Transistor count

https://en.wikipedia.org/wiki/Transistor_count

 

Transistor count

From Wikipedia, the free encyclopedia

Jump to navigationJump to search

Plot of MOS transistor counts for microprocessors against dates of in­tro­duction. The curve shows counts doubling every two years, per Moore's law.

The transistor count is the number of transistors on an integrated circuit (IC). It typically refers to the number of MOSFETs (metal-oxide-semiconductor field-effect transistors, or MOS transistors) on an IC chip, as all modern ICs use MOSFETs. It is the most common measure of IC complexity (although the majority of transistors in modern microprocessors are contained in the cache memories, which consist mostly of the same memory cell circuits replicated many times). The rate at which MOS transistor counts have increased generally follows Moore's law, which observed that the transistor count doubles approximately every two years.

As of 2019, the largest transistor count in a commercially available microprocessor is 39.54 billion MOSFETs, in AMD's Zen 2 based Epyc Rome, which is a 3D integrated circuit (with eight dies in a single package) fabricated using TSMC's 7 nm FinFET semiconductor manufacturing process.[1][2] As of 2018, the highest transistor count in a graphics processing unit (GPU) is Nvidia's GV100 Volta with 21.1 billion MOSFETs, manufactured using TSMC's 12 nm FinFET process.[3] As of 2019, the highest transistor count in any IC chip is Samsung's 1 TB eUFS (3D-stacked) V-NAND flash memory chip, with 2 trillion floating-gate MOSFETs (4 bits per transistor).[4] As of 2019, the highest transistor count in a non-memory chip is a deep learning engine called the Wafer Scale Engine by Cerebras, using a special design to route around any non-functional core on the device; it has 1.2 trillion MOSFETs, manufactured using TSMC's 16 nm FinFET process.[5][6][7][8]

In terms of computer systems that consist of numerous integrated circuits, the supercomputer with the highest transistor count as of 2016 is the Chinese-designed Sunway TaihuLight, which has for all CPUs/nodes (1012 for the 10 million cores and for RAM 1015 for the 1.3 million GB) combined "about 400 trillion transistors in the processing part of the hardware" and "the DRAM includes about 12 quadrillion transistors, and that's about 97 percent of all the transistors."[9] To compare, the smallest computer, as of 2018 dwarfed by a grain of sand, has on the order of 100,000 transistors, and the one, fully programmable, with the fewest transistors ever has 130 transistors or fewer.

In terms of the total number of transistors in existence, it has been estimated that a total of 13 sextillion (1.3×1022) MOSFETs have been manufactured worldwide between 1960 and 2018, accounting for at least 99.9% of all transistors. This makes the MOSFET the most widely manufactured device in history.[10]

 

Contents

Transistor count

 

Part of an IBM 7070 card cage populated with Standard Modular System cards

Among the earliest products to use transistors were portable transistor radios, introduced in 1954, which typically used 4 to 8 transistors, often advertising the number on the radio's case. However, early junction transistors were relatively bulky devices that were difficult to manufacture on a mass-production basis, limiting the transistor counts and restricting their usage to a number of specialised applications.[11]

The MOSFET (MOS transistor), invented by Mohamed Atalla and Dawon Kahng at Bell Labs in 1959,[12] was the first truly compact transistor that could be miniaturised and mass-produced for a wide range of uses.[11] The MOSFET made it possible to build high-density integrated circuits (ICs),[13] enabling Moore's law[14][15] and very large-scale integration.[16] Atalla first proposed the concept of the MOS integrated circuit (MOS IC) chip in 1960, followed by Kahng in 1961, both noting that the MOSFET's ease of fabrication made it useful for integrated circuits.[11][17] The earliest experimental MOS IC to be demonstrated was a 16-transistor chip built by Fred Heiman and Steven Hofstein at RCA Laboratories in 1962.[15] Further large-scale integration was made possible with an improvement in MOSFET semiconductor device fabrication, the CMOS process, developed by Chih-Tang Sah and Frank Wanlass at Fairchild Semiconductor in 1963.[18]

Microprocessors[edit]

See also: Microprocessor chronology and Microcontroller

This subsection needs additional citations for verification. Relevant discussion may be found on the talk page. Please help improve this article by adding citations to reliable sources. Unsourced material may be challenged and removed.
Find sources: "Transistor count" – news · newspapers · books · scholar · JSTOR (December 2019) (Learn how and when to remove this template message)

A microprocessor incorporates the functions of a computer's central processing unit on a single integrated circuit. It is a multi-purpose, programmable device that accepts digital data as input, processes it according to instructions stored in its memory, and provides results as output.

The development of MOS integrated circuit technology in the 1960s led to the development of the first microprocessors.[19] The 20-bit MP944, developed by Garrett AiResearch for the U.S. Navy's F-14 Tomcat fighter in 1970, is considered by its designer Ray Holt to be the first microprocessor.[20] It was a multi-chip microprocessor, fabricated on six MOS chips. However, it was classified by the Navy until 1998. The 4-bit Intel 4004, released in 1971, was the first single-chip microprocessor. It was made possible with an improvement in MOSFET design, MOS silicon-gate technology (SGT), developed in 1968 at Fairchild Semiconductor by Federico Faggin, who went on to use MOS SGT technology to develop the 4004 with Marcian Hoff, Stanley Mazor and Masatoshi Shima at Intel.[19]

All chips over e.g. a million transistors have lots of memory, usually cache memories in level 1 and 2 or more levels, accounting for most transistors on microprocessors in modern times, where large caches have become the norm. The level 1 caches of the Pentium Pro die accounted for over 14% of its transistors, while the much larger L2 cache was on a separate die, but on-package, so it's not included in the transistor count. Later chips included more levels, L2 or even L3 on-chip. The last DEC Alpha chip made has 90% of it for cache.[21]

While Intel's i960CA small cache of 1 KB, at about 50,000 transistors, isn't a big part of the chip, it alone would have been very large in early microprocessors. In the ARM 3 chip, with 4 KB, the cache was over 63% of the chip, and in the Intel 80486 its larger cache is only over a third of it because the rest of the chip is more complex. So cache memories are the largest factor, except for in early chips with smaller caches or even earlier chips with no cache at all. Then the inherent complexity, e.g. number of instructions, is the dominant factor, more than e.g. the memory the registers of the chip represent.

Processor MOS transistor count Date of
introduction		 Designer MOS
process		 Area
MP944 (20-bit, 6-chip) ? 1970[20][a] Garrett AiResearch ? ?
Intel 4004 (4-bit, 16-pin) 2,250 1971 Intel 10,000 nm 12 mm2
Intel 8008 (8-bit, 18-pin) 3,500 1972 Intel 10,000 nm 14 mm2
NEC μCOM-4 (4-bit, 42-pin) 2,500[22][23] 1973 NEC 7,500 nm[24] ?
Toshiba TLCS-12 (12-bit) 11,000+[25] 1973 Toshiba 6,000 nm 32 mm2
Intel 4040 (4-bit, 16-pin) 3,000 1974 Intel 10,000 nm 12 mm2
Motorola 6800 (8-bit, 40-pin) 4,100 1974 Motorola 6,000 nm 16 mm2
Intel 8080 (8-bit, 40-pin) 6,000 1974 Intel 6,000 nm 20 mm2
TMS 1000 (4-bit, 28-pin) 8,000 1974[26] Texas Instruments 8,000 nm 11 mm2
MOS Technology 6502 (8-bit, 40-pin) 4,528[b][27] 1975 MOS Technology 8,000 nm 21 mm2
Intersil IM6100 (12-bit, 40-pin; clone of PDP-8) 4,000 1975 Intersil    
CDP 1801 (8-bit, 2-chip, 40-pin) 5,000 1975 RCA    
RCA 1802 (8-bit, 40-pin) 5,000 1976 RCA 5,000 nm 27 mm2
Zilog Z80 (8-bit, 4-bit ALU, 40-pin) 8,500[c] 1976 Zilog 4,000 nm 18 mm2
Intel 8085 (8-bit, 40-pin) 6,500 1976 Intel 3,000 nm 20 mm2
TMS9900 (16-bit) 8,000 1976 Texas Instruments    
Motorola MC14500B (1-bit, 16-pin) ? 1977 Motorola ? ?
Bellmac-8 (8-bit) 7,000 1977 Bell Labs 5,000 nm  
Motorola 6809 (8-bit with some 16-bit features, 40-pin) 9,000 1978 Motorola 5,000 nm 21 mm2
Intel 8086 (16-bit, 40-pin) 29,000 1978 Intel 3,000 nm 33 mm2
Zilog Z8000 (16-bit) 17,500[28] 1979 Zilog    
Intel 8088 (16-bit, 8-bit data bus) 29,000 1979 Intel 3,000 nm 33 mm2
Motorola 68000 (16/32-bit, 32-bit registers, 16-bit ALU) 68,000[29] 1979 Motorola 3,500 nm 44 mm2
Intel 8051 (8-bit, 40-pin) 50,000 1980 Intel    
WDC 65C02 11,500[30] 1981 WDC 3,000 nm 6 mm2
ROMP (32-bit) 45,000 1981 IBM 2,000 nm  
Intel 80186 (16-bit, 68-pin) 55,000 1982 Intel 3,000 nm 60 mm2
Intel 80286 (16-bit, 68-pin) 134,000 1982 Intel 1,500 nm 49 mm2
WDC 65C816 (8/16-bit) 22,000[31] 1983 WDC 3,000 nm[32] 9 mm2
NEC V20 63,000 1984 NEC    
Motorola 68020 (32-bit; 114 pins used) 190,000[33] 1984 Motorola 2,000 nm 85 mm2
Intel 80386 (32-bit, 132-pin; no cache) 275,000 1985 Intel 1,500 nm 104 mm2
ARM 1 (32-bit; no cache) 25,000[33] 1985 Acorn 3,000 nm 50 mm2
Novix NC4016 (16-bit) 16,000[34] 1985[35] Harris Corporation 3,000 nm[36]  
SPARC MB86900 (32-bit; no cache) 110,000[37] 1986 Fujitsu 1,200 nm  
NEC V60[38] (32-bit; no cache) 375,000 1986 NEC 1,500 nm  
ARM 2 (32-bit, 84-pin; no cache) 27,000[39][33] 1986 Acorn 2,000 nm 30.25 mm2
Z80000 (32-bit; very small cache) 91,000 1986 Zilog    
NEC V70[38] (32-bit; no cache) 385,000 1987 NEC 1,500 nm  
Hitachi Gmicro/200[40] 730,000 1987 Hitachi 1,000 nm  
Motorola 68030 (32-bit, very small caches) 273,000 1987 Motorola 800 nm 102 mm2
TI Explorer's 32-bit Lisp machine chip 553,000[41] 1987 Texas Instruments 2,000 nm[42]  
DEC WRL MultiTitan 180,000[43] 1988 DEC WRL 1,500 nm 61 mm2
Intel i960 (32-bit, 33-bit memory subsystem, no cache) 250,000[44] 1988 Intel 1,500 nm[45]  
Intel i960CA (32-bit, cache) 600,000[45] 1989 Intel 800 nm 143 mm2
Intel i860 (32/64-bit, 128-bit SIMD, cache, VLIW) 1,000,000[46] 1989 Intel    
Intel 80486 (32-bit, 4 KB cache) 1,180,235 1989 Intel 1000 nm 173 mm2
ARM 3 (32-bit, 4 KB cache) 310,000 1989 Acorn 1,500 nm 87 mm2
Motorola 68040 (32-bit, 8 KB caches) 1,200,000 1990 Motorola 650 nm 152 mm2
R4000 (64-bit, 16 KB of caches) 1,350,000 1991 MIPS 1,000 nm 213 mm2
ARM 6 (32-bit, no cache for this 60 variant) 35,000 1991 ARM 800 nm  
Hitachi SH-1 (32-bit, no cache) 600,000[47] 1992[48] Hitachi 800 nm 10 mm2
Intel i960CF (32-bit, cache) 900,000[45] 1992 Intel   125 mm2
DEC Alpha 21064 (64-bit, 290-pin; 16 KB of caches) 1,680,000 1992 DEC 750 nm 233.52 mm2
Hitachi HARP-1 (32-bit, cache) 2,800,000[49] 1993 Hitachi 500 nm 267 mm2
Pentium (32-bit, 16 KB of caches) 3,100,000 1993 Intel 800 nm 294 mm2
ARM700 (32-bit; 8 KB cache) 578,977[50] 1994 ARM 700 nm 68.51 mm2
MuP21 (21-bit,[51] 40-pin; includes video) 7,000[52] 1994 Offete Enterprises 1200 nm  
Motorola 68060 (32-bit, 16 KB of caches) 2,500,000 1994 Motorola 600 nm 218 mm2
SA-110 (32-bit, 32 KB of caches) 2,500,000[33] 1995 Acorn/DEC/Apple 350 nm 50 mm2
Pentium Pro (32-bit, 16 KB of caches;[53] L2 cache on-package, but on separate die) 5,500,000[54] 1995 Intel 500 nm 307 mm2
AMD K5 (32-bit, caches) 4,300,000 1996 AMD 500 nm 251 mm2
Hitachi SH-4 (32-bit, caches) 10,000,000[55] 1997 Hitachi 200 nm[56] 42 mm2[57]
Pentium II Klamath (32-bit, 64-bit SIMD, caches) 7,500,000 1997 Intel 350 nm 195 mm2
AMD K6 (32-bit, caches) 8,800,000 1997 AMD 350 nm 162 mm2
F21 (21-bit; includes e.g. video) 15,000 1997[52] Offete Enterprises    
AVR (8-bit, 40-pin; w/memory) 140,000 (48,000 excl. memory[58]) 1997 Nordic VLSI/Atmel    
Pentium II Deschutes (32-bit, large cache) 7,500,000 1998 Intel 250 nm 113 mm2
ARM 9TDMI (32-bit, no cache) 111,000[33] 1999 Acorn 350 nm 4.8 mm2
Pentium III Katmai (32-bit, 128-bit SIMD, caches) 9,500,000 1999 Intel 250 nm 128 mm2
Emotion Engine (64-bit, 128-bit SIMD, cache) 13,500,000[59] 1999 Sony/Toshiba 180 nm[60] 240 mm2[61]
Pentium II Mobile Dixon (32-bit, caches) 27,400,000 1999 Intel 180 nm 180 mm2
AMD K6-III (32-bit, caches) 21,300,000 1999 AMD 250 nm 118 mm2
AMD K7 (32-bit, caches) 22,000,000 1999 AMD 250 nm 184 mm2
Gekko (32-bit, large cache) 21,000,000[62] 2000 IBM/Nintendo 180 nm 43 mm2
Pentium III Coppermine (32-bit, large cache) 21,000,000 2000 Intel 180 nm 80 mm2
Pentium 4 Willamette (32-bit, large cache) 42,000,000 2000 Intel 180 nm 217 mm2
SPARC64 V (64-bit, large cache) 191,000,000[63] 2001 Fujitsu 130 nm[64] 290 mm2
Pentium III Tualatin (32-bit, large cache) 45,000,000 2001 Intel 130 nm 81 mm2
Pentium 4 Northwood (32-bit, large cache) 55,000,000 2002 Intel 130 nm 145 mm2
Itanium 2 McKinley (64-bit, large cache) 220,000,000 2002 Intel 180 nm 421 mm2
DEC Alpha 21364 (64-bit, 946-pin, SIMD, very large caches) 152,000,000[21] 2003 DEC 180 nm 397 mm2
Barton (32-bit, large cache) 54,300,000 2003 AMD 130 nm 101 mm2
AMD K8 (64-bit, large cache) 105,900,000 2003 AMD 130 nm 193 mm2
Itanium 2 Madison 6M (64-bit) 410,000,000 2003 Intel 130 nm 374 mm2
Pentium 4 Prescott (32-bit, large cache) 112,000,000 2004 Intel 90 nm 110 mm2
SPARC64 V+ (64-bit, large cache) 400,000,000[65] 2004 Fujitsu 90 nm 294 mm2
Itanium 2 (64-bit;9 MB cache) 592,000,000 2004 Intel 130 nm 432 mm2
Pentium 4 Prescott-2M (32-bit, large cache) 169,000,000 2005 Intel 90 nm 143 mm2
Pentium D Smithfield (32-bit, large cache) 228,000,000 2005 Intel 90 nm 206 mm2
Xenon (64-bit, 128-bit SIMD, large cache) 165,000,000 2005 IBM 90 nm  
Cell (32-bit, cache) 250,000,000[66] 2005 Sony/IBM/Toshiba 90 nm 221 mm2
Pentium 4 Cedar Mill (32-bit, large cache) 184,000,000 2006 Intel 65 nm 90 mm2
Pentium D Presler (32-bit, large cache) 362,000,000 2006 Intel 65 nm 162 mm2
Core 2 Duo Conroe (dual-core 64-bit, large caches) 291,000,000 2006 Intel 65 nm 143 mm2
Dual-core Itanium 2 (64-bit, SIMD, large caches) 1,700,000,000[67] 2006 Intel 90 nm 596 mm2
AMD K10 quad-core 2M L3 (64-bit, large caches) 463,000,000[68] 2007 AMD 65 nm 283 mm2
ARM Cortex-A9 (32-bit, (optional) SIMD, caches) 26,000,000[69] 2007 ARM 45 nm 31 mm2
Core 2 Duo Wolfdale (dual-core 64-bit, SIMD, caches) 411,000,000 2007 Intel 45 nm 107 mm2
POWER6 (64-bit, large caches) 789,000,000 2007 IBM 65 nm 341 mm2
Core 2 Duo Allendale (dual-core 64-bit, SIMD, large caches) 169,000,000 2007 Intel 65 nm 111 mm2
Uniphier 250,000,000[70] 2007 Matsushita 45 nm ?
SPARC64 VI (64-bit, SIMD, large caches) 540,000,000 2007[71] Fujitsu 90 nm 421 mm2
Core 2 Duo Wolfdale 3M (dual-core 64-bit, SIMD, large caches) 230,000,000 2008 Intel 45 nm 83 mm2
Core i7 (quad-core 64-bit, SIMD, large caches) 731,000,000 2008 Intel 45 nm 263 mm2
AMD K10 quad-core 6M L3 (64-bit, SIMD, large caches) 758,000,000[68] 2008 AMD 45 nm 258 mm2
Atom (32-bit, large cache) 47,000,000 2008 Intel 45 nm 24 mm2
SPARC64 VII (64-bit, SIMD, large caches) 600,000,000 2008[72] Fujitsu 65 nm 445 mm2
Six-core Xeon 7400 (64-bit, SIMD, large caches) 1,900,000,000 2008 Intel 45 nm 503 mm2
Six-core Opteron 2400 (64-bit, SIMD, large caches) 904,000,000 2009 AMD 45 nm 346 mm2
SPARC64 VIIIfx (64-bit, SIMD, large caches) 760,000,000[73] 2009 Fujitsu 45 nm 513 mm2
16-core SPARC T3 (64-bit, SIMD, large caches) 1,000,000,000[74] 2010 Sun/Oracle 40 nm 377 mm2
Six-core Core i7 (Gulftown) 1,170,000,000 2010 Intel 32 nm 240 mm2
8-core POWER7 32M L3 (64-bit, SIMD, large caches) 1,200,000,000 2010 IBM 45 nm 567 mm2
Quad-core z196[75] (64-bit, very large caches) 1,400,000,000 2010 IBM 45 nm 512 mm2
Quad-core Itanium Tukwila (64-bit, SIMD, large caches) 2,000,000,000[76] 2010 Intel 65 nm 699 mm2
8-core Xeon Nehalem-EX (64-bit, SIMD, large caches) 2,300,000,000[77] 2010 Intel 45 nm 684 mm2
SPARC64 IXfx (64-bit, SIMD, large caches) 1,870,000,000[78] 2011 Fujitsu 40 nm 484 mm2
Quad-core + GPU Core i7 (64-bit, SIMD, large caches) 1,160,000,000 2011 Intel 32 nm 216 mm2
Six-core Core i7/8-core Xeon E5
(Sandy Bridge-E/EP) (64-bit, SIMD, large caches)		 2,270,000,000[79] 2011 Intel 32 nm 434 mm2
10-core Xeon Westmere-EX (64-bit, SIMD, large caches) 2,600,000,000 2011 Intel 32 nm 512 mm2
Atom "Medfield" (64-bit) 432,000,000[80] 2012 Intel 32 nm 64 mm2
SPARC64 X (64-bit, SIMD, caches) 2,990,000,000[81] 2012 Fujitsu 28 nm 600 mm2
8-core AMD Bulldozer (64-bit, SIMD, caches) 1,200,000,000[82] 2012 AMD 32 nm 315 mm2
Quad-core + GPU AMD Trinity (64-bit, SIMD, caches) 1,303,000,000 2012 AMD 32 nm 246 mm2
Quad-core + GPU Core i7 Ivy Bridge (64-bit, SIMD, caches) 1,400,000,000 2012 Intel 22 nm 160 mm2
8-core POWER7+ (64-bit, SIMD, 80 MB L3 cache) 2,100,000,000 2012 IBM 32 nm 567 mm2
Six-core zEC12 (64-bit, SIMD, large caches) 2,750,000,000 2012 IBM 32 nm 597 mm2
8-core Itanium Poulson (64-bit, SIMD, caches) 3,100,000,000 2012 Intel 32 nm 544 mm2
61-core Xeon Phi (32-bit, 512-bit SIMD, caches) 5,000,000,000[83] 2012 Intel 22 nm 720 mm2
Apple A7 (dual-core 64/32-bit ARM64, "mobile SoC", SIMD, caches) 1,000,000,000 2013 Apple 28 nm 102 mm2
Six-core Core i7 Ivy Bridge E (64-bit, SIMD, caches) 1,860,000,000 2013 Intel 22 nm 256 mm2
12-core POWER8 (64-bit, SIMD, caches) 4,200,000,000 2013 IBM 22 nm 650 mm2
Xbox One main SoC (64-bit, SIMD, caches) 5,000,000,000 2013 Microsoft/AMD 28 nm 363 mm2
Quad-core + GPU Core i7 Haswell (64-bit, SIMD, caches) 1,400,000,000[84] 2014 Intel 22 nm 177 mm2
Apple A8 (dual-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 2,000,000,000 2014 Apple 20 nm 89 mm2
8-core Core i7 Haswell-E (64-bit, SIMD, caches) 2,600,000,000[85] 2014 Intel 22 nm 355 mm2
Apple A8X (tri-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 3,000,000,000[86] 2014 Apple 20 nm 128 mm2
15-core Xeon Ivy Bridge-EX (64-bit, SIMD, caches) 4,310,000,000[87] 2014 Intel 22 nm 541 mm2
18-core Xeon Haswell-E5 (64-bit, SIMD, caches) 5,560,000,000[88] 2014 Intel 22 nm 661 mm2
Quad-core + GPU GT2 Core i7 Skylake K (64-bit, SIMD, caches) 1,750,000,000 2015 Intel 14 nm 122 mm2
Dual-core + GPU Iris Core i7 Broadwell-U (64-bit, SIMD, caches) 1,900,000,000[89] 2015 Intel 14 nm 133 mm2
Apple A9 (dual-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 2,000,000,000+ 2015 Apple 14 nm
(Samsung)		 96 mm2
(Samsung)
16 nm
(TSMC)		 104.5 mm2
(TSMC)
Apple A9X (dual core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 3,000,000,000+ 2015 Apple 16 nm 143.9 mm2
IBM z13 (64-bit, caches) 3,990,000,000 2015 IBM 22 nm 678 mm2
IBM z13 Storage Controller 7,100,000,000 2015 IBM 22 nm 678 mm2
32-core SPARC M7 (64-bit, SIMD, caches) 10,000,000,000[90] 2015 Oracle 20 nm  
Qualcomm Snapdragon 835 (octa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 3,000,000,000[91][92] 2016 Qualcomm 10 nm 72.3 mm2
10-core Core i7 Broadwell-E (64-bit, SIMD, caches) 3,200,000,000[93] 2016 Intel 14 nm 246 mm2[94]
Apple A10 Fusion (quad-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 3,300,000,000 2016 Apple 16 nm 125 mm2
HiSilicon Kirin 960 (octa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 4,000,000,000[95] 2016 Huawei 16 nm 110.00 mm2
22-core Xeon Broadwell-E5 (64-bit, SIMD, caches) 7,200,000,000[96] 2016 Intel 14 nm 456 mm2
72-core Xeon Phi (64-bit, 512-bit SIMD, caches) 8,000,000,000 2016 Intel 14 nm 683 mm2
Zip CPU (32-bit, for FPGAs) 1,286 6-LUTs[97] 2016 Gisselquist Technology    
Qualcomm Snapdragon 845 (octa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 5,300,000,000[98] 2017 Qualcomm 10 nm 94 mm2
Qualcomm Snapdragon 850 (octa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 5,300,000,000[99] 2017 Qualcomm 10 nm 94 mm2
Apple A11 Bionic (hexa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 4,300,000,000 2017 Apple 10 nm 89.23 mm2
Zeppelin SoC Ryzen (64-bit, SIMD, caches) 4,800,000,000[100] 2017 AMD 14 nm 192 mm2
Ryzen 5 1600 Ryzen (64-bit, SIMD, caches) 4,800,000,000[101] 2017 AMD 14 nm 213 mm2
Ryzen 5 1600 X Ryzen (64-bit, SIMD, caches) 4,800,000,000[102] 2017 AMD 14 nm 213 mm2
IBM z14 (64-bit, SIMD, caches) 6,100,000,000 2017 IBM 14 nm 696 mm2
IBM z14 Storage Controller (64-bit) 9,700,000,000 2017 IBM 14 nm 696 mm2
HiSilicon Kirin 970 (octa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 5,500,000,000[103] 2017 Huawei 10 nm 96.72 mm2
Xbox One X (Project Scorpio) main SoC (64-bit, SIMD, caches) 7,000,000,000[104] 2017 Microsoft/AMD 16 nm 360 mm2[104]
28-core Xeon Platinum 8180 (64-bit, SIMD, caches) 8,000,000,000[105][disputed – discuss] 2017 Intel 14 nm  
POWER9 (64-bit, SIMD, caches) 8,000,000,000 2017 IBM 14 nm 695 mm2
Freedom U500 Base Platform Chip (E51, 4×U54) RISC-V (64-bit, caches) 250,000,000[106] 2017 SiFive 28 nm ~30 mm2
SPARC64 XII (12-core 64-bit, SIMD, caches) 5,450,000,000[107] 2017 Fujitsu 20 nm 795 mm2
Apple A10X Fusion (hexa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 4,300,000,000[108] 2017 Apple 10 nm 96.40 mm2
Centriq 2400 (64/32-bit, SIMD, caches) 18,000,000,000[109] 2017 Qualcomm 10 nm 398 mm2
32-core AMD Epyc (64-bit, SIMD, caches) 19,200,000,000 2017 AMD 14 nm 768 mm2
Qualcomm Snapdragon 710 (octa-core ARM64 "mobile SoC", SIMD, caches) ? 2018 Qualcomm 10 nm ?
Qualcomm Snapdragon 675 (octa-core ARM64 "mobile SoC", SIMD, caches) ? 2018 Qualcomm 11 nm ?
Qualcomm Snapdragon 855 (octa-core ARM64 "mobile SoC", SIMD, caches) ? 2018 Qualcomm 7 nm 73.27 mm2
Qualcomm Snapdragon 8cx / SCX8180 (octa-core ARM64 "mobile SoC", SIMD, caches) 8,500,000,000[110] 2018 Qualcomm 7 nm 112 mm2
Apple A12 Bionic (hexa-core ARM64 "mobile SoC", SIMD, caches) 6,900,000,000[111][112] 2018 Apple 7 nm 83.27 mm2
HiSilicon Kirin 980 (octa-core ARM64 "mobile SoC", SIMD, caches) 6,900,000,000[113] 2018 Huawei 7 nm 74.13 mm2
HiSilicon Kirin 990 5G 10,300,000,000[114] 2019 Huawei 7 nm 113.31 mm2
HiSilicon Kirin 990 4G 8,000,000,000[115] 2019 Huawei 7 nm 90.00 mm2
HiSilicon Kirin 710 (octa-core ARM64 "mobile SoC", SIMD, caches) 5,500,000,000[116] 2018 Huawei 12 nm  
Apple A12X Bionic (octa-core 64/32-bit ARM64 "mobile SoC", SIMD, caches) 10,000,000,000[117] 2018 Apple 7 nm 122 mm2
Apple A13 (iPhone 11 Pro) 8,500,000,000[118][119] 2019 Apple 7 nm 98.48 mm2
Fujitsu A64FX (64/32-bit, SIMD, caches) 8,786,000,000[120] 2018[121] Fujitsu 7 nm  
Tegra Xavier SoC (64/32-bit) 9,000,000,000[122] 2018 Nvidia 12 nm 350 mm2
Samsung Exynos 9820 (octa-core ARM64 "mobile SoC", SIMD, caches) ? 2019 Samsung 8 nm 127 mm2
AMD Ryzen 7 3700X (64-bit, SIMD, caches, I/O die) 5,990,000,000[123][d] 2019 AMD 7&12 nm (TSMC) 199 (74+125) mm2
AMD Ryzen 9 3900X (64-bit, SIMD, caches, I/O die) 9,890,000,000[1][2] 2019 AMD 7 & 12 nm (TSMC) 273 mm2
AMD Epyc Rome (64-bit, SIMD, caches) 39,540,000,000[1][2] 2019 AMD 7 & 12 nm (TSMC) 1088 mm2
AWS Graviton2 (64-bit, 64-core ARM-based, SIMD, caches)[124][125] 30,000,000,000 2019 Amazon 7 nm  

GPUs[edit]

A graphics processing unit (GPU) is a specialized electronic circuit designed to rapidly manipulate and alter memory to accelerate the building of images in a frame buffer intended for output to a display.

The designer refers to the technology company that designs the logic of the integrated circuit chip (such as Nvidia and AMD). The manufacturer refers to the semiconductor company that fabricates the chip using its semiconductor manufacturing process at a foundry (such as TSMC and Samsung Semiconductor). The transistor count in a chip is dependent on a manufacturer's fabrication process, with smaller semiconductor nodes typically enabling higher transistor density and thus higher transistor counts.

The random-access memory (RAM) that comes with GPUs (such as VRAM, SGRAM or HBM) greatly increase the total transistor count, with the memory typically accounting for the majority of transistors in a graphics card. For example, Nvidia's Tesla P100 has 15 billion FinFETs (16 nm) in the GPU in addition to 16 GB of HBM2 memory, totaling about 150 billion MOSFETs on the graphics card.[126] The following table does not include the memory. For memory transistor counts, see the Memory section below.

Processor MOS transistor count Date of introduction Designer(s) Manufacturer(s) MOS process Area Ref
µPD7220 GDC 40,000 1982 NEC NEC 5,000 nm   [127]
ARTC HD63484 60,000 1984 Hitachi Hitachi     [128]
YM7101 VDP 100,000 1988 Sega Yamaha     [129]
Tom & Jerry 750,000 1993 Flare IBM     [129]
VDP1 1,000,000 1994 Sega Hitachi 500 nm   [130][131]
Sony GPU 1,000,000 1994 Toshiba LSI 500 nm   [132][133][134]
NV1 1,000,000 1995 Nvidia, Sega SGS 500 nm 90 mm2 [130]
Reality Coprocessor 2,600,000 1996 SGI NEC 350 nm 81 mm2 [135]
PowerVR 1,200,000 1996 VideoLogic NEC 350 nm   [136]
Voodoo Graphics 1,000,000 1996 3dfx TSMC 500 nm   [137][138]
Voodoo Rush 1,000,000 1997 3dfx TSMC 500 nm   [137][138]
NV3 3,500,000 1997 Nvidia SGS, TSMC 350 nm 90 mm2 [139][140]
PowerVR2 CLX2 10,000,000 1998 VideoLogic NEC 250 nm 116 mm2 [55][141][142][57]
i740 3,500,000 1998 Intel, Real3D Real3D 350 nm   [137][138]
Voodoo 2 4,000,000 1998 3dfx TSMC 350 nm  
Voodoo Rush 4,000,000 1998 3dfx TSMC 350 nm  
Riva TNT 7,000,000 1998 Nvidia TSMC 350 nm   [137][140]
PowerVR2 PMX1 6,000,000 1999 VideoLogic NEC 250 nm   [143]
Rage 128 8,000,000 1999 ATI TSMC, UMC 250 nm 70 mm2 [138]
Voodoo 3 8,100,000 1999 3dfx TSMC 250 nm   [144]
Graphics Synthesizer 43,000,000 1999 Sony, Toshiba Sony, Toshiba 180 nm 279 mm2 [62][60][59][61]
NV5 15,000,000 1999 Nvidia TSMC 250 nm   [138]
NV10 17,000,000 1999 Nvidia TSMC 220 nm 111 mm2 [145][140]
Voodoo 4 14,000,000 2000 3dfx TSMC 220 nm   [137][138]
NV11 20,000,000 2000 Nvidia TSMC 180 nm 65 mm2 [138]
NV15 25,000,000 2000 Nvidia TSMC 180 nm 81 mm2 [138]
Voodoo 5 28,000,000 2000 3dfx TSMC 220 nm   [137][138]
R100 30,000,000 2000 ATI TSMC 180 nm 97 mm2 [138]
Flipper 51,000,000 2000 ArtX NEC 180 nm 106 mm2 [62][146]
PowerVR3 KYRO 14,000,000 2001 Imagination ST 250 nm   [137][138]
PowerVR3 KYRO II 15,000,000 2001 Imagination ST 180 nm
NV2A 60,000,000 2001 Nvidia TSMC 150 nm   [137][147]
NV20 57,000,000 2001 Nvidia TSMC 150 nm 128 mm2 [138]
R200 60,000,000 2001 ATI TSMC 150 nm 68 mm2
NV25 63,000,000 2002 Nvidia TSMC 150 nm 142 mm2
R300 107,000,000 2002 ATI TSMC 150 nm 218 mm2
R360 117,000,000 2003 ATI TSMC 150 nm 218 mm2
NV38 135,000,000 2003 Nvidia TSMC 130 nm 207 mm2
R480 160,000,000 2004 ATI TSMC 130 nm 297 mm2
NV40 222,000,000 2004 Nvidia IBM 130 nm 305 mm2
Xenos 232,000,000 2005 ATI TSMC 90 nm 182 mm2 [148][149]
RSX Reality Synthesizer 300,000,000 2005 Nvidia, Sony Sony 90 nm 186 mm2 [150][151]
G70 303,000,000 2005 Nvidia TSMC, Chartered 110 nm 333 mm2 [138]
R520 321,000,000 2005 ATI TSMC 90 nm 288 mm2
R580 384,000,000 2006 ATI TSMC 90 nm 352 mm2
G80 681,000,000 2006 Nvidia TSMC 90 nm 480 mm2
G86 Tesla 210,000,000 2007 Nvidia TSMC 80 nm 127 mm2
G84 Tesla 289,000,000 2007 Nvidia TSMC 80 nm 169 mm2
R600 700,000,000 2007 ATI TSMC 80 nm 420 mm2
G92 754,000,000 2007 Nvidia TSMC, UMC 65 nm 324 mm2
G98 Tesla 210,000,000 2008 Nvidia TSMC 65 nm 86 mm2
RV710 242,000,000 2008 ATI TSMC 55 nm 73 mm2
G96 Tesla 314,000,000 2008 Nvidia TSMC 55 nm 121 mm2
G94 Tesla 505,000,000 2008 Nvidia TSMC 65 nm 240 mm2
RV730 514,000,000 2008 ATI TSMC 55 nm 146 mm2
RV670 666,000,000 2008 ATI TSMC 55 nm 192 mm2
RV770 956,000,000 2008 ATI TSMC 55 nm 256 mm2
RV790 959,000,000 2008 ATI TSMC 55 nm 282 mm2 [152][138]
GT200b Tesla 1,400,000,000 2008 Nvidia TSMC, UMC 55 nm 470 mm2 [138]
GT200 Tesla 1,400,000,000 2008 Nvidia TSMC 65 nm 576 mm2 [153][138]
GT218 Tesla 260,000,000 2009 Nvidia TSMC 40 nm 57 mm2 [138]
GT216 Tesla 486,000,000 2009 Nvidia TSMC 40 nm 100 mm2
GT215 Tesla 727,000,000 2009 Nvidia TSMC 40 nm 144 mm2
RV740 826,000,000 2009 ATI TSMC 40 nm 137 mm2
Juniper RV840 1,040,000,000 2009 ATI TSMC 40 nm 166 mm2
Cypress RV870 2,154,000,000 2009 ATI TSMC 40 nm 334 mm2 [154]
Cedar RV810 292,000,000 2010 AMD (formerly ATI) TSMC 40 nm 59 mm2 [138]
Redwood RV830 627,000,000 2010 AMD TSMC 40 nm 104 mm2
GF106 Fermi 1,170,000,000 2010 Nvidia TSMC 40 nm 238 mm2
Barts RV940 1,700,000,000 2010 AMD TSMC 40 nm 255 mm2
Cayman RV970 2,640,000,000 2010 AMD TSMC 40 nm 389 mm2
GF100 Fermi 3,200,000,000 March 2010 Nvidia TSMC 40 nm 526 mm2 [155]
GF110 Fermi 3,000,000,000 November 2010 Nvidia TSMC 40 nm 520 mm2 [155]
GF119 Fermi 292,000,000 2011 Nvidia TSMC 40 nm 79 mm2 [138]
Caicos RV910 370,000,000 2011 AMD TSMC 40 nm 67 mm2
GF108 Fermi 585,000,000 2011 Nvidia TSMC 40 nm 116 mm2
Turks RV930 716,000,000 2011 AMD TSMC 40 nm 118 mm2
GF104 Fermi 1,950,000,000 2011 Nvidia TSMC 40 nm 332 mm2
Tahiti 4,312,711,873 2011 AMD TSMC 28 nm 365 mm2 [156]
GK107 Kepler 1,270,000,000 2012 Nvidia TSMC 28 nm 118 mm2 [138]
Cape Verde 1,500,000,000 2012 AMD TSMC 28 nm 123 mm2
GK106 Kepler 2,540,000,000 2012 Nvidia TSMC 28 nm 221 mm2
Pitcairn 2,800,000,000 2012 AMD TSMC 28 nm 212 mm2
GK104 Kepler 3,540,000,000 2012 Nvidia TSMC 28 nm 294 mm2 [157]
GK110 Kepler 7,080,000,000 2012 Nvidia TSMC 28 nm 561 mm2 [158][159]
Oland 1,040,000,000 2013 AMD TSMC 28 nm 90 mm2 [138]
Bonaire 2,080,000,000 2013 AMD TSMC 28 nm 160 mm2
Durango (Xbox One) 5,000,000,000 2013 AMD TSMC 28 nm 363 mm2 [160]
Liverpool (PlayStation 4) Unknown 2013 AMD TSMC 28 nm 348 mm2 [161]
Hawaii 6,300,000,000 2013 AMD TSMC 28 nm 438 mm2 [138]
GM107 Maxwell 1,870,000,000 2014 Nvidia TSMC 28 nm 148 mm2
GM206 Maxwell 2,940,000,000 2014 Nvidia TSMC 28 nm 228 mm2
Tonga 5,000,000,000 2014 AMD TSMC, GlobalFoundries 28 nm 366 mm2
GM204 Maxwell 5,200,000,000 2014 Nvidia TSMC 28 nm 398 mm2
GM200 Maxwell 8,000,000,000 2015 Nvidia TSMC 28 nm 601 mm2
Fiji 8,900,000,000 2015 AMD TSMC 28 nm 596 mm2
Polaris 11 "Baffin" 3,000,000,000 2016 AMD Samsung, GlobalFoundries 14 nm 123 mm2 [138][162]
GP108 Pascal 4,400,000,000 2016 Nvidia TSMC 16 nm 200 mm2 [138]
Durango 2 (Xbox One S) 5,000,000,000 2016 AMD TSMC 16 nm 240 mm2 [163]
Neo (PlayStation 4 Pro) 5,700,000,000 2016 AMD TSMC 16 nm 325 mm2 [164]
Polaris 10 "Ellesmere" 5,700,000,000 2016 AMD Samsung, GlobalFoundries 14 nm 232 mm2 [165]
GP104 Pascal 7,200,000,000 2016 Nvidia TSMC 16 nm 314 mm2 [138]
GP100 Pascal 15,300,000,000 2016 Nvidia TSMC, Samsung 16 nm 610 mm2 [166]
GP108 Pascal 1,850,000,000 2017 Nvidia Samsung 14 nm 74 mm2 [138]
Polaris 12 "Lexa" 2,200,000,000 2017 AMD Samsung, GlobalFoundries 14 nm 101 mm2 [138][162]
GP107 Pascal 3,300,000,000 2017 Nvidia Samsung 14 nm 132 mm2 [138]
Scorpio (Xbox One X) 7,000,000,000 2017 AMD TSMC 16 nm 359 mm2 [167]
GP102 Pascal 11,800,000,000 2017 Nvidia TSMC, Samsung 16 nm 471 mm2 [138]
Vega 10 12,500,000,000 2017 AMD Samsung, GlobalFoundries 14 nm 484 mm2 [168]
GV100 Volta 21,100,000,000 2017 Nvidia TSMC 12 nm 815 mm2 [3]
TU106 Turing 10,800,000,000 2018 Nvidia TSMC 12 nm 445 mm2  
Vega 20 13,230,000,000 2018 AMD TSMC 7 nm 331 mm2 [138]
TU104 Turing 13,600,000,000 2018 Nvidia TSMC 12 nm 545 mm2  
TU102 Turing 18,600,000,000 2018 Nvidia TSMC 12 nm 754 mm2 [169]
TU117 Turing 4,700,000,000 2019 Nvidia TSMC 12 nm 200 mm2 [170]
TU116 Turing 6,600,000,000 2019 Nvidia TSMC 12 nm 284 mm2 [171]
Navi 14 6,400,000,000 2019 AMD TSMC 7 nm 158 mm2 [172]
Navi 10 10,300,000,000 2019 AMD TSMC 7 nm 251 mm2 [173]

FPGA[edit]

A field-programmable gate array (FPGA) is an integrated circuit designed to be configured by a customer or a designer after manufacturing.

FPGA MOS transistor count Date of introduction Designer Manufacturer MOS process Area Ref
Virtex 70,000,000 1997 Xilinx        
Virtex-E 200,000,000 1998 Xilinx      
Virtex-II 350,000,000 2000 Xilinx   130 nm  
Virtex-II PRO 430,000,000 2002 Xilinx      
Virtex-4 1,000,000,000 2004 Xilinx   90 nm  
Virtex-5 1,100,000,000 2006 Xilinx TSMC 65 nm   [174]
Stratix IV 2,500,000,000 2008 Altera TSMC 40 nm   [175]
Stratix V 3,800,000,000 2011 Altera TSMC 28 nm   [176]
Arria 10 5,300,000,000 2014 Altera TSMC 20 nm   [177]
Virtex-7 2000T 6,800,000,000 2011 Xilinx TSMC 28 nm   [178]
Stratix 10 SX 2800 17,000,000,000 TBD Intel Intel 14 nm 560 mm2 [179][180]
Virtex-Ultrascale VU440 20,000,000,000 Q1 2015 Xilinx TSMC 20 nm   [181][182]
Virtex-Ultrascale+ VU19P 35,000,000,000 2020 Xilinx TSMC 16 nm 900 mm2 [e] [183][184][185]
Versal VC1902 37,000,000,000 2H 2019 Xilinx TSMC 7 nm   [186][187][188]
Stratix 10 GX 10M 43,300,000,000 Q4 2019 Intel Intel 14 nm 1400 mm2 [e] [189][190]

Memory[edit]

See also: Random-access memory § Timeline, flash memory § Timeline, and read-only memory § Timeline

Semiconductor memory is an electronic data storage device, often used as computer memory, implemented on integrated circuits. Nearly all semiconductor memory since the 1970s have used MOSFETs (MOS transistors), replacing earlier bipolar junction transistors. There are two major types of semiconductor memory, random-access memory (RAM) and non-volatile memory (NVM). In turn, there are two major RAM types, dynamic random-access memory (DRAM) and static random-access memory (SRAM), as well as two major NVM types, flash memory and read-only memory (ROM).

Typical CMOS SRAM consists of six transistors per cell. For DRAM, 1T1C, which means one transistor and one capacitor structure, is common. Capacitor charged or not is used to store 1 or 0. For flash memory, the data is stored in floating gate, and the resistance of the transistor is sensed to interpret the data stored. Depending on how fine scale the resistance could be separated, one transistor could store up to 3-bits, meaning eight distinctive level of resistance possible per transistor. However, the fine the scale comes with cost of repeatability therefore reliability. Typically, low grade 2-bits MLC flash is used for flash drives, so a 16 GB flash drive contains roughly 64 billion transistors.

For SRAM chips, six-transistor cells (six transistors per bit) was the standard.[191] DRAM chips during the early 1970s had three-transistor cells (three transistors per bit), before single-transistor cells (one transistor per bit) became standard since the era of 4 Kb DRAM in the mid-1970s.[192][193] In single-level flash memory, each cell contains one floating-gate MOSFET (one transistor per bit),[194] whereas multi-level flash contains 2, 3 or 4 bits per transistor.

Flash memory chips are commonly stacked up in layers, up to 128-layer in production,[195] and 136-layer managed,[196] and available in end-user devices up to 69-layer from manufacturers.

Random-access memory (RAM)
Chip name Capacity (bits) RAM type Transistor count Date of introduction Manufacturer(s) MOS process Area Ref
N/A 1-bit SRAM (cell) 6 1963 Fairchild N/A N/A [197]
N/A 1-bit DRAM (cell) 1 1965 Toshiba N/A N/A [198][199]
? 8-bit SRAM (bipolar) 48 1965 SDS, Signetics ? ? [197]
SP95 16-bit SRAM (bipolar) 80 1965 IBM ? ? [200]
TMC3162 16-bit SRAM (TTL) 96 1966 Transitron N/A ? [193]
? ? SRAM (MOS) ? 1966 NEC ? ? [192]
256-bit DRAM (IC) 256 1968 Fairchild ? ? [193]
64-bit SRAM (PMOS) 384 1968 Fairchild ? ? [192]
144-bit SRAM (NMOS) 864 1968 NEC
1101 256-bit SRAM (PMOS) 1,536 1969 Intel 12,000 nm ? [201][202][203]
1102 1 Kb DRAM (PMOS) 3,072 1970 Intel, Honeywell ? ? [192]
1103 1 Kb DRAM (PMOS) 3,072 1970 Intel 8,000 nm 10 mm2 [204][191][205][193]
μPD403 1 Kb DRAM (NMOS) 3,072 1971 NEC ? ? [206]
? 2 Kb DRAM (PMOS) 6,144 1971 General Instrument ? 12.7 mm2 [207]
2102 1 Kb SRAM (NMOS) 6,144 1972 Intel ? ? [201][208]
? 8 Kb DRAM (PMOS) 8,192 1973 IBM ? 18.8 mm2 [207]
5101 1 Kb SRAM (CMOS) 6,144 1974 Intel ? ? [201]
2116 16 Kb DRAM (NMOS) 16,384 1975 Intel ? ? [209][193]
2114 4 Kb SRAM (NMOS) 24,576 1976 Intel ? ? [201][210]
? 4 Kb SRAM (CMOS) 24,576 1977 Toshiba ? ? [202]
64 Kb DRAM (NMOS) 65,536 1977 NTT ? 35.4 mm2 [207]
DRAM (VMOS) 65,536 1979 Siemens ? 25.2 mm2 [207]
16 Kb SRAM (CMOS) 98,304 1980 Hitachi, Toshiba ? ? [211]
256 Kb DRAM (NMOS) 262,144 1980 NEC 1,500 nm 41.6 mm2 [207]
NTT 1,000 nm 34.4 mm2 [207]
64 Kb SRAM (CMOS) 393,216 1980 Matsushita ? ? [211]
288 Kb DRAM 294,912 1981 IBM ? 25 mm2 [212]
64 Kb SRAM (NMOS) 393,216 1982 Intel 1,500 nm ? [211]
256 Kb SRAM (CMOS) 1,572,864 1984 Toshiba 1,200 nm ? [211][203]
8 Mb DRAM 8,388,608 January 5, 1984 Hitachi ? ? [213][214]
16 Mb DRAM (CMOS) 16,777,216 1987 NTT 700 nm 148 mm2 [207]
4 Mb SRAM (CMOS) 25,165,824 1990 NEC, Toshiba, Hitachi, Mitsubishi ? ? [211]
64 Mb DRAM (CMOS) 67,108,864 1991 Matsushita, Mitsubishi, Fujitsu, Toshiba 400 nm
KM48SL2000 16 Mb SDRAM 16,777,216 1992 Samsung ? ? [215][216]
? 16 Mb SRAM (CMOS) 100,663,296 1992 Fujitsu, NEC 400 nm ? [211]
256 Mb DRAM (CMOS) 268,435,456 1993 Hitachi, NEC 250 nm
1 Gb DRAM 1,073,741,824 January 9, 1995 NEC 250 nm ? [217][218]
Hitachi 160 nm ?
SDRAM 1,073,741,824 1996 Mitsubishi 150 nm ? [211]
SDRAM (SOI) 1,073,741,824 1997 Hyundai ? ? [219]
4 Gb DRAM (4-bit) 1,073,741,824 1997 NEC 150 nm ? [211]
DRAM 4,294,967,296 1998 Hyundai ? ? [219]
8 Gb SDRAM (DDR3) 8,589,934,592 April 2008 Samsung 50 nm ? [220]
16 Gb SDRAM (DDR3) 17,179,869,184 2008
32 Gb SDRAM (HBM2) 34,359,738,368 2016 Samsung 20 nm ? [221]
64 Gb SDRAM (HBM2) 68,719,476,736 2017
128 Gb SDRAM (DDR4) 137,438,953,472 2018 Samsung 10 nm ? [222]
? RRAM[223] (3DSoC)[224] ? 2019 Skywater[225] 90 nm ?  
Flash memory
Chip name Capacity (bits) Flash type FGMOS transistor count Date of introduction Manufacturer(s) MOS process Area Ref
? 256 Kb NOR 262,144 1985 Toshiba 2,000 nm ? [211]
1 Mb NOR 1,048,576 1989 Seeq, Intel ?
4 Mb NAND 4,194,304 1989 Toshiba 1,000 nm
16 Mb NOR 16,777,216 1991 Mitsubishi 600 nm
DD28F032SA 32 Mb NOR 33,554,432 1993 Intel ? 280 mm2 [201][226]
? 64 Mb NOR 67,108,864 1994 NEC 400 nm ? [211]
NAND 67,108,864 1996 Hitachi
128 Mb NAND 134,217,728 1996 Samsung, Hitachi ?
256 Mb NAND 268,435,456 1999 Hitachi, Toshiba 250 nm
512 Mb NAND 536,870,912 2000 Toshiba ? ? [227]
1 Gb 2-bit NAND 536,870,912 2001 Samsung ? ? [211]
Toshiba, SanDisk 160 nm ? [228]
2 Gb NAND 2,147,483,648 2002 Samsung, Toshiba ? ? [229][230]
8 Gb NAND 8,589,934,592 2004 Samsung 60 nm ? [229]
16 Gb NAND 17,179,869,184 2005 Samsung 50 nm ? [231]
32 Gb NAND 34,359,738,368 2006 Samsung 40 nm
THGAM 128 Gb Stacked NAND 128,000,000,000 April 2007 Toshiba 56 nm 252 mm2 [232]
THGBM 256 Gb Stacked NAND 256,000,000,000 2008 Toshiba 43 nm 353 mm2 [233]
THGBM2 1 Tb Stacked 4-bit NAND 256,000,000,000 2010 Toshiba 32 nm 374 mm2 [234]
KLMCG8GE4A 512 Gb Stacked 2-bit NAND 256,000,000,000 2011 Samsung ? 192 mm2 [235]
KLUFG8R1EM 4 Tb Stacked 3-bit V-NAND 1,365,333,333,504 2017 Samsung ? 150 mm2 [236]
eUFS (1 TB) 8 Tb Stacked 4-bit V-NAND 2,048,000,000,000 2019 Samsung ? 150 mm2 [4][237]
Read-only memory (ROM)
Chip name Capacity (bits) ROM type Transistor count Date of introduction Manufacturer(s) MOS process Area Ref
? ? PROM ? 1956 Arma N/A ? [238][239]
1 Kb ROM (MOS) 1,024 1965 General Microelectronics ? ? [240]
3301 1 Kb ROM (bipolar) 1,024 1969 Intel N/A ? [240]
1702 2 Kb EPROM (MOS) 2,048 1971 Intel ? 15 mm2 [241]
? 4 Kb ROM (MOS) 4,096 1974 AMD, General Instrument ? ? [240]
2708 8 Kb EPROM (MOS) 8,192 1975 Intel ? ? [201]
? 2 Kb EEPROM (MOS) 2,048 1976 Toshiba ? ? [242]
µCOM-43 ROM 16 Kb PROM (PMOS) 16,000 1977 NEC ? ? [243]
2716 16 Kb EPROM (TTL) 16,384 1977 Intel N/A ? [204][244]
EA8316F 16 Kb ROM (NMOS) 16,384 1978 Electronic Arrays ? 436 mm2 [240][245]
2732 32 Kb EPROM 32,768 1978 Intel ? ? [201]
2364 64 Kb ROM 65,536 1978 Intel ? ? [246]
2764 64 Kb EPROM 65,536 1981 Intel 3,500 nm ? [201][211]
27128 128 Kb EPROM 131,072 1982 Intel ?
27256 256 Kb EPROM (HMOS) 262,144 1983 Intel ? ? [201][247]
? 256 Kb EPROM (CMOS) 262,144 1983 Fujitsu ? ? [248]
512 Kb EPROM (NMOS) 524,288 1984 AMD 1,700 nm ? [211]
27512 512 Kb EPROM (HMOS) 524,288 1984 Intel ? ? [201][249]
? 1 Mb EPROM (CMOS) 1,048,576 1984 NEC 1,200 nm ? [211]
4 Mb EPROM (CMOS) 4,194,304 1987 Toshiba 800 nm
16 Mb EPROM (CMOS) 16,777,216 1990 NEC 600 nm
MROM 16,777,216 1995 AKM, Hitachi ? ? [218]

Transistor computers[edit]

Before transistors were invented, relays were used in early computers. The world's first working programmable, fully automatic digital computer,[250] the 1941 Z3 22-bit word length computer, had 2,600 relays, and operated at a clock frequency of about 4–5 Hz. The 1940 Complex Number Computer had fewer than 500 relays,[251] but it was not fully programmable.

The second generation of computers were transistor computers that featured boards filled with discrete transistors and magnetic memory cores. The experimental 1953 48-bit Transistor Computer, developed at the University of Manchester, is widely believed to be the first transistor computer to come into operation anywhere in the world (the prototype had 92 point-contact transistors and 550 diodes).[252] A later version the 1955 machine had a total of 250 junction transistors and 1300 point diodes. The Computer also used a small number of tubes in its clock generator, so it was not the first fully transistorized. The ETL Mark III, developed at the Electrotechnical Laboratory in 1956, may have been the first transistor-based electronic computer using the stored program method. It had about "130 point-contact transistors and about 1,800 germanium diodes were used for logic elements, and these were housed on 300 plug-in packages which could be slipped in and out.[253] The 1958 decimal architecture IBM 7070 was the first transistor computer to be fully programmable. It had about 30,000 alloy-junction germanium transistors and 22,000 germanium diodes, on approximately 14,000 Standard Modular System (SMS) cards. The 1959 MOBIDIC, short for "MOBIle DIgital Computer", at 12,000 pounds (6.0 short tons) mounted in the trailer of a semi-trailer truck, was a transistorized computer for battlefield data.

The third generation of computers used integrated circuits (ICs).[254] The 1962 15-bit Apollo Guidance Computer used "about 4,000 "Type-G" (3-input NOR gate) circuits" for about 12,000 transistors plus 32,000 resistors.[255] The first commercial IC-based computer was the IBM System/360 in 1964.[254] The 1965 12-bit PDP-8 CPU had 1409 transistors and over 10,000 diodes. It was not a microprocessor, as it used discrete transistors on many cards; but later microprocessors, such as the Intersil 6100 reimplemented it, see below.[256]

The next generation of computers were the microcomputers, also known as home computers or personal computers (PC), which used MOS microprocessors, in the 1970s. This list includes early transistorized computers (second generation) and IC-based computers (third generation) from the 1950s and 1960s.

Computer Transistor count Year Manufacturer Notes Ref
Transistor Computer 92 1953 University of Manchester Point-contact transistors [252]
TRADIC 700 1954 Bell Labs Point-contact transistors [252]
Transistor Computer (full size) 250 1955 University of Manchester Discrete point-contact transistors [252]
ETL Mark III 130 1956 Electrotechnical Laboratory Point-contact transistors [252][253]
Metrovick 950 200 1956 Metropolitan-Vickers Discrete junction transistors  
NEC NEAC-2201 600 1958 NEC Germanium transistors [257]
Hitachi MARS-1 1,000 1958 Hitachi   [258]
IBM 7070 30,000 1958 IBM Alloy-junction germanium transistors [259]
Matsushita MADIC-I 400 1959 Matsushita Bipolar transistors [260]
NEC NEAC-2203 2,579 1959 NEC   [261]
Toshiba TOSBAC-2100 5,000 1959 Toshiba   [262]
IBM 7090 50,000 1959 IBM Discrete germanium transistors [263]
PDP-1 2,700 1959 Digital Equipment Corporation Discrete transistors  
Mitsubishi MELCOM 1101 3,500 1960 Mitsubishi Germanium transistors [264]
M18 FADAC 1,600 1960 Autonetics Discrete transistors  
D-17B 1,521 1962 Autonetics Discrete transistors  
NEC NEAC-L2 16,000 1964 NEC Ge transistors [265]
IBM System/360 ? 1964 IBM Integrated circuits  
PDP-8/I 1409 1968 Digital Equipment Corporation 74 series TTL circuits  
Apollo Guidance Computer Block II 12,300 1966 Raytheon / MIT Instrumentation Laboratory 4,100 ICs, each containing a 3-transistor, 3-input NOR gate  

Logic functions[edit]

Transistor count for generic logic functions is based on static CMOS implementation.[266]

Function Transistor count Ref
NOT 2  
Buffer 4
NAND 2-input 4
NOR 2-input 4
AND 2-input 6
OR 2-input 6
NAND 3-input 6
NOR 3-input 6
XOR 2-input 6
XNOR 2-input 8
MUX 2-input with TG 6
MUX 4-input with TG 18
NOT MUX 2-input 8
MUX 4-input 24
1-bit adder full 28
1-bit adder–subtractor 48
AND-OR-INVERT 6 [267]
Latch, D gated 8  
Flip-flop, edge triggered dynamic D with reset 12
8-bit multiplier 3,000  
16-bit multiplier 9,000  
32-bit multiplier 21,000 [268]
small-scale integration 2–100 [269]
medium-scale integration 100–500 [269]
large-scale integration 500–20,000 [269]
very-large-scale integration 20,000–1,000,000 [269]
ultra-large scale integration >1,000,000  

Parallel systems[edit]

Historically, each processing element in earlier parallel systems—like all CPUs of that time—was a serial computer built out of multiple chips. As transistor counts per chip increases, each processing element could be built out of fewer chips, and then later each multi-core processor chip could contain more processing elements.[270]

Goodyear MPP: (1983?) 8 pixel processors per chip, 3,000 to 8,000 transistors per chip.[270]

Brunel University Scape (single-chip array-processing element): (1983) 256 pixel processors per chip, 120,000 to 140,000 transistors per chip.[270]

Cell Broadband Engine: (2006) with 9 cores per chip, had 234 million transistors per chip.[271]

Other devices[edit]

Device type Device name Transistor count Date of introduction Designer(s) Manufacturer(s) MOS process Area Ref
Deep learning engine / IPU[f] Colossus GC2 23,600,000,000 2018 Graphcore TSMC 16 nm ~800 mm2 [272][273][better source needed]
Deep learning engine / IPU Wafer Scale Engine 1,200,000,000,000 2019 Cerebras TSMC 16 nm 46,225 mm2 [5][6][7][8]

Transistor density

Semiconductor
device
fabrication
 

MOSFET scaling
(process nodes)

Future

The transistor density is the number of transistors that are fabricated per unit area, typically measured in terms of the number of transistors per square millimeter (mm2). The transistor density usually correlates with the gate length of a semiconductor node (also known as a semiconductor manufacturing process), typically measured in nanometers (nm). As of 2019, the semiconductor node with the highest transistor density is TSMC's 5 nanometer node, with 171.3 million transistors per square millimeter.[274]

MOSFET nodes[edit]

Further information: List of semiconductor scale examples

Semiconductor nodes
Node name Transistor density (transistors/mm2) Production year Process MOSFET Manufacturer(s) Ref
? ? 1960 20,000 nm PMOS Bell Labs [275][276]
? ? 1960 20,000 nm NMOS
? ? 1963 ? CMOS Fairchild [18]
? ? 1964 ? PMOS General Microelectronics [277]
? ? 1968 20,000 nm CMOS RCA [278]
? ? 1969 12,000 nm PMOS Intel [211][203]
? ? 1970 10,000 nm CMOS RCA [278]
? 300 1970 8,000 nm PMOS Intel [205][193]
? ? 1971 10,000 nm PMOS Intel [279]
? 480 1971 ? PMOS General Instrument [207]
? ? 1973 ? NMOS Texas Instruments [207]
? 220 1973 ? NMOS Mostek [207]
? ? 1973 7,500 nm NMOS NEC [24][23]
? ? 1973 6,000 nm PMOS Toshiba [25][280]
? ? 1976 5,000 nm NMOS Hitachi, Intel [207]
? ? 1976 5,000 nm CMOS RCA  
? ? 1976 4,000 nm NMOS Zilog  
? ? 1976 3,000 nm NMOS Intel [281]
? 1,850 1977 ? NMOS NTT [207]
? ? 1978 3,000 nm CMOS Hitachi [282]
? ? 1978 2,500 nm NMOS Texas Instruments [207]
? ? 1978 2,000 nm NMOS NEC, NTT
? 2,600 1979 ? VMOS Siemens
? 7,280 1979 1,000 nm NMOS NTT
? 7,620 1980 1,000 nm NMOS NTT
? ? 1983 2,000 nm CMOS Toshiba [211]
? ? 1983 1,500 nm CMOS Intel [207]
? ? 1983 1,200 nm CMOS Intel
? ? 1984 800 nm CMOS NTT
? ? 1987 700 nm CMOS Fujitsu
? ? 1989 600 nm CMOS Mitsubishi, NEC, Toshiba [211]
? ? 1989 500 nm CMOS Hitachi, Mitsubishi, NEC, Toshiba
? ? 1991 400 nm CMOS Matsushita, Mitsubishi, Fujitsu, Toshiba
? ? 1993 350 nm CMOS Sony
? ? 1993 250 nm CMOS Hitachi, NEC
3LM 32,000 1994 350 nm CMOS NEC [135]
? ? 1995 160 nm CMOS Hitachi [211]
? ? 1996 150 nm CMOS Mitsubishi
TSMC 180 nm ? 1998 180 nm CMOS TSMC [283]
CS80 ? 1999 180 nm CMOS Fujitsu [284]
? ? 1999 180 nm CMOS Intel, Sony, Toshiba [201][60]
CS85 ? 1999 170 nm CMOS Fujitsu [285]
Samsung 140 nm ? 1999 140 nm CMOS Samsung [211]
? ? 2001 130 nm CMOS Fujitsu, Intel [284][201]
Samsung 100 nm ? 2001 100 nm CMOS Samsung [211]
? ? 2002 90 nm CMOS Sony, Toshiba, Samsung [60][229]
CS100 ? 2003 90 nm CMOS Fujitsu [284]
Intel 90 nm 1,450,000 2004 90 nm CMOS Intel [286][201]
Samsung 80 nm ? 2004 80 nm CMOS Samsung [287]
? ? 2004 65 nm CMOS Fujitsu, Toshiba [288]
Samsung 60 nm ? 2004 60 nm CMOS Samsung [229]
TSMC 45 nm ? 2004 45 nm CMOS TSMC  
Elpida 90 nm ? 2005 90 nm CMOS Elpida Memory [289]
CS200 ? 2005 65 nm CMOS Fujitsu [290][284]
Samsung 50 nm ? 2005 50 nm CMOS Samsung [231]
Intel 65 nm 2,080,000 2006 65 nm CMOS Intel [286]
Samsung 40 nm ? 2006 40 nm CMOS Samsung [231]
Toshiba 56 nm ? 2007 56 nm CMOS Toshiba [232]
Matsushita 45 nm ? 2007 45 nm CMOS Matsushita [70]
Intel 45 nm 3,300,000 2008 45 nm CMOS Intel [291]
Toshiba 43 nm ? 2008 43 nm CMOS Toshiba [233]
TSMC 40 nm ? 2008 40 nm CMOS TSMC [292]
Toshiba 32 nm ? 2009 32 nm CMOS Toshiba [293]
Intel 32 nm 7,500,000 2010 32 nm CMOS Intel [291]
? ? 2010 20 nm CMOS Hynix, Samsung [294][231]
Intel 22 nm 15,300,000 2012 22 nm CMOS Intel [291]
IMFT 20 nm ? 2012 20 nm CMOS IMFT [295]
Toshiba 19 nm ? 2012 19 nm CMOS Toshiba
Hynix 16 nm ? 2013 16 nm FinFET SK Hynix [294]
TSMC 16 nm 28,880,000 2013 16 nm FinFET TSMC [296][297]
Samsung 10 nm 51,820,000 2013 10 nm FinFET Samsung [298][299]
Intel 14 nm 37,500,000 2014 14 nm FinFET Intel [291]
14LP 32,940,000 2015 14 nm FinFET Samsung [298]
TSMC 10 nm 52,510,000 2016 10 nm FinFET TSMC [296][300]
12LP 36,710,000 2017 12 nm FinFET GlobalFoundries, Samsung [162]
N7FF 96,500,000 2017 7 nm FinFET TSMC [301][302][303]
8LPP 61,180,000 2018 8 nm FinFET Samsung [298]
7LPE 95,300,000 2018 7 nm FinFET Samsung [302]
Intel 10 nm 100,760,000 2018 10 nm FinFET Intel [304]
5LPE 126,530,000 2018 5 nm FinFET Samsung [305][306]
N7FF+ 113,900,000 2019 7 nm FinFET TSMC [301][302]
CLN5FF 171,300,000 2019 5 nm FinFET TSMC [274]
TSMC 3 nm ? ? 3 nm ? TSMC [307]
Samsung 3 nm ? ? 3 nm GAAFET Samsung [308]

See also[edit]

Notes[edit]

  1. ^ Declassified 1998
  2. ^ 3,510 without depletion mode pull-up transistors
  3. ^ 6,813 without depletion mode pull-up transistors
  4. ^ 3,900,000,000 core chiplet die, 2,090,000,000 I/O die
  5. ^ Jump up to:a b Estimate
  6. ^ "Intelligence Processing Unit"

References[edit]

  1. ^ Jump up to:a b c Broekhuijsen, Niels (October 23, 2019). "AMD's 64-Core EPYC and Ryzen CPUs Stripped: A Detailed Inside Look". Retrieved October 24, 2019.
  2. ^ Jump up to:a b c Mujtaba, Hassan (October 22, 2019). "AMD 2nd Gen EPYC Rome Processors Feature A Gargantuan 39.54 Billion Transistors, IO Die Pictured in Detail". Retrieved October 24, 2019.
  3. ^ Jump up to:a b Durant, Luke; Giroux, Olivier; Harris, Mark; Stam, Nick (May 10, 2017). "Inside Volta: The World's Most Advanced Data Center GPU". Nvidia developer blog.
  4. ^ Jump up to:a b Manners, David (January 30, 2019). "Samsung makes 1TB flash eUFS module". Electronics Weekly. Retrieved June 23,2019.
  5. ^ Jump up to:a b Hruska, Joel (August 2019). "Cerebras Systems Unveils 1.2 Trillion Transistor Wafer-Scale Processor for AI". extremetech.com. Retrieved September 6, 2019.
  6. ^ Jump up to:a b Feldman, Michael (August 2019). "Machine Learning chip breaks new ground with waferscale integration". nextplatform.com. Retrieved September 6, 2019.
  7. ^ Jump up to:a b Cutress, Ian (August 2019). "Hot Chips 31 Live Blogs: Cerebras' 1.2 Trillion Transistor Deep Learning Processor". anandtech.com. Retrieved September 6, 2019.
  8. ^ Jump up to:a b "A Look at Cerebras Wafer-Scale Engine: Half Square Foot Silicon Chip". WikiChip Fuse. November 16, 2019. Retrieved December 2, 2019.
  9. ^ "John Gustafson's answer to How many individual transistors are in the world's most powerful supercomputer?". Quora. Retrieved August 22, 2019.
  10. ^ "13 Sextillion & Counting: The Long & Winding Road to the Most Frequently Manufactured Human Artifact in History". Computer History Museum. April 2, 2018. Retrieved July 28, 2019.
  11. ^ Jump up to:a b c Moskowitz, Sanford L. (2016). Advanced Materials Innovation: Managing Global Technology in the 21st century. John Wiley & Sons. pp. 165–168. ISBN 9780470508923.
  12. ^ "1960 - Metal Oxide Semiconductor (MOS) Transistor Demonstrated". The Silicon Engine. Computer History Museum.
  13. ^ "Who Invented the Transistor?". Computer History Museum. December 4, 2013. Retrieved July 20, 2019.
  14. ^ "Transistors Keep Moore's Law Alive". EETimes. December 12, 2018. Retrieved July 18, 2019.
  15. ^ Jump up to:a b "Tortoise of Transistors Wins the Race – CHM Revolution". Computer History Museum. Retrieved July 22, 2019.
  16. ^ Hittinger, William C. (1973). "Metal-Oxide-Semiconductor Technology". Scientific American. 229 (2): 48–59. Bibcode:1973SciAm.229b..48H. doi:10.1038/scientificamerican0873-48. ISSN 0036-8733. JSTOR 24923169.
  17. ^ Bassett, Ross Knox (2007). To the Digital Age: Research Labs, Start-up Companies, and the Rise of MOS Technology. Johns Hopkins University Press. p. 22. ISBN 9780801886393.
  18. ^ Jump up to:a b "1963: Complementary MOS Circuit Configuration is Invented". Computer History Museum. Retrieved July 6, 2019.
  19. ^ Jump up to:a b "1971: Microprocessor Integrates CPU Function onto a Single Chip". The Silicon Engine. Computer History Museum. Retrieved September 4, 2019.
  20. ^ Jump up to:a b Holt, Ray. "World's First Microprocessor". Retrieved March 5,2016. 1st fully integrated chip set microprocessor
  21. ^ Jump up to:a b "Alpha 21364 - Microarchitectures - Compaq - WikiChip". en.wikichip.org. Retrieved September 8, 2019.
  22. ^ Ryoichi Mori; Hiroaki Tajima; Morihiko Tajima; Yoshikuni Okada (October 1977). "Microprocessors in Japan". Euromicro Newsletter. 3 (4): 50–7. doi:10.1016/0303-1268(77)90111-0.
  23. ^ Jump up to:a b "NEC 751 (uCOM-4)". The Antique Chip Collector's Page. Archived from the original on May 25, 2011. Retrieved June 11,2010.
  24. ^ Jump up to:a b "1970s: Development and evolution of microprocessors"(PDF). Semiconductor History Museum of Japan. Retrieved June 27,2019.
  25. ^ Jump up to:a b "1973: 12-bit engine-control microprocessor (Toshiba)"(PDF). Semiconductor History Museum of Japan. Retrieved June 27,2019.
  26. ^ "Low Bandwidth Timeline – Semiconductor". Texas Instruments. Retrieved June 22, 2016.
  27. ^ "The MOS 6502 and the Best Layout Guy in the World". research.swtch.com. January 3, 2011. Retrieved September 3, 2019.
  28. ^ "Digital History: ZILOG Z8000 (APRIL 1979)". OLD-COMPUTERS.COM : The Museum. Retrieved June 19, 2019.
  29. ^ "Chip Hall of Fame: Motorola MC68000 Microprocessor". IEEE Spectrum. Institute of Electrical and Electronics Engineers. June 30, 2017. Retrieved June 19, 2019.
  30. ^ Microprocessors: 1971 to 1976 Christiansen
  31. ^ "Microprocessors 1976 to 1981". weber.edu. Retrieved August 9,2014.
  32. ^ "W65C816S 16-bit Core". www.westerndesigncenter.com. Retrieved September 12, 2017.
  33. ^ Jump up to:a b c d e Demone, Paul (November 9, 2000). "ARM's Race to World Domination". real world technologies. Retrieved July 20, 2015.
  34. ^ Hand, Tom. "The Harris RTX 2000 Microcontroller" (PDF). mpeforth.com. Retrieved August 9, 2014.
  35. ^ "Forth chips list". UltraTechnology. March 15, 2001. Retrieved August 9, 2014.
  36. ^ Koopman, Philip J. (1989). "4.4 Architecture of the Novix NC4016". Stack Computers: the new wave. Ellis Horwood Series in Computers and Their Applications. Carnegie Mellon University. ISBN 978-0745804187. Retrieved August 9, 2014.
  37. ^ "Fujitsu SPARC". cpu-collection.de. Retrieved June 30, 2019.
  38. ^ Jump up to:a b Kimura S, Komoto Y, Yano Y (1988). "Implementation of the V60/V70 and its FRM function". IEEE Micro. 8 (2): 22–36. doi:10.1109/40.527.
  39. ^ "VL2333 - VTI - WikiChip". en.wikichip.org. Retrieved August 31,2019.
  40. ^ Inayoshi H, Kawasaki I, Nishimukai T, Sakamura K (1988). "Realization of Gmicro/200". IEEE Micro. 8 (2): 12–21. doi:10.1109/40.526.
  41. ^ Bosshart, P.; Hewes, C.; Mi-Chang Chang; Kwok-Kit Chau; Hoac, C.; Houston, T.; Kalyan, V.; Lusky, S.; Mahant-Shetti, S.; Matzke, D.; Ruparel, K.; Ching-Hao Shaw; Sridhar, T.; Stark, D. (October 1987). "A 553K-Transistor LISP Processor Chip". IEEE Journal of Solid-State Circuits. 22 (5): 202–3. doi:10.1109/ISSCC.1987.1157084.
  42. ^ Fahlén, Lennart E.; Stockholm International Peace Research Institute (1987). "3. Hardware requirements for artificial intelligence § Lisp Machines: TI Explorer". Arms and Artificial Intelligence: Weapon and Arms Control Applications of Advanced Computing. SIPRI Monograph Series. Oxford University Press. p. 57. ISBN 978-0-19-829122-0.
  43. ^ Jouppi, Norman P.; Tang, Jeffrey Y. F. (July 1989). "A 20-MIPS Sustained 32-bit CMOS Microprocessor with High Ratio of Sustained to Peak Performance". IEEE Journal of Solid-State Circuits. 24 (5): i. Bibcode:1989IJSSC..24.1348J. CiteSeerX 10.1.1.85.988. doi:10.1109/JSSC.1989.572612. WRL Research Report 89/11.
  44. ^ "The CPU shack museum". CPUshack.com. May 15, 2005. Retrieved August 9, 2014.
  45. ^ Jump up to:a b c "Intel i960 Embedded Microprocessor". National High Magnetic Field Laboratory. Florida State University. March 3, 2003. Retrieved June 29, 2019.
  46. ^ Venkatasawmy, Rama (2013). The Digitization of Cinematic Visual Effects: Hollywood's Coming of Age. Rowman & Littlefield. p. 198. ISBN 9780739176214.
  47. ^ "SH Microprocessor Leading the Nomadic Era" (PDF). Semiconductor History Museum of Japan. Retrieved June 27, 2019.
  48. ^ "SH2: A Low Power RISC Micro for Consumer Applications"(PDF). Hitachi. Retrieved June 27, 2019.
  49. ^ "HARP-1: A 120 MHz Superscalar PA-RISC Processor" (PDF). Hitachi. Retrieved June 19, 2019.
  50. ^ "ARM7 Statistics". Poppyfields.net. May 27, 1994. Retrieved August 9, 2014.
  51. ^ "Forth Multiprocessor Chip MuP21". www.ultratechnology.com. Retrieved September 6, 2019. MuP21 has a 21-bit CPU core, a memory coprocessor, and a video coprocessor
  52. ^ Jump up to:a b "F21 CPU". www.ultratechnology.com. Retrieved September 6, 2019. F21 offers video I/O, analog I/O, serial network I/O, and a parallel I/O port on chip. F21 has a transistor count of about 15,000 vs about 7,000 for MuP21.
  53. ^ "Intel Pentium Pro 180". hw-museum.cz. Retrieved September 8, 2019.
  54. ^ "PC Guide Intel Pentium Pro ("P6")". PCGuide.com. April 17, 2001. Retrieved August 9, 2014.[dead link]
  55. ^ Jump up to:a b "Remembering the Sega Dreamcast". Bit-Tech. September 29, 2009. Retrieved June 18, 2019.
  56. ^ "Entertainment Systems and High-Performance Processor SH-4" (PDF). Hitachi Review. Hitachi. 48 (2): 58–63. 1999. Retrieved June 27, 2019.
  57. ^ Jump up to:a b Hagiwara, Shiro; Oliver, Ian (November–December 1999). "Sega Dreamcast: Creating a Unified Entertainment World". IEEE Micro. IEEE Computer Society. 19 (6): 29–35. doi:10.1109/40.809375. Retrieved June 27, 2019.
  58. ^ Ulf Samuelsson. "Transistor count of common uCs?". www.embeddedrelated.com. Retrieved September 8, 2019. IIRC, The AVR core is 12,000 gates, and the megaAVR core is 20,000 gates. Each gate is 4 transistors. The chip is considerably larger since the memory uses quite a lot.
  59. ^ Jump up to:a b Hennessy, John L.; Patterson, David A. (May 29, 2002). Computer Architecture: A Quantitative Approach (3 ed.). Morgan Kaufmann. p. 491. ISBN 978-0-08-050252-6. Retrieved April 9,2013.
  60. ^ Jump up to:a b c d "EMOTION ENGINE® AND GRAPHICS SYNTHESIZER USED IN THE CORE OF PLAYSTATION® BECOME ONE CHIP"(PDF). Sony. April 21, 2003. Retrieved June 26, 2019.
  61. ^ Jump up to:a b Diefendorff, Keith (April 19, 1999). "Sony's Emotionally Charged Chip: Killer Floating-Point "Emotion Engine" To Power PlayStation 2000" (PDF). Microprocessor Report. 13 (5). Retrieved June 19,2019.
  62. ^ Jump up to:a b c "NVIDIA GeForce 7800 GTX GPU Review". PC Perspective. June 22, 2005. Retrieved June 18, 2019.
  63. ^ Ando, H.; Yoshida, Y.; Inoue, A.; Sugiyama, I.; Asakawa, T.; Morita, K.; Muta, T.; otokurumada, T.; Okada, S.; Yamashita, H.; Satsukawa, Y.; Konmoto, A.; Yamashita, R.; Sugiyama, H. (2003). A 1.3GHz fifth eneration SPARC64 microprocessor. Design Automation Conference. pp. 702–705. doi:10.1145/775832.776010. ISBN 1-58113-688-9.
  64. ^ Krewell, Kevin (21 October 2002). "Fujitsu's SPARC64 V Is Real Deal". Microprocessor Report.
  65. ^ Fujitsu Limited (August 2004). SPARC64 V Processor For UNIX Server.
  66. ^ "A Glimpse Inside The Cell Processor". Gamasutra. July 13, 2006. Retrieved June 19, 2019.
  67. ^ "PRESS KIT — Dual-core Intel Itanium Processor". Intel. Retrieved August 9, 2014.
  68. ^ Jump up to:a b Toepelt, Bert (January 8, 2009). "AMD Phenom II X4: 45nm Benchmarked — The Phenom II And AMD's Dragon Platform". TomsHardware.com. Retrieved August 9, 2014.
  69. ^ "ARM (Advanced RISC Machines) Processors". EngineersGarage.com. Retrieved August 9, 2014.
  70. ^ Jump up to:a b "Panasonic starts to sell a New-generation UniPhier System LSI". Panasonic. October 10, 2007. Retrieved July 2, 2019.
  71. ^ "SPARC64 VI Extensions" page 56, Fujitsu Limited, Release 1.3, 27 March 2007
  72. ^ Morgan, Timothy Prickett (17 July 2008). "Fujitsu and Sun Flex Their Quads with New Sparc Server Lineup". The Unix Guardian, Vol. 8, No. 27.
  73. ^ Takumi Maruyama (2009). SPARC64 VIIIfx: Fujitsu's New Generation Octo Core Processor for PETA Scale computing(PDF). Proceedings of Hot Chips 21. IEEE Computer Society. Archived from the original (PDF) on October 8, 2010. Retrieved June 30, 2019.
  74. ^ Stokes, Jon (February 10, 2010). "Sun's 1 billion-transistor, 16-core Niagara 3 processor". ArsTechnica.com. Retrieved August 9,2014.
  75. ^ "IBM to Ship World's Fastest Microprocessor". IBM. September 1, 2010. Retrieved August 9, 2014.
  76. ^ "Intel to deliver first computer chip with two billion transistors". AFP. February 5, 2008. Archived from the original on May 20, 2011. Retrieved February 5, 2008.
  77. ^ "Intel Previews Intel Xeon 'Nehalem-EX' Processor." May 26, 2009. Retrieved on May 28, 2009.
  78. ^ Morgan, Timothy Prickett (November 21, 2011), "Fujitsu parades 16-core Sparc64 super stunner", The Register, retrieved December 8, 2011
  79. ^ Angelini, Chris (November 14, 2011). "Intel Core i7-3960X Review: Sandy Bridge-E And X79 Express". TomsHardware.com. Retrieved August 9, 2014.
  80. ^ "IDF2012 Mark Bohr, Intel Senior Fellow" (PDF).
  81. ^ "Images of SPARC64" (PDF). fujitsu.com. Retrieved August 29,2017.
  82. ^ "Intel's Atom Architecture: The Journey Begins". AnandTech. Retrieved April 4, 2010.
  83. ^ "Intel Xeon Phi SE10X". TechPowerUp. Retrieved July 20, 2015.
  84. ^ Shimpi, Lal. "The Haswell Review: Intel Core i7-4770K & i5-4670K Tested". anandtech. Retrieved November 20, 2014.
  85. ^ "Dimmick, Frank (August 29, 2014). "Intel Core i7 5960X Extreme Edition Review". Overclockers Club. Retrieved August 29, 2014.
  86. ^ "Apple A8X". NotebookCheck. Retrieved July 20, 2015.
  87. ^ "Intel Readying 15-core Xeon E7 v2". AnandTech. Retrieved August 9, 2014.
  88. ^ "Intel Xeon E5-2600 v3 Processor Overview: Haswell-EP Up to 18 Cores". pcper. Retrieved January 29, 2015.
  89. ^ "Intel's Broadwell-U arrives aboard 15W, 28W mobile processors". TechReport. Retrieved January 5, 2015.
  90. ^ http://www.enterprisetech.com/2014/08/13/oracle-cranks-cores-32-sparc-m7-chip/
  91. ^ "Qualcomm Snapdragon 835 (8998)". NotebookCheck. Retrieved September 23, 2017.
  92. ^ Takahashi, Dean (January 3, 2017). "Qualcomm's Snapdragon 835 will debut with 3 billion transistors and a 10nm manufacturing process". VentureBeat.
  93. ^ "Broadwell-E: Intel Core i7-6950X, 6900K, 6850K & 6800K Review". Tom's Hardware. May 30, 2016. Retrieved April 12,2017.
  94. ^ "The Broadwell-E Review". PC Gamer. July 8, 2016. Retrieved April 12, 2017.
  95. ^ "HUAWEI TO UNVEIL KIRIN 970 SOC WITH AI UNIT, 5.5 BILLION TRANSISTORS AND 1.2 GBPS LTE SPEED AT IFA 2017". firstpost.com. September 1, 2017. Retrieved November 18,2018.
  96. ^ "Broadwell-EP Architecture - Intel Xeon E5-2600 v4 Broadwell-EP Review". Tom's Hardware. March 31, 2016. Retrieved April 4,2016.
  97. ^ "About the ZipCPU". zipcpu.com. Retrieved September 10,2019. As of ORCONF, 2016, the ZipCPU used between 1286 and 4926 6-LUTs, depending upon how it is configured.
  98. ^ "Qualcomm Snapdragon 1000 for laptops could pack 8.5 billion transistors". techradar. Retrieved September 23, 2017.
  99. ^ "Spotted: Qualcomm Snapdragon 8cx Wafer on 7nm". AnandTech. Retrieved December 6, 2018.
  100. ^ Cutress, Ian (February 22, 2017). "AMD Launches Zen". Anandtech.com. Retrieved February 22, 2017.
  101. ^ "Ryzen 5 1600 - AMD". Wikichip.org. April 20, 2018. Retrieved December 9, 2018.
  102. ^ "Ryzen 5 1600X – AMD". Wikichip.org. October 26, 2018. Retrieved December 9, 2018.
  103. ^ "Kirin 970 – HiSilicon". Wikichip. March 1, 2018. Retrieved November 8, 2018.
  104. ^ Jump up to:a b Leadbetter, Richard (April 6, 2017). "Inside the next Xbox: Project Scorpio tech revealed". Eurogamer. Retrieved May 3,2017.
  105. ^ "Intel Xeon Platinum 8180". TechPowerUp. December 1, 2018. Retrieved December 2, 2018.
  106. ^ Lee, Y. "SiFive Freedom SoCs : Industry's First Open Source RISC V Chips" (PDF). HotChips 29 IOT/Embedded.
  107. ^ "Documents at Fujitsu" (PDF). fujitsu.com. Retrieved August 29,2017.
  108. ^ Schmerer, Kai (November 5, 2018). "iPad Pro 2018: A12X-Prozessor bietet deutlich mehr Leistung". ZDNet.de (in German).
  109. ^ "Qualcomm Datacenter Technologies Announces Commercial Shipment of Qualcomm Centriq 2400 – The World's First 10nm Server Processor and Highest Performance Arm-based Server Processor Family Ever Designed". Qualcomm. Retrieved November 9, 2017.
  110. ^ "Qualcomm Snapdragon 8180: 7nm SoC SDM1000 With 8.5 Billion Transistors To Challenge Apple A12 Bionic Chipset". dailyhunt. Retrieved September 21, 2018.
  111. ^ Yang, Daniel; Wegner, Stacy (September 21, 2018). "Apple iPhone Xs Max Teardown". TechInsights. Retrieved September 21, 2018.
  112. ^ "Apple's A12 Bionic is the first 7-nanometer smartphone chip". Engadget. Retrieved September 26, 2018.
  113. ^ "Kirin 980 – HiSilicon". Wikichip. November 8, 2018. Retrieved November 8, 2018.
  114. ^ Friedman, Alan. "5nm Kirin 1020 SoC tipped for next year's Huawei Mate 40 line". Phone Arena. Retrieved December 23,2019.
  115. ^ Frumusanu, Andrei. "The Huawei Mate 30 Pro Review: Top Hardware without Google?". AnandTech. Retrieved January 2,2020.
  116. ^ "HiSilicon Kirin 710". Notebookcheck. September 19, 2018. Retrieved November 24, 2018.
  117. ^ Zafar, Ramish (October 30, 2018). "Apple's A12X Has 10 Billion Transistors, 90% Performance Boost & 7-Core GPU". Wccftech.
  118. ^ Zafar, Ramish (September 10, 2019). "Apple A13 For iPhone 11 Has 8.5 Billion Transistors, Quad-Core GPU". Wccftech. Retrieved September 11, 2019.
  119. ^ Introducing iPhone 11 Pro — Apple Youtube Video, retrieved September 11, 2019
  120. ^ "Fujitsu began to produce Japan's billions of super-calculations with the strongest ARM processor A64FX". firstxw.com. April 16, 2019. Retrieved June 19, 2019.
  121. ^ "Fujitsu Successfully Triples the Power Output of Gallium-Nitride Transistors". Fujitsu. August 22, 2018. Retrieved June 19, 2019.
  122. ^ "Hot Chips 30: Nvidia Xavier SoC". fuse.wikichip.org. September 18, 2018. Retrieved December 6, 2018.
  123. ^ "AMD Ryzen 9 3900X and Ryzen 7 3700X Review: Zen 2 and 7nm Unleashed". Tom's Hardware. July 7, 2019. Retrieved October 19,2019.
  124. ^ CPUs, Arne Verheyde 2019-12-05T19:12:44Z. "Amazon Compares 64-core ARM Graviton2 to Intel's Xeon". Tom's Hardware. Retrieved December 6, 2019.
  125. ^ Morgan, Timothy Prickett (December 3, 2019). "Finally: AWS Gives Servers A Real Shot In The Arm". The Next Platform. Retrieved December 6, 2019.
  126. ^ Williams, Chris. "Nvidia's Tesla P100 has 15 billion transistors, 21TFLOPS". www.theregister.co.uk. Retrieved August 12, 2019.
  127. ^ "Famous Graphics Chips: NEC µPD7220 Graphics Display Controller". IEEE Computer Society. Institute of Electrical and Electronics Engineers. August 22, 2018. Retrieved June 21, 2019.
  128. ^ "GPU History: Hitachi ARTC HD63484. The second graphics processor". IEEE Computer Society. Institute of Electrical and Electronics Engineers. Retrieved June 21, 2019.
  129. ^ Jump up to:a b "30 Years of Console Gaming". Klinger Photography. August 20, 2017. Retrieved June 19, 2019.
  130. ^ Jump up to:a b "Diamond Edge 3D (nVidia NV1+Sega Saturn)". Naver. February 24, 2017. Retrieved June 19, 2019.
  131. ^ "Sega Saturn". MAME. Retrieved July 18, 2019.
  132. ^ "ASIC CHIPS ARE INDUSTRY'S GAME WINNERS". The Washington Post. September 18, 1995. Retrieved June 19, 2019.
  133. ^ "Is it Time to Rename the GPU?". Jon Peddie Research. IEEE Computer Society. July 9, 2018. Retrieved June 19, 2019.
  134. ^ "FastForward Sony Taps LSI Logic for PlayStation Video Game CPU Chip". FastForward. Retrieved January 29, 2014.
  135. ^ Jump up to:a b "Reality Co-Processor − The Power In Nintendo64" (PDF). Silicon Graphics. August 26, 1997. Retrieved June 18, 2019.
  136. ^ "Imagination PowerVR PCX2 GPU". VideoCardz.net. Retrieved June 19, 2019.
  137. ^ Jump up to:a b c d e f g h Lilly, Paul (May 19, 2009). "From Voodoo to GeForce: The Awesome History of 3D Graphics". PC Gamer. Retrieved June 19, 2019.
  138. ^ Jump up to:a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad "3D accelerator database". Vintage 3D. Retrieved July 21, 2019.
  139. ^ "RIVA128 Datasheet". SGS Thomson Microelectronics. Retrieved July 21, 2019.
  140. ^ Jump up to:a b c Singer, Graham (April 3, 2013). "History of the Modern Graphics Processor, Part 2". TechSpot. Retrieved July 21, 2019.
  141. ^ Weinberg, Neil (September 7, 1998). "Comeback kid". Forbes. Retrieved June 19, 2019.
  142. ^ Charles, Bertie (1998). "Sega's New Dimension". Forbes. Forbes Incorporated. 162 (5–9): 206. The chip, etched in 0.25-micron detail — state-of-the-art for graphics processors — fits 10 million transistors
  143. ^ "VideoLogic Neon 250 4MB". VideoCardz.net. Retrieved June 19, 2019.
  144. ^ Shimpi, Anand Lal (November 21, 1998). "Fall Comdex '98 Coverage". AnandTech. Retrieved June 19, 2019.
  145. ^ "NVIDIA NV10 A3 GPU Specs". TechPowerUp. Retrieved June 19, 2019.
  146. ^ IGN Staff (November 4, 2000). "Gamecube Versus PlayStation 2". IGN. Retrieved November 22, 2015.
  147. ^ "NVIDIA NV2A GPU Specs". TechPowerUp. Retrieved July 21,2019.
  148. ^ "ATI Xenos GPU Specs". TechPowerUp. Retrieved June 21,2019.
  149. ^ International, GamesIndustry (July 14, 2005). "TSMC to manufacture X360 GPU". Eurogamer. Retrieved August 22, 2006.
  150. ^ "NVIDIA Playstation 3 RSX 65nm Specs". TechPowerUp. Retrieved June 21, 2019.
  151. ^ "PS3 Graphics Chip Goes 65nm in Fall". Edge Online. June 26, 2008.
  152. ^ "The Radeon HD 4850 & 4870: AMD Wins at $199 and $299". AnandTech.com. Retrieved August 9, 2014.
  153. ^ "NVIDIA's 1.4 Billion Transistor GPU: GT200 Arrives as the GeForce GTX 280 & 260". AnandTech.com. Retrieved August 9,2014.
  154. ^ "Radeon 5870 specifications". AMD. Retrieved August 9, 2014.
  155. ^ Jump up to:a b Glaskowsky, Peter. "ATI and Nvidia face off-obliquely". CNET. Retrieved August 9, 2014.[dead link]
  156. ^ Woligroski, Don (December 22, 2011). "AMD Radeon HD 7970". TomsHardware.com. Retrieved August 9, 2014.
  157. ^ "Whitepaper: NVIDIA GeForce GTX 680" (PDF). NVIDIA. 2012. Archived from the original (PDF) on April 17, 2012.
  158. ^ http://www.nvidia.com/content/PDF/kepler/NVIDIA-Kepler-GK110-Architecture-Whitepaper.pdf
  159. ^ Smith, Ryan (November 12, 2012). "NVIDIA Launches Tesla K20 & K20X: GK110 Arrives At Last". AnandTech.
  160. ^ "AMD Xbox One GPU". www.techpowerup.com. Retrieved February 5, 2020.
  161. ^ "AMD PlayStation 4 GPU". www.techpowerup.com. Retrieved February 5, 2020.
  162. ^ Jump up to:a b c Schor, David (July 22, 2018). "VLSI 2018: GlobalFoundries 12nm Leading-Performance, 12LP". WikiChip Fuse. Retrieved May 31, 2019.
  163. ^ "AMD Xbox One S GPU". www.techpowerup.com. Retrieved February 5, 2020.
  164. ^ "AMD PlayStation 4 Pro GPU". www.techpowerup.com. Retrieved February 5, 2020.
  165. ^ Smith, Ryan (June 29, 2016). "The AMD RX 480 Preview". Anandtech.com. Retrieved February 22, 2017.
  166. ^ Harris, Mark (April 5, 2016). "Inside Pascal: NVIDIA's Newest Computing Platform". Nvidia developer blog.
  167. ^ "AMD Xbox One X GPU". www.techpowerup.com. Retrieved February 5, 2020.
  168. ^ "Radeon's next-generation Vega architecture" (PDF).
  169. ^ "NVIDIA TURING GPU ARCHITECTURE: Graphics Reinvented" (PDF). Nvidia. 2018. Retrieved June 28, 2019.
  170. ^ "NVIDIA GeForce GTX 1660 Ti". www.techpowerup.com. Retrieved February 5, 2020.
  171. ^ "NVIDIA GeForce GTX 1650". www.techpowerup.com. Retrieved February 5, 2020.
  172. ^ "AMD Radeon RX 5500 XT". www.techpowerup.com. Retrieved February 5, 2020.
  173. ^ "AMD Radeon RX 5700 XT". www.techpowerup.com. Retrieved February 5, 2020.
  174. ^ "Taiwan Company UMC Delivers 65nm FPGAs to Xilinx." SDA-ASIA Thursday, November 9, 2006.
  175. ^ ""Altera's new 40nm FPGAs — 2.5 billion transistors!". pldesignline.com.
  176. ^ "Altera unveils 28-nm Stratix V FPGA family". April 20, 2010. Retrieved April 20, 2010.
  177. ^ "Design of a High-Density SoC FPGA at 20nm" (PDF). 2014. Retrieved July 16, 2017.
  178. ^ Maxfield, Clive (October 2011). "New Xilinx Virtex-7 2000T FPGA provides equivalent of 20 million ASIC gates". EETimes. AspenCore. Retrieved September 4, 2019.
  179. ^ Greenhill, D.; Ho, R.; Lewis, D.; Schmit, H.; Chan, K. H.; Tong, A.; Atsatt, S.; How, D.; McElheny, P. (February 2017). "3.3 A 14nm 1GHz FPGA with 2.5D transceiver integration". 2017 IEEE International Solid-State Circuits Conference (ISSCC): 54–55. doi:10.1109/ISSCC.2017.7870257. ISBN 978-1-5090-3758-2.
  180. ^ "3.3 A 14nm 1GHz FPGA with 2.5D transceiver integration | DeepDyve". May 17, 2017. Archived from the original on May 17, 2017. Retrieved September 19, 2019.
  181. ^ Santarini, Mike (May 2014). "Xilinx Ships Industry's First 20-nm All Programmable Devices" (PDF). Xcell journal. No. 86. Xilinx. p. 14. Retrieved June 3, 2014.
  182. ^ Gianelli, Silvia (January 2015). "Xilinx Delivers the Industry's First 4M Logic Cell Device, Offering >50M Equivalent ASIC Gates and 4X More Capacity than Competitive Alternatives". www.xilinx.com. Retrieved August 22, 2019.
  183. ^ Sims, Tara (August 2019). "Xilinx Announces the World's Largest FPGA Featuring 9 Million System Logic Cells". www.xilinx.com. Retrieved August 22, 2019.
  184. ^ Verheyde, Arne (August 2019). "Xilinx Introduces World's Largest FPGA With 35 Billion Transistors". www.tomshardware.com. Retrieved August 23, 2019.
  185. ^ Cutress, Ian (August 2019). "Xilinx Announces World Largest FPGA: Virtex Ultrascale+ VU19P with 9m Cells". www.anandtech.com. Retrieved September 25, 2019.
  186. ^ Abazovic, Fuad (May 2019). "Xilinx 7nm Versal taped out last year". Retrieved September 30, 2019.
  187. ^ Cutress, Ian (August 2019). "Hot Chips 31 Live Blogs: Xilinx Versal AI Engine". Retrieved September 30, 2019.
  188. ^ Krewell, Kevin (August 2019). "Hot Chips 2019 highlights new AI strategies". Retrieved September 30, 2019.
  189. ^ Leibson, Steven (November 6, 2019). "Intel announces Intel Stratix 10 GX 10M FPGA, worlds highest capacity with 10.2 million logic elements". Retrieved November 7, 2019.
  190. ^ Verheyde, Arne (November 6, 2019). "Intel Introduces World's Largest FPGA With 43.3 Billion Transistors". Retrieved November 7, 2019.
  191. ^ Jump up to:a b The DRAM memory of Robert Dennard history-computer.com
  192. ^ Jump up to:a b c d "Late 1960s: Beginnings of MOS memory" (PDF). Semiconductor History Museum of Japan. January 23, 2019. Retrieved June 27, 2019.
  193. ^ Jump up to:a b c d e f "1970: Semiconductors compete with magnetic cores". Computer History Museum. Retrieved June 19, 2019.
  194. ^ "2.1.1 Flash Memory". TU Wien. Retrieved June 20, 2019.
  195. ^ Shilov, Anton. "SK Hynix Starts Production of 128-Layer 4D NAND, 176-Layer Being Developed". www.anandtech.com. Retrieved September 16, 2019.
  196. ^ "Samsung Begins Production of 100+ Layer Sixth-Generation V-NAND Flash". PC Perspective. August 11, 2019. Retrieved September 16, 2019.
  197. ^ Jump up to:a b "1966: Semiconductor RAMs Serve High-speed Storage Needs". Computer History Museum. Retrieved June 19, 2019.
  198. ^ "Specifications for Toshiba "TOSCAL" BC-1411". Old Calculator Web Museum. Archived from the original on July 3, 2017. Retrieved May 8, 2018.
  199. ^ "Toshiba "Toscal" BC-1411 Desktop Calculator". Old Calculator Web Museum. Archived from the original on May 20, 2007.
  200. ^ "IBM first in IC memory". Computer History Museum. Retrieved June 19, 2019.
  201. ^ Jump up to:a b c d e f g h i j k l m "A chronological list of Intel products. The products are sorted by date" (PDF). Intel museum. Intel Corporation. July 2005. Archived from the original (PDF) on August 9, 2007. Retrieved July 31, 2007.
  202. ^ Jump up to:a b "1970s: SRAM evolution" (PDF). Semiconductor History Museum of Japan. Retrieved June 27, 2019.
  203. ^ Jump up to:a b c Pimbley, J. (2012). Advanced CMOS Process Technology. Elsevier. p. 7. ISBN 9780323156806.
  204. ^ Jump up to:a b "Intel: 35 Years of Innovation (1968–2003)" (PDF). Intel. 2003. Retrieved June 26, 2019.
  205. ^ Jump up to:a b Lojek, Bo (2007). History of Semiconductor Engineering. Springer Science & Business Media. pp. 362–363. ISBN 9783540342588. The i1103 was manufactured on a 6-mask silicon-gate P-MOS process with 8 μm minimum features. The resulting product had a 2,400 µm2 memory cell size, a die size just under 10 mm2, and sold for around $21.
  206. ^ "Manufacturers in Japan enter the DRAM market and integration densities are improved" (PDF). Semiconductor History Museum of Japan. Retrieved June 27, 2019.
  207. ^ Jump up to:a b c d e f g h i j k l m n Gealow, Jeffrey Carl (August 10, 1990). "Impact of Processing Technology on DRAM Sense Amplifier Design" (PDF). CORE. Massachusetts Institute of Technology. pp. 149–166. Retrieved June 25, 2019.
  208. ^ "Silicon Gate MOS 2102A". Intel. Retrieved June 27, 2019.
  209. ^ "One of the Most Successful 16K Dynamic RAMs: The 4116". National Museum of American History. Smithsonian Institution. Retrieved June 20, 2019.
  210. ^ Component Data Catalog (PDF). Intel. 1978. pp. 3–94. Retrieved June 27, 2019.
  211. ^ Jump up to:a b c d e f g h i j k l m n o p q r s t "Memory". STOL (Semiconductor Technology Online). Retrieved June 25, 2019.
  212. ^ "The Cutting Edge of IC Technology: The First 294,912-Bit (288K) Dynamic RAM". National Museum of American History. Smithsonian Institution. Retrieved June 20, 2019.
  213. ^ "Computer History for 1984". Computer Hope. Retrieved June 25, 2019.
  214. ^ "Japanese Technical Abstracts". Japanese Technical Abstracts. University Microfilms. 2 (3–4): 161. 1987. The announcement of 1M DRAM in 1984 began the era of megabytes.
  215. ^ "KM48SL2000-7 Datasheet". Samsung. August 1992. Retrieved June 19, 2019.
  216. ^ "Electronic Design". Electronic Design. Hayden Publishing Company. 41 (15–21). 1993. The first commercial synchronous DRAM, the Samsung 16-Mbit KM48SL2000, employs a single-bank architecture that lets system designers easily transition from asynchronous to synchronous systems.
  217. ^ Breaking the gigabit barrier, DRAMs at ISSCC portend major system-design impact. (dynamic random access memory; International Solid-State Circuits Conference; Hitachi Ltd. and NEC Corp. research and development) Highbeam Business, January 9, 1995
  218. ^ Jump up to:a b "Japanese Company Profiles" (PDF). Smithsonian Institution. 1996. Retrieved June 27, 2019.
  219. ^ Jump up to:a b "History: 1990s". SK Hynix. Retrieved July 6, 2019.
  220. ^ "Samsung 50nm 2GB DDR3 chips are industry's smallest". SlashGear. September 29, 2008. Retrieved June 25, 2019.
  221. ^ Shilov, Anton (July 19, 2017). "Samsung Increases Production Volumes of 8 GB HBM2 Chips Due to Growing Demand". AnandTech. Retrieved June 29, 2019.
  222. ^ "Samsung Unleashes a Roomy DDR4 256GB RAM". Tom's Hardware. September 6, 2018. Retrieved June 21, 2019.
  223. ^ "First 3D Nanotube and RRAM ICs Come Out of Foundry". IEEE Spectrum: Technology, Engineering, and Science News. July 19, 2019. Retrieved September 16, 2019. This wafer was made just last Friday… and it's the first monolithic 3D IC ever fabricated within a foundry
  224. ^ "Three Dimensional Monolithic System-on-a-Chip". www.darpa.mil. Retrieved September 16, 2019.
  225. ^ "DARPA 3DSoC Initiative Completes First Year, Update Provided at ERI Summit on Key Steps Achieved to Transfer Technology into SkyWater's 200mm U.S. Foundry". Skywater Technology Foundry(Press release). July 25, 2019. Retrieved September 16, 2019.
  226. ^ "DD28F032SA Datasheet". Intel. Retrieved June 27, 2019.
  227. ^ "TOSHIBA ANNOUNCES 0.13 MICRON 1Gb MONOLITHIC NAND FEATURING LARGE BLOCK SIZE FOR IMPROVED WRITE/ERASE SPEED PERFORMANCE". Toshiba. September 9, 2002. Retrieved March 11, 2006.
  228. ^ "TOSHIBA AND SANDISK INTRODUCE A ONE GIGABIT NAND FLASH MEMORY CHIP, DOUBLING CAPACITY OF FUTURE FLASH PRODUCTS". Toshiba. November 12, 2001. Retrieved June 20, 2019.
  229. ^ Jump up to:a b c d "Our Proud Heritage from 2000 to 2009". Samsung Semiconductor. Samsung. Retrieved June 25, 2019.
  230. ^ "TOSHIBA ANNOUNCES 1 GIGABYTE COMPACTFLASH™CARD". Toshiba. September 9, 2002. Retrieved March 11, 2006.
  231. ^ Jump up to:a b c d "History". Samsung Electronics. Samsung. Retrieved June 19, 2019.
  232. ^ Jump up to:a b "TOSHIBA COMMERCIALIZES INDUSTRY'S HIGHEST CAPACITY EMBEDDED NAND FLASH MEMORY FOR MOBILE CONSUMER PRODUCTS". Toshiba. April 17, 2007. Retrieved November 23, 2010.
  233. ^ Jump up to:a b "Toshiba Launches the Largest Density Embedded NAND Flash Memory Devices". Toshiba. August 7, 2008. Retrieved June 21,2019.
  234. ^ "Toshiba Launches Industry's Largest Embedded NAND Flash Memory Modules". Toshiba. June 17, 2010. Retrieved June 21,2019.
  235. ^ "Samsung e·MMC Product family" (PDF). Samsung Electronics. December 2011. Retrieved July 15, 2019.
  236. ^ Shilov, Anton (December 5, 2017). "Samsung Starts Production of 512 GB UFS NAND Flash Memory: 64-Layer V-NAND, 860 MB/s Reads". AnandTech. Retrieved June 23, 2019.
  237. ^ Tallis, Billy (October 17, 2018). "Samsung Shares SSD Roadmap for QLC NAND And 96-layer 3D NAND". AnandTech. Retrieved June 27, 2019.
  238. ^ Han-Way Huang (December 5, 2008). Embedded System Design with C805. Cengage Learning. p. 22. ISBN 978-1-111-81079-5. Archived from the original on April 27, 2018.
  239. ^ Marie-Aude Aufaure; Esteban Zimányi (January 17, 2013). Business Intelligence: Second European Summer School, eBISS 2012, Brussels, Belgium, July 15-21, 2012, Tutorial Lectures. Springer. p. 136. ISBN 978-3-642-36318-4. Archived from the original on April 27, 2018.
  240. ^ Jump up to:a b c d "1965: Semiconductor Read-Only-Memory Chips Appear". Computer History Museum. Retrieved June 20, 2019.
  241. ^ "1971: Reusable semiconductor ROM introduced". The Storage Engine. Computer History Museum. Retrieved June 19, 2019.
  242. ^ Iizuka, H.; Masuoka, F.; Sato, Tai; Ishikawa, M. (1976). "Electrically alterable avalanche-injection-type MOS READ-ONLY memory with stacked-gate structure". IEEE Transactions on Electron Devices. 23(4): 379–387. Bibcode:1976ITED...23..379I. doi:10.1109/T-ED.1976.18415. ISSN 0018-9383.
  243. ^ µCOM-43 SINGLE CHIP MICROCOMPUTER: USERS' MANUAL (PDF). NEC Microcomputers. January 1978. Retrieved June 27, 2019.
  244. ^ "2716: 16K (2K x 8) UV ERASABLE PROM" (PDF). Intel. Retrieved June 27, 2019.
  245. ^ "1982 CATALOG" (PDF). NEC Electronics. Retrieved June 20,2019.
  246. ^ Component Data Catalog (PDF). Intel. 1978. pp. 1–3. Retrieved June 27, 2019.
  247. ^ "27256 Datasheet" (PDF). Intel. Retrieved July 2, 2019.
  248. ^ "History of Fujitsu's Semiconductor Business". Fujitsu. Retrieved July 2, 2019.
  249. ^ "D27512-30 Datasheet" (PDF). Intel. Retrieved July 2, 2019.
  250. ^ "A Computer Pioneer Rediscovered, 50 Years On". The New York Times. April 20, 1994. Archived from the original on November 4, 2016.
  251. ^ "History of Computers and Computing, Birth of the modern computer, Relays computer, George Stibitz". history-computer.com. Retrieved August 22, 2019. Initially the 'Complex Number Computer' performed only complex multiplication and division, but later a simple modification enabled it to add and subtract as well. It used about 400-450 binary relays, 6-8 panels, and ten multiposition, multipole relays called "crossbars" for temporary storage of numbers.
  252. ^ Jump up to:a b c d e "1953: Transistorized Computers Emerge". Computer History Museum. Retrieved June 19, 2019.
  253. ^ Jump up to:a b "ETL Mark III Transistor-Based Computer". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  254. ^ Jump up to:a b "Brief History". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  255. ^ "1962: Aerospace systems are first the applications for ICs in computers | The Silicon Engine | Computer History Museum". www.computerhistory.org. Retrieved September 2, 2019.
  256. ^ "PDP-8 (Straight 8) Computer Functional Restoration". www.pdp8.net. Retrieved August 22, 2019. backplanes contain 230 cards, approximately 10,148 diodes, 1409 transistors, 5615 resistors, and 1674 capacitors
  257. ^ "【NEC】 NEAC-2201". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  258. ^ "【Hitachi and Japanese National Railways】 MARS-1". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  259. ^ The IBM 7070 Data Processing System. Avery et al. (page 167)
  260. ^ "【Matsushita Electric Industrial】 MADIC-I transistor-based computer". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  261. ^ "【NEC】 NEAC-2203". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  262. ^ "【Toshiba】 TOSBAC-2100". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  263. ^ 7090 Data Processing System
  264. ^ "【Mitsubishi Electric】 MELCOM 1101". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  265. ^ "【NEC】NEAC-L2". IPSJ Computer Museum. Information Processing Society of Japan. Retrieved June 19, 2019.
  266. ^ Jan M. Rabaey, Digital Integrated Circuits, Fall 2001: Course Notes, Chapter 6: Designing Combinatorial Logic Gates in CMOS, retrieved October 27, 2012.
  267. ^ Richard F. Tinder (January 2000). Engineering Digital Design. Academic Press. ISBN 978-0-12-691295-1.
  268. ^ Asadi, P. (2007). "Energy-efficient 32 × 32-bit multiplier in tunable near-zero threshold CMOS" (PDF). World Applied Sciences Journal. 2 (4): 341–7. ISSN 1818-4952.
  269. ^ Jump up to:a b c d Engineers, Institute of Electrical Electronics (2000). IEEE Standard 100: The Authoritative Dictionary of IEEE Standards Terms(7th ed.). doi:10.1109/IEEESTD.2000.322230. ISBN 978-0-7381-2601-2. IEEE Std 100-2000.
  270. ^ Jump up to:a b c Smith, Kevin (August 11, 1983). "Image processor handles 256 pixels simultaneously". Electronics.
  271. ^ Kanellos, Michael (February 9, 2005). "Cell chip: Hit or hype?". CNET News. Archived from the original on October 25, 2012.
  272. ^ Kennedy, Patrick (June 2019). "Hands-on With a Graphcore C2 IPU PCIe Card at Dell Tech World". servethehome.com. Retrieved December 29, 2019.
  273. ^ "Colossus – Graphcore". en.wikichip.org. Retrieved December 29, 2019.
  274. ^ Jump up to:a b Schor, David (April 6, 2019). "TSMC Starts 5-Nanometer Risk Production". WikiChip Fuse. Retrieved April 7, 2019.
  275. ^ "1960: Metal Oxide Semiconductor (MOS) Transistor Demonstrated". Computer History Museum. Retrieved July 17,2019.
  276. ^ Lojek, Bo (2007). History of Semiconductor Engineering. Springer Science & Business Media. pp. 321–3. ISBN 9783540342588.
  277. ^ "1964: First Commercial MOS IC Introduced". Computer History Museum. Retrieved July 17, 2019.
  278. ^ Jump up to:a b Lojek, Bo (2007). History of Semiconductor Engineering. Springer Science & Business Media. p. 330. ISBN 9783540342588.
  279. ^ Lambrechts, Wynand; Sinha, Saurabh; Abdallah, Jassem Ahmed; Prinsloo, Jaco (2018). Extending Moore's Law through Advanced Semiconductor Design and Processing Techniques. CRC Press. p. 59. ISBN 9781351248655.
  280. ^ Belzer, Jack; Holzman, Albert G.; Kent, Allen (1978). Encyclopedia of Computer Science and Technology: Volume 10 – Linear and Matrix Algebra to Microorganisms: Computer-Assisted Identification. CRC Press. p. 402. ISBN 9780824722609.
  281. ^ "Intel® Microprocessor Quick Reference Guide". Intel. Retrieved June 27, 2019.
  282. ^ "1978: Double-well fast CMOS SRAM (Hitachi)" (PDF). Semiconductor History Museum of Japan. Retrieved July 5, 2019.
  283. ^ "0.18-micron Technology". TSMC. Retrieved June 30, 2019.
  284. ^ Jump up to:a b c d 65nm CMOS Process Technology
  285. ^ Diefendorff, Keith (15 November 1999). "Hal Makes Sparcs Fly". Microprocessor Report, Volume 13, Number 5.
  286. ^ Jump up to:a b Cutress, Ian. "Intel's 10nm Cannon Lake and Core i3-8121U Deep Dive Review". AnandTech. Retrieved June 19, 2019.
  287. ^ "Samsung Shows Industry's First 2-Gigabit DDR2 SDRAM". Samsung Semiconductor. Samsung. September 20, 2004. Retrieved June 25, 2019.
  288. ^ Williams, Martyn (July 12, 2004). "Fujitsu, Toshiba begin 65nm chip trial production". InfoWorld. Retrieved June 26, 2019.
  289. ^ Elpida's presentation at Via Technology Forum 2005 and Elpida 2005 Annual Report
  290. ^ Fujitsu Introduces World-class 65-Nanometer Process Technology for Advanced Server, Mobile Applications
  291. ^ Jump up to:a b c d "Intel Now Packs 100 Million Transistors in Each Square Millimeter". IEEE Spectrum: Technology, Engineering, and Science News. Retrieved November 14, 2018.
  292. ^ "40nm Technology". TSMC. Retrieved June 30, 2019.
  293. ^ "Toshiba Makes Major Advances in NAND Flash Memory with 3-bit-per-cell 32nm generation and with 4-bit-per-cell 43nm technology". Toshiba. February 11, 2009. Retrieved June 21,2019.
  294. ^ Jump up to:a b "History: 2010s". SK Hynix. Retrieved July 8, 2019.
  295. ^ Shimpi, Anand Lal (June 8, 2012). "SandForce Demos 19nm Toshiba & 20nm IMFT NAND Flash". AnandTech. Retrieved June 19, 2019.
  296. ^ Jump up to:a b Schor, David (April 16, 2019). "TSMC Announces 6-Nanometer Process". WikiChip Fuse. Retrieved May 31, 2019.
  297. ^ "16/12nm Technology". TSMC. Retrieved June 30, 2019.
  298. ^ Jump up to:a b c "VLSI 2018: Samsung's 8nm 8LPP, a 10nm extension". WikiChip Fuse. July 1, 2018. Retrieved May 31, 2019.
  299. ^ "Samsung Mass Producing 128Gb 3-bit MLC NAND Flash". Tom's Hardware. April 11, 2013. Retrieved June 21, 2019.
  300. ^ "10nm Technology". TSMC. Retrieved June 30, 2019.
  301. ^ Jump up to:a b Jones, Scotten (May 3, 2019). "TSMC and Samsung 5nm Comparison". Semiwiki. Retrieved July 30, 2019.
  302. ^ Jump up to:a b c Nenni, Daniel (January 2, 2019). "Samsung vs TSMC 7nm Update". Semiwiki. Retrieved July 6, 2019.
  303. ^ "7nm Technology". TSMC. Retrieved June 30, 2019.
  304. ^ Schor, David (June 15, 2018). "A Look at Intel's 10nm Std Cell as TechInsights Reports on the i3-8121U, finds Ruthenium". WikiChip Fuse. Retrieved May 31, 2019.
  305. ^ Jones, Scotten, 7nm, 5nm and 3nm Logic, current and projected processes
  306. ^ Shilov, Anton. "Samsung Completes Development of 5nm EUV Process Technology". AnandTech. Retrieved May 31, 2019.
  307. ^ "TSMC Plans New Fab for 3nm". EE Times. December 12, 2016. Retrieved September 26, 2019.
  308. ^ Armasu, Lucian (January 11, 2019), "Samsung Plans Mass Production of 3nm GAAFET Chips in 2021", www.tomshardware.com
posted @ 2020-03-09 11:00  张同光  阅读(526)  评论(0)  编辑  收藏  举报