FLOPS

In computing, FLOPS (or flops or flop/s) is an acronym meaning FLoating point Operations Per Second. The FLOPS is a measure of a computer's performance, especially in fields of scientific calculations that make heavy use of floating point calculations, similar to the older, simpler, instructions per second. Since the final S stands for "second", conservative speakers consider "FLOPS" as both the singular and plural of the term, although the singular "FLOP" is frequently encountered. Alternatively, the singular FLOP (or flop) is used as an abbreviation for "FLoating-point OPeration", and a flop count is a count of these operations (e.g., required by a given algorithm or computer program). In this context, "flops" is simply the plural rather than a rate.

NEC's SX-9 supercomputer was the world's first vector processor to exceed 100 gigaFLOPS per single core. IBM's supercomputer dubbed Blue Gene/P is designed to eventually operate at three petaFLOPS.^[1] Now, Jaguar, The world's fastest supercomputer is Cray XT5, also known as Jaguar. Jaguar bags the no. 1 spot, beating IBM's Roadrunner, who has been holding the top crown since past 18 months. Jaguar recently upgraded its quad-core CPUs to hex-core Opteron processors, which meant a 2.3 petaflop per second theoretical performance peak (”nearly a quarter of a million cores”), and 1.75 petaflops measured by the Linpack benchmark. This surpasses Roadrunner's 1.04 petaflop/s. A petaflop/s refers to 1 quadrillion calculations per second. Jaguar, located at the US Department of Energy's Oak Ridge Leadership Computing Facility, came close to beating Roadrunner in the two previous Top500 lists. This time, however, Roadrunner's performance fell from 1.105 petaflop/s in June due to a repartitioning of the system.

A basic calculator performs relatively few FLOPS. Each calculation request to a typical calculator requires only a single operation, so there is rarely any need for its response time to exceed that needed by the operator. A response time below 0.1 second in a calculation context is usually perceived as instantaneous by a human operator,^[2] so a simple calculator with multiplication and division needs only about 10 FLOPS.

Measuring performance

In order for FLOPS to be useful as a measure of floating-point performance, a standard benchmark must be available on all computers of interest. One example is the LINPACK benchmark.

There are many factors in computer performance other than raw floating-point computation speed, such as I/O performance, interprocessor communication, cache coherence, and the memory hierarchy. This means that supercomputers are in general only capable of a fraction of their "theoretical peak" FLOPS throughput (obtained by adding together the theoretical peak FLOPS performance of every element of the system). Even when operating on large highly parallel problems, their performance will be bursty, mostly due to the residual effects of Amdahl's law. Real benchmarks therefore measure both peak actual FLOPS performance as well as sustained FLOPS performance.

Supercomputer ratings, like TOP500, usually derive theoretical peak FLOPS as a product of number of cores, cycles per second each core runs at, and number of double-precision FLOPS each core can ideally perform, thanks to SIMD or otherwise. Despite different processor architectures can achieve different parallelism on single core, most mainstream ones, like recent Xeon and Itanium models, claim a factor of four. Some ratings adopted the factor as a given constant, and use it to compute peak values for all architectures, often leading to huge difference from sustained performance.

For ordinary (non-scientific) applications, integer operations (measured in MIPS) are far more common. Measuring floating point operation speed, therefore, does not predict accurately how the processor will perform on just any problem. However, for many scientific jobs such as analysis of data, a FLOPS rating is effective.

Historically, the earliest reliably documented serious use of the Floating Point Operation as a metric appears to be AEC justification to Congress for purchasing a Control Data CDC 6600 in the mid-1960s.

The terminology is currently so confusing that until April 24, 2006 U.S. export control was based upon measurement of "Composite Theoretical Performance" (CTP) in millions of "Theoretical Operations Per Second" or MTOPS. On that date, however, the U.S. Department of Commerce's Bureau of Industry and Security amended the Export Administration Regulations to base controls on Adjusted Peak Performance (APP) in Weighted TeraFLOPS (WT).

Records

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 TFLOPS at 3.13 GHz. The 80-core chip can increase this result to 2 TFLOPS at 6.26 GHz, although the thermal dissipation at this frequency exceeds 190 watts^[3].

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.^[4]

In June 2007, Top500.org reported the fastest computer in the world to be the IBM Blue Gene/L supercomputer, measuring a peak of 596 TFLOPS^[5]. The Cray XT4 hit second place with 101.7 TFLOPS.

On October 25, 2007, NEC Corporation of Japan issued a press release^[6] 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 opened full scale research runs on an AMD, Sun supercomputer Ranger, the most powerful supercomputing system in the world for open science research, which operates at sustained speed of half a petaflop.

On May 25, 2008, an American military supercomputer built by IBM, named 'Roadrunner', reached the computing milestone of one petaflop by processing more than 1.026 quadrillion calculations per second. It headed the June, 2008^[7] and November, 2008^[8] TOP500 list of the most powerful supercomputers (excluding grid computers). The computer's name refers to the state bird of New Mexico, the Greater Roadrunner.^[9]

In June 2008, AMD released ATI Radeon HD4800 series, which are reported to be the first GPUs to achieve one teraFLOP scale. On August 12, 2008 AMD released the ATI Radeon HD 4870X2 graphics card with two Radeon R770 GPUs totalling 2.4 teraFLOPs.

In November 2008, the latest upgrade to the Cray XT Jaguar supercomputer at the Department of Energy’s (DOE’s) Oak Ridge National Laboratory (ORNL) has increased the system's computing power to a peak 1.64 “petaflops,” or a quadrillion mathematical calculations per second, making Jaguar the world’s first petaflop 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 most recent TOP500 list, which is the global standard for ranking supercomputers. Just recently Kraken was upgraded and can operate faster and is more powerful.

As of 2008, the fastest PC processors (quad-core) perform over 70 GFLOPS (Intel Core i7 965 XE) in double precision^[10]. GPUs are considerably more powerful, for example, nVidia's Tesla C1060 GPU computing card performs around 933 GFLOPS in single precision calculations^[11] while AMD's FireStream 9270 reaches 1200 GFLOPS ^[12]. However, whereas GPUs are highly efficient at single precision calculations they are not nearly as fast in double precision (DP) operations with the same Nvidia Tesla C1060 capable of 78 GFLOPS in double precision ^[11] and the AMD FireStream 9270 capable of 240 GFLOPS in double precision ^[12].

In 2009, the Cray Jaguar performed at 1.75 petaFLOPs, besting the IBM Roadrunner for the number one spot on the TOP500 list.^[13]

Distributed computing uses the Internet to link personal computers to achieve more flops:

• Folding@Home is of April 2009 sustaining over 8.1 PFLOPS ^[14], the first computing project of any kind to cross the four petaFLOPS milestone. This level of performance is primarily enabled by the cumulative effort of a vast array of PlayStation 3, CPU, and powerful GPU units.^[15]
• The entire BOINC averages over 2.2 PFLOPS as of August 13, 2009^[16].
• SETI@Home computes data averages more than 694 TFLOPS^[17]
• Einstein@Home is crunching more than 150 TFLOPS^[18]
• As of August 2008^[update], GIMPS is sustaining 27 TFLOPS.^[19]

Future developments

In May 2008 a collaboration was announced between NASA, SGI and Intel to build a 1 petaflops computer, Pleiades, in 2009, scaling up to 10 petaflops by 2012.^[20] At the same time, IBM intends to build a 20 petaflops supercomputer, Sequoia, at Lawrence Livermore National Laboratory until 2011. IBM also intends to have an exaflop supercomputer functional by the 2010s to support the Square Kilometre Array. It will be built in either South Africa or Australia, depending on who wins the bid for the SKA. http://www.computerworld.com.au/article/319128/ska_telescope_provide_billion_pcs_worth_processing

Given the current speed of progress, Supercomputers are projected to reach 1 Exaflops in 2019.^[21] Erik P. DeBenedictis of Sandia National Laboratories theorizes that a Zettaflop computer is required to accomplish full weather modeling, which could cover a two week time span accurately.^[22] Such systems might be built around 2030.^[23]

Cost of computing

Hardware costs

The following is a list of examples of computers that demonstrates how performance has increased drastically and price has decreased drastically. The "cost per GFLOPS" is the cost for a set of hardware that would theoretically operate at one billion floating point operations per second. During the era when no single computation platform was able to achieve one GFLOPS, this table lists the total cost for multiple instances of a fast computation platform whose speed sums to one GFLOPS. Otherwise, the least expensive computing platform able to achieve one GFLOPS is listed.

The trend toward a higher and higher number of transistors that can be placed inexpensively on an integrated circuit follows Moore's law. This trend explains the increasing speed and decreasing cost of computer processing.

Operation costs

In energy cost, according to the Green500 list, as of November 2008 the most efficient TOP500 supercomputer runs at 536.24 MFLOPS per watt. This translates to an energy requirement of 1.86 watts per GFLOPS, however this energy requirement will be much greater for less efficient supercomputers.

Hardware costs for low cost supercomputers may be less significant than energy costs when running continuously for several years. A Playstation 3 (PS3) 120 GiB (65 nm Cell) costs $299 (as of September 2009) and consumes 250 watts^[30] or $219 of electricity each year if operated 24 hours per day, conservatively assuming U.S. national average residential electric rates of $0.10/kWh^[31] (0.250 kW × 24 h × 365 d × 0.10 $/kWh = $219 per year).

See also

Computer Science portal

• Gordon Bell Prize

References

External links

• Current Einstein@Home benchmark
• BOINC projects global benchmark
• Current GIMPS throughput
• Top500.org
• LinuxHPC.org Linux High Performance Computing and Clustering Portal
• WinHPC.org Windows High Performance Computing and Clustering Portal
• Oscar Linux-cluster ranking list by CPUs/types and respective FLOPS
• Information on how to calculate "Composite Theoretical Performance" (CTP)
• Information on the Oak Ridge National Laboratory Cray XT system.
• Infiscale Cluster Portal - Free GPL HPC
• Source code, pre-compiled versions and results for PCs - Linpack, Livermore Loops, Whetstone MFLOPS
• PC CPU Performance Comparisons %MFLOPS/MHz - CPU, Caches and RAM
• Xeon export compliance metrics, including GFLOPS