This article needs additional citations for verification. (February 2015) (Learn how and when to remove this template message)
In computing, floating point operations per second (FLOPS, flops or flop/s) is a measure of computer performance, useful in fields of scientific computations that require floating-point calculations. For such cases it is a more accurate measure than measuring instructions per second.
The similar term FLOP is often used for floating-point operation, for example as a unit of counting floating-point operations carried out by an algorithm or computer hardware.
Floating-point arithmetic is needed for very large or very small real numbers, or computations that require a large dynamic range. Floating-point representation is similar to scientific notation, except everything is carried out in base two, rather than base ten. The encoding scheme stores the sign, the exponent (in base two for Cray and VAX, base two or ten for IEEE floating point formats, and base 16 for IBM Floating Point Architecture) and the Significand (number after the radix point). While several similar formats are in use, the most common is ANSI/IEEE Std. 754-1985. This standard defines the format for 32-bit numbers called single precision, as well as 64-bit numbers called double precision and longer numbers called extended precision (used for intermediate results). Floating-point representations can support a much wider range of values than fixed-point, with the ability to represent very small numbers and very large numbers.
Dynamic range and precision
The exponentiation inherent in floating-point computation assures a much larger dynamic range – the largest and smallest numbers that can be represented – which is especially important when processing data sets where some of the data may have extremely large range of numerical values or where the range may be unpredictable. As such, floating-point processors are ideally suited for computationally intensive applications.
FLOPS and MIPS are units of measure for the numerical computing performance of a computer. Floating-point operations are typically used in fields such as scientific computational research. The unit MIPS measures integer performance of a computer. Examples of integer operation include data movement (A to B) or value testing (If A = B, then C). MIPS as a performance benchmark is adequate when a computer is used in database queries, word processing, spreadsheets, or to run multiple virtual operating systems. Frank H. McMahon, of the Lawrence Livermore National Laboratory, invented the terms FLOPS and MFLOPS (megaFLOPS) so that he could compare the supercomputers of the day by the number of floating-point calculations they performed per second. This was much better than using the prevalent MIPS to compare computers as this statistic usually had little bearing on the arithmetic capability of the machine.
FLOPS on an HPC-system can be calculated using this equation:
This can be simplified to the most common case: a computer that has exactly 1 CPU:
FLOPs per cycle for various processors
|Intel Atom (Bonnell, Saltwell, Silvermont and Goldmont)||SSE3 (64-bit)||2||4||0|
|Intel Core (Merom, Penryn)
Intel Nehalem (Nehalem, Westmere)
|Intel Sandy Bridge (Sandy Bridge, Ivy Bridge)||AVX (256-bit)||8||16||0|
|Intel Haswell (Haswell, Devil's Canyon, Broadwell)
Intel Skylake (Skylake, Kaby Lake, Coffee Lake, Whiskey lake, Amber lake)
|AVX2 & FMA (256-bit)||16||32||0|
|Intel Xeon Phi (Knights Corner)||SSE & FMA (256-bit)||16||32||0|
Intel Xeon Phi (Knights Landing, Knights Mill)
|AVX-512 & FMA (512-bit)||32||64||0|
|AMD Bobcat||AMD64 (64-bit)||2||4||0|
|AMD K10||SSE4/4a (128-bit)||4||8||0|
|AMD Bulldozer (Piledriver, Steamroller, Excavator)||AVX (128-bit) Bulldozer-Steamroller
AVX2 (128-bit) Excavator
|AMD Zen (Ryzen 1000 series, Threadripper 1000 series, Epyc Naples)
AMD Zen+ (Ryzen 2000 series, Threadripper 2000 series)
|AVX2 & FMA (128-bit, 256-bit decoding)||8||16||0|
|AMD Zen 2 (Ryzen 3000 series, Threadripper 3000 series, Epyc Rome)||AVX2 & FMA (256-bit)||16||32||0|
|ARM Cortex-A7, A9, A15||ARMv7||1||8||0|
|ARM Cortex-A32, A35, A53, A55, A72, A73, A75||ARMv8||2||8||0|
|ARM Cortex-A76, A77||ARMv8||8||16||0|
|Qualcomm Kryo (1xx - 3xx)||ARMv8||2||8||0|
|Qualcomm Kryo (4xx)||ARMv8||8||16||0|
|Samsung Exynos M1 and M2||ARMv8||2||8||0|
|Samsung Exynos M3 and M4||ARMv8||3||12||0|
|IBM PowerPC A2 (Blue Gene/Q)||?||8||8 (as FP64)||0|
|Nvidia Fermi (only GeForce GTX 465-480, 560 Ti, 570-590)||PTX||1/4 (locked by driver, 1 in hardware)||2||0|
|Nvidia Fermi (only Quadro 600-2000)||PTX||1/8||2||0|
|Nvidia Fermi (only Quadro 4000-7000, Tesla)||PTX||1||2||0|
|Nvidia Kepler (GeForce (except Titan and Titan Black), Quadro (except K6000), Tesla K10)||PTX||1/12 (for GK110: locked by driver, 2/3 in hardware)||2||0|
|Nvidia Kepler (GeForce GTX Titan and Titan Black, Quadro K6000, Tesla (except K10))||PTX||2/3||2||0|
Nvidia Pascal (all except Quadro GP100 and Tesla P100)
|Nvidia Pascal (only Quadro GP100 and Tesla P100)||PTX||1||2||4|
|Nvidia Turing (only GeForce 16XX)||PTX||1/16||2 (FP32) + 2 (INT32)||4|
|Nvidia Turing (all except GeForce 16XX)||PTX||1/16||2 (FP32) + 2 (INT32)||16|
|AMD GCN (only Radeon Pro WX 2100-7100)||GCN||1/8||2||2|
|AMD GCN (all except Radeon VII, Instinct MI50 and MI60, Radeon Pro WX 2100-7100)||GCN||1/8||2||4|
|AMD GCN Vega 20 (only Radeon VII)||GCN||1/2 (locked by driver, 1 in hardware)||2||4|
|AMD GCN Vega 20 (only Radeon Instinct MI50 and MI60)||GCN||1||2||4|
|Graphcore Colossus GC2 (values estimated)||?||0||18||72|
Single computer records
In June 1997, Intel's ASCI Red was the world's first computer to achieve one teraFLOPS and beyond. Sandia director Bill Camp said that ASCI Red had the best reliability of any supercomputer ever built, and "was supercomputing's high-water mark in longevity, price, and performance".
For comparison, a handheld calculator performs relatively few FLOPS. A computer response time below 0.1 second in a calculation context is usually perceived as instantaneous by a human operator, so a simple calculator needs only about 10 FLOPS to be considered functional.
In June 2006, a new computer was announced by Japanese research institute RIKEN, the MDGRAPE-3. The computer's performance tops out at one petaFLOPS, almost two times faster than the Blue Gene/L, but MDGRAPE-3 is not a general purpose computer, which is why it does not appear in the Top500.org list. It has special-purpose pipelines for simulating molecular dynamics.
By 2007, Intel Corporation unveiled the experimental multi-core POLARIS chip, which achieves 1 teraFLOPS at 3.13 GHz. The 80-core chip can raise this result to 2 teraFLOPS at 6.26 GHz, although the thermal dissipation at this frequency exceeds 190 watts.
On June 26, 2007, IBM announced the second generation of its top supercomputer, dubbed Blue Gene/P and designed to continuously operate at speeds exceeding one petaFLOPS. When configured to do so, it can reach speeds in excess of three petaFLOPS.
On October 25, 2007, NEC Corporation of Japan issued a press release announcing its SX series model SX-9, claiming it to be the world's fastest vector supercomputer. The SX-9 features the first CPU capable of a peak vector performance of 102.4 gigaFLOPS per single core.
On February 4, 2008, the NSF and the University of Texas at Austin opened full scale research runs on an AMD, Sun supercomputer named Ranger, the most powerful supercomputing system in the world for open science research, which operates at sustained speed of 0.5 petaFLOPS.
On May 25, 2008, an American supercomputer built by IBM, named 'Roadrunner', reached the computing milestone of one petaFLOPS. It headed the June 2008 and November 2008 TOP500 list of the most powerful supercomputers (excluding grid computers). The computer is located at Los Alamos National Laboratory in New Mexico. The computer's name refers to the New Mexico state bird, the greater roadrunner (Geococcyx californianus).
In June 2008, AMD released ATI Radeon HD 4800 series, which are reported to be the first GPUs to achieve one teraFLOPS. On August 12, 2008, AMD released the ATI Radeon HD 4870X2 graphics card with two Radeon R770 GPUs totaling 2.4 teraFLOPS.
In November 2008, an upgrade to the Cray Jaguar supercomputer at the Department of Energy's (DOE's) Oak Ridge National Laboratory (ORNL) raised the system's computing power to a peak 1.64 petaFLOPS, making Jaguar the world's first petaFLOPS system dedicated to open research. In early 2009 the supercomputer was named after a mythical creature, Kraken. Kraken was declared the world's fastest university-managed supercomputer and sixth fastest overall in the 2009 TOP500 list. In 2010 Kraken was upgraded and can operate faster and is more powerful.
As of 2010[update] the fastest PC processor reached 109 gigaFLOPS (Intel Core i7 980 XE) in double precision calculations. GPUs are considerably more powerful. For example, Nvidia Tesla C2050 GPU computing processors perform around 515 gigaFLOPS in double precision calculations, and the AMD FireStream 9270 peaks at 240 gigaFLOPS.
In November 2011, it was announced that Japan had achieved 10.51 petaFLOPS with its K computer. It has 88,128 SPARC64 VIIIfx processors in 864 racks, with theoretical performance of 11.28 petaFLOPS. It is named after the Japanese word "kei", which stands for 10 quadrillion, corresponding to the target speed of 10 petaFLOPS.
On November 15, 2011, Intel demonstrated a single x86-based processor, code-named "Knights Corner", sustaining more than a teraFLOPS on a wide range of DGEMM operations. Intel emphasized during the demonstration that this was a sustained teraFLOPS (not "raw teraFLOPS" used by others to get higher but less meaningful numbers), and that it was the first general purpose processor to ever cross a teraFLOPS.
On June 18, 2012, IBM's Sequoia supercomputer system, based at the U.S. Lawrence Livermore National Laboratory (LLNL), reached 16 petaFLOPS, setting the world record and claiming first place in the latest TOP500 list.
On November 12, 2012, the TOP500 list certified Titan as the world's fastest supercomputer per the LINPACK benchmark, at 17.59 petaFLOPS. It was developed by Cray Inc. at the Oak Ridge National Laboratory and combines AMD Opteron processors with "Kepler" NVIDIA Tesla graphic processing unit (GPU) technologies.
On June 20, 2016, China's Sunway TaihuLight was ranked the world's fastest with 93 petaFLOPS on the LINPACK benchmark (out of 125 peak petaFLOPS). The system, which is almost exclusively based on technology developed in China, is installed at the National Supercomputing Center in Wuxi, and represents more performance than the next five most powerful systems on the TOP500 list combined.
In June 2018, Summit, an IBM-built supercomputer now running at the Department of Energy's (DOE) Oak Ridge National Laboratory (ORNL), captured the number one spot with a performance of 122.3 petaFLOPS on High Performance Linpack (HPL), the benchmark used to rank the TOP500 list. Summit has 4,356 nodes, each one equipped with two 22-core Power9 CPUs, and six NVIDIA Tesla V100 GPUs.
Distributed computing records
- As of March 2020[update], the Folding@home network has over 767 petaFLOPS of total computing power. It was the first computing project of any kind to cross the 1, 2, 3, 4, and 5 native petaFLOPS milestones. This level of performance is primarily enabled by the cumulative effort of a vast array of powerful GPU and CPU units.
- As of June 2018[update], the entire BOINC network averages about 20 petaFLOPS.
- As of June 2018[update], SETI@Home, employing the BOINC software platform, averages 896 teraFLOPS.
- As of June 2018[update], Einstein@Home, a project using the BOINC network, is crunching at 3 petaFLOPS.
- As of June 2018[update], MilkyWay@Home, using the BOINC infrastructure, computes at 847 teraFLOPS.
- As of June 2018[update], GIMPS, searching for Mersenne primes, is sustaining 306 teraFLOPS.
Cost of computing
|Date||Approximate cost per GFLOPS||Approximate cost per GFLOPS (2019 US dollars)||Approximate cost per TFLOPS (2017 US dollars)||Platform providing the lowest cost per GFLOPS||Comments|
|1961||$18.7 billion||$160 billion||$160 trillion||A basic installation of IBM 7030 Stretch had a cost at the time of US$7.78 million each.||The IBM 7030 Stretch performs one floating-point multiply every 2.4 microseconds.|
|1984||$18,750,000||$46,140,000||$44.2 billion||Cray X-MP/48||$15,000,000 / 0.8 GFLOPS|
|1997||$30,000||$48,000||$46,000,000||Two 16-processor Beowulf clusters with Pentium Pro microprocessors|
|April 2000||$1,000||$1,510||$1,440,000||Bunyip Beowulf cluster||Bunyip was the first sub-US$1/MFLOPS computing technology. It won the Gordon Bell Prize in 2000.|
|May 2000||$640||$964||$922,000||KLAT2||KLAT2 was the first computing technology which scaled to large applications while staying under US-$1/MFLOPS.|
|August 2003||$82||$114||$109,000||KASY0||KASY0 was the first sub-US$100/GFLOPS computing technology.|
|August 2007||$48||$59||$57,000||Microwulf||As of August 2007, this 26.25 GFLOPS "personal" Beowulf cluster can be built for $1256.|
|March 2011||$1.80||$2.07||$1,980||HPU4Science||This $30,000 cluster was built using only commercially available "gamer" grade hardware.|
|August 2012||$0.75||$0.84||$800||Quad AMD Radeon 7970 GHz System||A quad AMD Radeon 7970 desktop computer reaching 16 TFLOPS of single-precision, 4 TFLOPS of double-precision computing performance. Total system cost was $3000; Built using only commercially available hardware.|
|June 2013||$0.22||$0.24||$230||Sony PlayStation 4||The Sony PlayStation 4 is listed as having a peak performance of 1.84 TFLOPS, at a price of $400|
|November 2013||$0.16||$0.18||$170||AMD Sempron 145 & GeForce GTX 760 System||Built using commercially available parts, a system using one AMD Sempron 145 and three Nvidia GeForce GTX 760 reaches a total of 6.771 TFLOPS for a total cost of $1090.66.|
|December 2013||$0.12||$0.13||$130||Pentium G550 & Radeon R9 290 System||Built using commercially available parts. Intel Pentium G550 and AMD Radeon R9 290 tops out at 4.848 TFLOPS grand total of US$681.84.|
|January 2015||$0.08||$0.09||$80||Celeron G1830 & Radeon R9 295X2 System||Built using commercially available parts. Intel Celeron G1830 and AMD Radeon R9 295X2 tops out at over 11.5 TFLOPS at a grand total of US$902.57.|
|June 2017||$0.06||$0.06||$60||AMD Ryzen 7 1700 & AMD Radeon Vega Frontier Edition||Built using commercially available parts. AMD Ryzen 7 1700 CPU combined with AMD Radeon Vega FE cards in CrossFire tops out at over 50 TFLOPS at just under US$3,000 for the complete system.|
|October 2017||$0.03||$0.03||$30||Intel Celeron G3930 & AMD RX Vega 64||Built using commercially available parts. Three AMD RX Vega 64 graphics cards provide just over 75 TFLOPS half precision (38 TFLOPS SP or 2.6 TFLOPS DP when combined with the CPU) at ~$2,050 for the complete system.|
- Floating Point Retrieved on December 25, 2009.
- Summary: Fixed-point (integer) vs floating-point Retrieved on December 25, 2009.
- Fixed versus floating point. Retrieved on December 25, 2009.
- Data manipulation and math calculation. Retrieved on December 25, 2009.
- "Nodes, Sockets, Cores and FLOPS, Oh, My" by Dr. Mark R. Fernandez, Ph.D.
- Dolbeau, Romain (2017). "Theoretical Peak FLOPS per instruction set: a tutorial". Journal of Supercomputing. 74 (3): 1341–1377. doi:10.1007/s11227-017-2177-5.
- "New instructions support for Bulldozer (FMA3) and Piledriver (FMA3+4 and CVT,BMI,TBM)" (PDF).
- "Agner's CPU blog - Test results for AMD Ryzen".
- https://arstechnica.com/gadgets/2017/03/amds-moment-of-zen-finally-an-architecture-that-can-compete/2/ "each core now has a pair of 128-bit FMA units of its own"
- Mike Clark (August 23, 2016). A New x86 Core Architecture for the Next Generation of Computing (PDF). HotChips 28. AMD. page 7
- "The microarchitecture of Intel and AMD CPUs" (PDF).
- "AMD CEO Lisa Su's COMPUTEX 2019 Keynote". www.youtube.com.
- "Entertainment Systems and High-Performance Processor SH-4" (PDF). Hitachi Review. Hitachi. 48 (2): 58–63. 1999. Retrieved June 21, 2019.
- "SH-4 Next-Generation DSP Architecture for VoIP" (PDF). Hitachi. 2000. Retrieved June 21, 2019.
- "Alles zu Navi: Radeon RX 5700 XT ist RDNA mit GDDR6".
- "AMD Radeon RX 5700 XT".
- "6 threads per core imply that IPC is a multiple of 6, 1216 cores per chip". www.youtube.com.
- "250 TFLOPs/s for two chips with FP16 mixed precision". www.youtube.com.
- "Estimation via power consumption that FP32 is 1/4 of FP16 and that clock frequency is below 1.5GHz". www.youtube.com.
- "Floating-Point Operations Per Second (FLOPS)".
- "Sandia's ASCI Red, world's first teraflop supercomputer, is decommissioned" (PDF). Archived from the original (PDF) on November 5, 2010. Retrieved November 17, 2011.
- "Response Times: The Three Important Limits". Jakob Nielsen. Retrieved June 11, 2008.
- Richard Swinburne (April 30, 2007). "The Arrival of TeraFLOP Computing". bit-tech.net. Retrieved February 9, 2012.
- "June 2008". TOP500. Retrieved July 8, 2008.
- "29th TOP500 List of World's Fastest Supercomputers Released". Top500.org. June 23, 2007. Archived from the original on May 9, 2008. Retrieved July 8, 2008.
- "NEC Launches World's Fastest Vector Supercomputer, SX-9". NEC. October 25, 2007. Retrieved July 8, 2008.
- "University of Texas at Austin, Texas Advanced Computing Center". Archived from the original on August 1, 2009. Retrieved September 13, 2010.
Any researcher at a U.S. institution can submit a proposal to request an allocation of cycles on the system.
- Sharon Gaudin (June 9, 2008). "IBM's Roadrunner smashes 4-minute mile of supercomputing". Computerworld. Archived from the original on December 24, 2008. Retrieved June 10, 2008.
- "Austin ISC08". Top500.org. November 14, 2008. Archived from the original on February 22, 2012. Retrieved February 9, 2012.
- Fildes, Jonathan (June 9, 2008). "Supercomputer sets petaflop pace". BBC News. Retrieved July 8, 2008.
- Greenberg, Andy (November 16, 2009). "Cray Dethrones IBM In Supercomputing". Forbes.
- "China claims supercomputer crown". BBC News. October 28, 2010.
- Dillow, Clay (October 28, 2010). "China Unveils 2507 Petaflop Supercomputer, the World's Fastest". Popsci.com. Retrieved February 9, 2012.
- "Intel's Core i7-980X Extreme Edition – Ready for Sick Scores?: Mathematics: Sandra Arithmetic, Crypto, Microsoft Excel". Techgage. March 10, 2010. Retrieved February 9, 2012.
- "NVIDIA Tesla Personal Supercomputer". Nvidia.com. Retrieved February 9, 2012.
- "AMD FireStream 9270 GPU Compute Accelerator". Amd.com. Retrieved February 9, 2012.
- "'K computer' Achieves Goal of 10 Petaflops". Fujitsu.com. Retrieved February 9, 2012.
- See Japanese numbers
- "Intel's Knights Corner: 50+ Core 22nm Co-processor". November 16, 2011. Retrieved November 16, 2011.
- "Intel unveils 1 TFLOP/s Knight's Corner". Retrieved November 16, 2011.
- Clark, Don (June 18, 2012). "IBM Computer Sets Speed Record". The Wall Street Journal. Retrieved June 18, 2012.
- "BBC News – US Titan supercomputer clocked as world's fastest". BBC News. Bbc.co.uk. November 12, 2012. Retrieved February 28, 2013.
- "Oak Ridge Claims No. 1 Position on Latest TOP500 List with Titan | TOP500 Supercomputer Sites". Top500.org. November 12, 2012. Retrieved February 28, 2013.
- Montalbano, Elizabeth (October 11, 2011). "Oak Ridge Labs Builds Fastest Supercomputer". Informationweek. Retrieved February 9, 2012.
- Tibken, Shara (October 29, 2012). "Titan supercomputer debuts for open scientific research | Cutting Edge – CNET News". News.cnet.com. Retrieved February 28, 2013.
- "Chinese Supercomputer Is Now The World's Fastest – By A Lot". Forbes Magazine. June 17, 2013. Retrieved June 17, 2013.
- Feldman, Michael. "China Races Ahead in TOP500 Supercomputer List, Ending US Supremacy". TOP500.org. Retrieved December 31, 2016.
- "June 2018 | TOP500 Supercomputer Sites". www.top500.org. Retrieved July 17, 2018.
- "Closing in on 100 Petaflops". Folding@Home. May 11, 2016. Retrieved July 17, 2016.
- "Folding@home team stats pages". Folding@Home. Retrieved October 14, 2016.
- "Sony Computer Entertainment's Support for Folding@home Project on PlayStation™3 Receives This Year's "Good Design Gold Award"" (Press release). Sony Computer Entertainment Inc. November 6, 2008. Archived from the original on January 31, 2009. Retrieved December 11, 2008.
- "Computering Power". BOINC. Retrieved June 15, 2018.
- "SETI@Home Credit overview". BOINC. Retrieved June 15, 2018.
- "Einstein@Home Credit overview". BOINC. Retrieved June 15, 2018.
- "MilkyWay@Home Credit overview". BOINC. Retrieved June 15, 2018.
- "Internet PrimeNet Server Distributed Computing Technology for the Great Internet Mersenne Prime Search". GIMPS. Retrieved June 15, 2018.
- Federal Reserve Bank of Minneapolis. "Consumer Price Index (estimate) 1800–". Retrieved January 1, 2020.
- "The IBM 7030 (STRETCH)". Norman Hardy. Retrieved February 24, 2017.
- "Loki and Hyglac". Loki-www.lanl.gov. July 13, 1997. Archived from the original on July 21, 2011. Retrieved February 9, 2012.
- "Kentucky Linux Athlon Testbed 2 (KLAT2)". The Aggregate. Retrieved February 9, 2012.
- "KASY0". The Aggregate. August 22, 2003. Retrieved February 9, 2012.
- "Microwulf: A Personal, Portable Beowulf Cluster". Archived from the original on September 12, 2007. Retrieved February 9, 2012.
- Adam Stevenson, Yann Le Du, and Mariem El Afrit. "High-performance computing on gamer PCs." Ars Technica. March 31, 2011.
- Tom Logan (January 9, 2012). "HD7970 Quadfire Eyefinity Review". OC3D.net.
- "Sony Sparks Price War With PS4 Priced at $399." CNBC. June 11, 2013.
- "Radeon R9 295X2 8 GB Review: Project Hydra Gets Liquid Cooling". April 8, 2014.
- Perez, Carol E. (July 13, 2017). "Building a 50 Teraflops AMD Vega Deep Learning Box for Under $3K". Intuition Machine. Retrieved July 26, 2017.
- "lowest_$/fp16 - mattebaughman's Saved Part List - Celeron G3930 2.9GHz Dual-Core, Radeon RX VEGA 64 8GB (3-Way CrossFire), XON-350_BK ATX Mid Tower - PCPartPicker". pcpartpicker.com. Retrieved September 13, 2017.