling Chipidea
2013 bought by Imagination Technologies (PowerVR)
2022 Adopted RISC-V
1997 Lexra based on MIPS-I instruction set (Not the core license)
1998 Patent on "unaligned loads and stores"
2003 Lexra bankrupt
2006 RTL8196E Realtek continues to use Lexra Cores (RLX4181)

ARM

Cortex-A Microprocessors, with an MMU, for Rich OS e.g. BSD/Linux/Windows)

Cortex-R Realtime processors

Cortex-M Microcontrollers for RTOS Task Scheduling


ARMv7-M ISA

M0+ von neuman (instruction and data share the same bus)

M3

M4

M7


ARMv8-M

M23 Trustzone

M33

M35P


Armv8.1-MM55

M85

Segger

SEGGER J-Link EDU Mini - JTAG/SWD Debugger

https://thepihut.com/products/segger-j-link-edu-mini-jtag-swd-debugger

https://www.adafruit.com/product/3571

https://1bitsquared.com/products/black-magic-probe