Computer performance by orders of magnitude

This list compares various amounts of computing power in instructions per second organized by order of magnitude in FLOPS. == Milliscale computing (10−3) == 2 × 10−3: average human multiplication of two 10-digit numbers using pen and paper without aids == Deciscale computing (10−1) == 1 × 10−1: multiplication of two 10-digit numbers by a 1940s electromechanical desk calculator 3 × 10−1: multiplication on Zuse Z3 and Z4, first programmable digital computers, 1941 and 1945 respectively 5 × 10−1: computing power of the average human mental calculation for multiplication using pen and paper == Scale computing (100) == 1.2 OP/S: addition on Z3, 1941, and multiplication on Bell Model V, 1946 2.4 OP/S: addition on Z4, 1945 == Decascale computing (101) == 1.8 × 101: ENIAC, first programmable electronic digital computer, 1945 5 × 101: upper end of serialized human perception computation (light bulbs do not flicker to the human observer) 7 × 101: Whirlwind I 1951 vacuum tube computer and IBM 1620 1959 transistorized scientific minicomputer == Hectoscale computing (102) == 1.3 × 102: PDP-4 commercial minicomputer, 1962 2 × 102: IBM 602 electromechanical calculator (then called computer), 1946 6 × 102: Manchester Mark 1 electronic general-purpose stored-program digital computer, 1949 == Kiloscale computing (103) == 2 × 103: UNIVAC I, first American commercially available electronic general-purpose stored program digital computer, 1951 3 × 103: PDP-1 commercial minicomputer, 1959 15 × 103: IBM Naval Ordnance Research Calculator, 1954 24 × 103: AN/FSQ-7 Combat Direction Central, 1957 30 × 103: IBM 1130 commercial minicomputer, 1965 40 × 103: multiplication on Hewlett-Packard 9100A early desktop electronic calculator, 1968 53 × 103: Lincoln TX-2 transistor-based computer, 1958 92 × 103: Intel 4004, first commercially available full function CPU on a chip, released in 1971 5 × 103: Colossus computer vacuum tube cryptanalytic supercomputer, 1943 == Megascale computing (106) == 1 × 106: i486 microprocessor at 25 MHz using Linpack, 1989 1.2 × 106: IBM 7030 "Stretch" transistorized supercomputer, 1961 3 × 106: computing power of the 25 MHz Motorola 68040 using Linpack, 1990 5 × 106: CDC 6600, first commercially successful supercomputer, 1964 11 × 106: Intel i386 microprocessor at 33 MHz, 1985 14 × 106: CDC 7600 supercomputer, 1967 86 × 106: Cray 1 supercomputer, 1978 100 × 106: Pentium (i586) microprocessor, 1993 400 × 106: Cray X-MP, 1982 == Gigascale computing (109) == 1 × 109: ILLIAC IV 1972 supercomputer does first computational fluid dynamics problems 1.3 × 109: NEC SX supercomputer, 1983 1.4 × 109: Intel Pentium III microprocessor, 1999 1.6 × 109: PowerVR MBX Lite 3D GPU on iPhone 1, 2007 8 × 109: PowerVR SGX535 GPU on iPad 1, 2010 136 × 109: PowerVR GXA6450 GPU on iPhone 6 and iPhone SE, 2014 148 × 109: Intel Core i7-980X Extreme Edition commercial computing 2010 == Terascale computing (1012) == 1 × 1012: NEC SX-4 supercomputer, 1994 1.34 × 1012: Intel ASCI Red 1997 supercomputer 1.344 × 1012 GeForce GTX 480 in 2010 from Nvidia at its peak performance 2.15 × 1012: iPhone 15 Pro September 2023 A17 Pro processor 4.64 × 1012: Radeon HD 5970 in 2009 from AMD (under ATI branding) at its peak performance 5.152 × 1012: S2050/S2070 1U GPU Computing System from Nvidia 11.3 × 1012: GeForce GTX 1080 Ti in 2017 13.7 × 1012: Radeon RX Vega 64 in 2017 15.0 × 1012: Nvidia Titan V in 2017 80 × 1012: IBM Watson 170 × 1012: Nvidia DGX-1 The initial Pascal based DGX-1 delivered 170 teraflops of half precision processing. 478.2 × 1012 IBM BlueGene/L 2007 Supercomputer 960 × 1012 Nvidia DGX-1 The Volta-based upgrade increased calculation power of Nvidia DGX-1 to 960 teraflops.

Source: Wikipedia — Computer performance by orders of magnitude (CC BY-SA 4.0)

Computer performance by orders of magnitude

This list compares various amounts of computing power in instructions per second organized by order of magnitude in FLOPS. == Milliscale computing (10−3) == 2 × 10−3: average human multiplication of two 10-digit numbers using pen and paper without aids == Deciscale computing (10−1) == 1 × 10−1: multiplication of two 10-digit numbers by a 1940s electromechanical desk calculator 3 × 10−1: multiplication on Zuse Z3 and Z4, first programmable digital computers, 1941 and 1945 respectively 5 × 10−1: computing power of the average human mental calculation for multiplication using pen and paper == Scale computing (100) == 1.2 OP/S: addition on Z3, 1941, and multiplication on Bell Model V, 1946 2.4 OP/S: addition on Z4, 1945 == Decascale computing (101) == 1.8 × 101: ENIAC, first programmable electronic digital computer, 1945 5 × 101: upper end of serialized human perception computation (light bulbs do not flicker to the human observer) 7 × 101: Whirlwind I 1951 vacuum tube computer and IBM 1620 1959 transistorized scientific minicomputer == Hectoscale computing (102) == 1.3 × 102: PDP-4 commercial minicomputer, 1962 2 × 102: IBM 602 electromechanical calculator (then called computer), 1946 6 × 102: Manchester Mark 1 electronic general-purpose stored-program digital computer, 1949 == Kiloscale computing (103) == 2 × 103: UNIVAC I, first American commercially available electronic general-purpose stored program digital computer, 1951 3 × 103: PDP-1 commercial minicomputer, 1959 15 × 103: IBM Naval Ordnance Research Calculator, 1954 24 × 103: AN/FSQ-7 Combat Direction Central, 1957 30 × 103: IBM 1130 commercial minicomputer, 1965 40 × 103: multiplication on Hewlett-Packard 9100A early desktop electronic calculator, 1968 53 × 103: Lincoln TX-2 transistor-based computer, 1958 92 × 103: Intel 4004, first commercially available full function CPU on a chip, released in 1971 5 × 103: Colossus computer vacuum tube cryptanalytic supercomputer, 1943 == Megascale computing (106) == 1 × 106: i486 microprocessor at 25 MHz using Linpack, 1989 1.2 × 106: IBM 7030 "Stretch" transistorized supercomputer, 1961 3 × 106: computing power of the 25 MHz Motorola 68040 using Linpack, 1990 5 × 106: CDC 6600, first commercially successful supercomputer, 1964 11 × 106: Intel i386 microprocessor at 33 MHz, 1985 14 × 106: CDC 7600 supercomputer, 1967 86 × 106: Cray 1 supercomputer, 1978 100 × 106: Pentium (i586) microprocessor, 1993 400 × 106: Cray X-MP, 1982 == Gigascale computing (109) == 1 × 109: ILLIAC IV 1972 supercomputer does first computational fluid dynamics problems 1.3 × 109: NEC SX supercomputer, 1983 1.4 × 109: Intel Pentium III microprocessor, 1999 1.6 × 109: PowerVR MBX Lite 3D GPU on iPhone 1, 2007 8 × 109: PowerVR SGX535 GPU on iPad 1, 2010 136 × 109: PowerVR GXA6450 GPU on iPhone 6 and iPhone SE, 2014 148 × 109: Intel Core i7-980X Extreme Edition commercial computing 2010 == Terascale computing (1012) == 1 × 1012: NEC SX-4 supercomputer, 1994 1.34 × 1012: Intel ASCI Red 1997 supercomputer 1.344 × 1012 GeForce GTX 480 in 2010 from Nvidia at its peak performance 2.15 × 1012: iPhone 15 Pro September 2023 A17 Pro processor 4.64 × 1012: Radeon HD 5970 in 2009 from AMD (under ATI branding) at its peak performance 5.152 × 1012: S2050/S2070 1U GPU Computing System from Nvidia 11.3 × 1012: GeForce GTX 1080 Ti in 2017 13.7 × 1012: Radeon RX Vega 64 in 2017 15.0 × 1012: Nvidia Titan V in 2017 80 × 1012: IBM Watson 170 × 1012: Nvidia DGX-1 The initial Pascal based DGX-1 delivered 170 teraflops of half precision processing. 478.2 × 1012 IBM BlueGene/L 2007 Supercomputer 960 × 1012 Nvidia DGX-1 The Volta-based upgrade increased calculation power of Nvidia DGX-1 to 960 teraflops.

Source: Wikipedia "Computer performance by orders of magnitude" · CC BY-SA 4.0

Share this article: X · Bluesky
Privacy Policy