Unum (number format)

The first version of unums, formally known as Type I unum, was introduced in Gustafson's book The End of Error as a superset of the IEEE-754 floating-point format. The defining features of the Type I unum format are:

• a variable-width storage format for both the significand and exponent, and
• a u-bit, which determines whether the unum corresponds to an exact number (u = 0), or an interval between consecutive exact unums (u = 1). In this way, the unums cover the entire extended real number line [−∞,+∞].

For computation with the format, Gustafson proposed using interval arithmetic with a pair of unums, what he called a ubound, providing the guarantee that the resulting interval contains the exact solution.

William M. Kahan and Gustafson debated unums at the Arith23 conference.

Type II Unums were introduced in 2016 as a redesign of Unums that broke IEEE-754 compatibility. In addition to the sign bit and the interval bit mentioned earlier, the Type II Unum uses a bit to indicate inversion. These three operations make it possible, starting from a finite set of points between one and infinity, to quantify the entire projective line except for four points: the two exceptions, 0 and ∞, and then 1 and −1. This set of points is chosen arbitrarily, and arithmetic operations involving them are not performed logically but rather by using a lookup table. The size of such a table becomes prohibitive for an encoding format spanning multiple bytes. This challenge necessitated the development of the Type III Unum, known as the posit, discussed below.

In February 2017, Gustafson officially introduced Type III unums (posits), for fixed floating-point-like values and valids for interval arithmetic. In March 2022, a standard was ratified and published by the Posit Working Group.

Posits are a hardware-friendly version of unum where difficulties faced in the original type I unum due to its variable size are resolved. Compared to IEEE 754 floats of similar size, posits offer a bigger dynamic range and more fraction bits for values with magnitude near 1 (but fewer fraction bits for very large or very small values), and Gustafson claims that they offer better accuracy. Studies confirm that for some applications, posits with quire out-perform floats in accuracy. Posits have superior accuracy in the range near one, where most computations occur. This makes it very attractive to the current trend in deep learning to minimize the number of bits used. It potentially helps any application to accelerate by enabling the use of fewer bits (since it has more fraction bits for accuracy) reducing network and memory bandwidth and power requirements.

The format of an n-bit posit is given a label of "posit" followed by the decimal digits of n (e.g., the 16-bit posit format is "posit16") and consists of four sequential fields:

1. sign: 1 bit, representing an unsigned integer s
2. regime: at least 2 bits and up to (n − 1), representing an unsigned integer r as described below
3. exponent: generally 2 bits as available after regime, representing an unsigned integer e
4. fraction: all remaining bits available after exponent, representing a non-negative real dyadic rational f less than 1

The regime field uses unary coding of k identical bits, followed by a bit of opposite value if any remaining bits are available, to represent an unsigned integer r that is −k if the first bit is 0 or k − 1 if the first bit is 1. The sign, exponent, and fraction fields are analogous to IEEE 754 sign, exponent, and significand fields (respectively), except that the posit exponent and fraction fields may be absent or truncated and implicitly extended with zeroes—an absent exponent is treated as 00₂ (representing 0), a one-bit exponent E₁ is treated as E10₂ (representing the integer 0 if E₁ is 0 or 2 if E₁ is 1), and an absent fraction is treated as 0. Negative numbers (s is 1) are encoded as 2's complements.

The two encodings in which all non-sign bits are 0 have special interpretations:

• If the sign bit is 1, the posit value is NaR ("not a real")
• If the sign bit is 0, the posit value is 0 (which is unsigned and the only value for which the sign function returns 0)

Otherwise, the posit value is equal to $((1\; -\; 3s)\; +\; f)\; \backslash times\; 2^\{(1\; -\; 2s)\; \backslash times\; (4r\; +\; e\; +\; s)\}$, in which r scales by powers of 16, e scales by powers of 2, f distributes values uniformly between adjacent combinations of (r, e), and s adjusts the sign symmetrically about 0.

Note: 32-bit posit is expected to be sufficient to solve almost all classes of applications{{Citation needed|reason=There are many realistic problems which require more than 32 bits of information and thus would be impossible to solve using a 32-bit posit, no matter how detailed. For example, dealing with long-running times where nanosecond or greater accuracy is needed.|date=December 2018}}.

For each positn type of precision $n$, the standard defines a corresponding "quire" type quiren of precision $16\; \backslash times\; n$, used to accumulate exact sums of products of those posits without rounding or overflow in dot products for vectors of up to 2³¹ or more elements (the exact limit is $2^\{23\; +\; 4n\}$). The quire format is a two's complement signed integer, interpreted as a multiple of units of magnitude $2^\{16\; -\; 8n\}$ except for the special value with a leading sign bit of 1 and all other bits equal to 0 (which represents NaR). Quires are based on the work of Ulrich W. Kulisch and Willard L. Miranker.

Valids are described as a Type III Unum mode that bounds results in a given range.

Several software and hardware solutions implement posits. The first complete parameterized posit arithmetic hardware generator was proposed in 2018.

Unum implementations have been explored in Julia and MATLAB. A C++ version with support for any posit sizes combined with any number of exponent bits is available. A fast implementation in C, SoftPosit, provided by the NGA research team based on Berkeley SoftFloat adds to the available software implementations.

{{sticky header}}

SoftPosit is a software implementation of posits based on Berkeley SoftFloat. It allows software comparison between posits and floats. It currently supports

• Add
• Subtract
• Multiply
• Divide
• Fused-multiply-add
• Fused-dot-product (with quire)
• Square root
• Convert posit to signed and unsigned integer
• Convert signed and unsigned integer to posit
• Convert posit to another posit size
• Less than, equal, less than equal comparison
• Round to nearest integer

• convert double to posit
• convert posit to double
• cast unsigned integer to posit

It works for 16-bit posits with one exponent bit and 8-bit posit with zero exponent bit. Support for 32-bit posits and flexible type (2-32 bits with two exponent bits) is pending validation. It supports x86_64 systems. It has been tested on GNU gcc (SUSE Linux) 4.8.5 Apple LLVM version 9.1.0 (clang-902.0.39.2).

Add with posit8_t

1. include "softposit.h"

int main(int argc, char *argv[]) {

posit8_t pA, pB, pZ;

pA = castP8(0xF2);

pB = castP8(0x23);

pZ = p8_add(pA, pB);

// To check answer by converting it to double

double dZ = convertP8ToDouble(pZ);

printf("dZ: %.15f\n", dZ);

// To print result in binary (warning: non-portable code)

uint8_t uiZ = castUI8(pZ);

printBinary((uint64_t*)&uiZ, 8);

return 0;

}

Fused dot product with quire16_t

// Convert double to posit

posit16_t pA = convertDoubleToP16(1.02783203125);

posit16_t pB = convertDoubleToP16(0.987060546875);

posit16_t pC = convertDoubleToP16(0.4998779296875);

posit16_t pD = convertDoubleToP16(0.8797607421875);

quire16_t qZ;

// Set quire to 0

qZ = q16_clr(qZ);

// Accumulate products without roundings

qZ = q16_fdp_add(qZ, pA, pB);

qZ = q16_fdp_add(qZ, pC, pD);

// Convert back to posit

posit16_t pZ = q16_to_p16(qZ);

// To check answer

double dZ = convertP16ToDouble(pZ);

William M. Kahan, the principal architect of IEEE 754-1985 criticizes type I unums on the following grounds (some are addressed in type II and type III standards):

• The description of unums sidesteps using calculus for solving physics problems.
• Unums can be expensive in terms of time and power consumption.
• Each computation in unum space is likely to change the bit length of the structure. This requires either unpacking them into a fixed-size space, or data allocation, deallocation, and garbage collection during unum operations, similar to the issues for dealing with variable-length records in mass storage.
• Unums provide only two kinds of numerical exception, quiet and signaling NaN (Not-a-Number).
• Unum computation may deliver overly loose bounds from the selection of an algebraically correct but numerically unstable algorithm.
• The benefits of unum over short precision floating point for problems requiring low precision are not obvious.
• Solving differential equations and evaluating integrals with unums guarantee correct answers but may not be as fast as methods that usually work.

• Karlsruhe Accurate Arithmetic (KAA)
• Q (number format)
• Significant figures
• Floating-point error mitigation
• Elias gamma (γ) code
• Tapered floating point (TFP)

{{Reflist|refs=

{{cite book |author-last=Gustafson |author-first=John L. |author-link=John Gustafson (scientist) |title=The End of Error: Unum Computing |publisher=CRC Press |edition=2nd corrected printing, 1st |date=2016-02-04 |orig-date=2015-02-05 |isbn=978-1-4822-3986-7 |series=Chapman & Hall / CRC Computational Science |volume=24 |url=https://www.crcpress.com/The-End-of-Error-Unum-Computing/Gustafson/p/book/9781482239867 |access-date=2016-05-30}} [https://books.google.com/books?id=fZsXBgAAQBAJ] [http://www.crcpress.com/product/isbn/9781482239867]

{{cite web |url=https://github.com/jwmerrill/Pnums.jl |title=jwmerrill/Pnums.jl |work=Github.com |access-date=2017-01-30}}

{{cite web |title=Implementing Unums in Julia |author-first=Simon |author-last=Byrne |date=2016-03-29 |url=http://juliacomputing.com/blog/2016/03/29/unums.html |access-date=2016-05-30}}

{{cite web |title=Unum arithmetic in Julia: Unums.jl |website=GitHub |url=https://github.com/JuliaComputing/Unums.jl |access-date=2016-05-30}}

{{cite web |title=Julia Implementation of Unums: README |website=GitHub |url=https://github.com/REX-Computing/unumjl |access-date=2016-05-30}}

{{cite web |title=Unum (Universal Number) types and operations: Unums |website=GitHub |url=https://github.com/tbreloff/Unums.jl |access-date=2016-05-30}}

{{cite web |title=Reducing Memory Footprints in Explicit Model Predictive Control using Universal Numbers. Submitted to the IFAC World Congress 2017 |author-first1=Deepak |author-last1=Ingole |author-first2=Michal |author-last2=Kvasnica |author-first3=Himeshi |author-last3=De Silva |author-first4=John L. |author-last4=Gustafson |author-link4=John Gustafson (scientist) |url=https://bitbucket.org/kvasnica/munum |access-date=2016-11-15}}

{{cite web |title=MATLAB Prototype of unum (munum) |author-first1=Deepak |author-last1=Ingole |author-first2=Michal |author-last2=Kvasnica |author-first3=Himeshi |author-last3=De Silva |author-first4=John L. |author-last4=Gustafson |author-link4=John Gustafson (scientist) |url=https://bitbucket.org/kvasnica/munum |access-date=2016-11-15}}

{{cite web |title=Program: Special Session: The Great Debate: John Gustafson and William Kahan |work=Arith23: 23rd IEEE Symposium on Computer Arithmetic |location=Silicon Valley, USA |date=2016-07-12 |url=http://arith23.gforge.inria.fr/program.html |access-date=2016-05-30 |url-status=live |archive-url=https://web.archive.org/web/20160530120059/http://arith23.gforge.inria.fr/program.html |archive-date=2016-05-30}}

{{cite AV media |title=The Great Debate @ARITH23: John Gustafson and William Kahan (1:34:41) |type=video |author-first1=John L. |author-last1=Gustafson |author-link1=John Gustafson (scientist) |author-first2=William M. |author-last2=Kahan |author-link2=William Kahan |language=en |date=2016-07-12 |url=https://www.youtube.com/watch?v=LZAeZBVAzVw |access-date=2016-07-20}}

{{cite web |title=A Critique of John L. Gustafson's THE END of ERROR — Unum Computation and his A Radical Approach to Computation with Real Numbers |author-first=William M. |author-last=Kahan |author-link=William Kahan |location=Santa Clara, CA, USA |publisher=IEEE Symposium on Computer Arithmetic, ARITH 23 |date=2016-07-16 |orig-date=2016-07-12 |url=http://people.eecs.berkeley.edu/~wkahan/UnumSORN.pdf |access-date=2016-07-25 |url-status=live |archive-url=https://web.archive.org/web/20160725181213/https://people.eecs.berkeley.edu/~wkahan/UnumSORN.pdf |archive-date=2016-07-25}} [http://arith23.gforge.inria.fr/slides/Kahan.pdf]

{{cite web |title="The Great Debate": Unum arithmetic position paper |author-first=John L. |author-last=Gustafson |author-link=John Gustafson (scientist) |location=Santa Clara, CA, USA |publisher=IEEE Symposium on Computer Arithmetic, ARITH 23 |date=2016-07-12 |url=https://groups.google.com/group/unum-computing/attach/b9d9a10a1e53e/PositionSlides-builds.pdf?part=0.1 |access-date=2016-07-20}} [http://arith23.gforge.inria.fr/slides/Gustafson.pdf]

{{cite web |title=Prof. W. Kahan's Commentary on "THE END of ERROR — Unum Computing" by John L. Gustafson, (2015) CRC Press |date=2016-07-15 |author-first=William M. |author-last=Kahan |author-link=William Kahan |url=https://people.eecs.berkeley.edu/~wkahan/EndErErs.pdf |access-date=2016-08-01 |url-status=live |archive-url=https://web.archive.org/web/20160801162357/https://people.eecs.berkeley.edu/~wkahan/EndErErs.pdf |archive-date=2016-08-01}}

{{cite journal |author-first=Walter F. |author-last=Tichy |title=The End of (Numeric) Error: An interview with John L. Gustafson |journal=Ubiquity |author-link=Walter F. Tichy |publisher=Association for Computing Machinery (ACM) |volume=2016 |number=April |date=April 2016 |pages=1–14 |doi=10.1145/2913029 |quote=JG: The word "unum" is short for "universal number," the same way the word "bit" is short for "binary digit." |doi-access=free}}

{{cite web |author-last=Tichy |author-first=Walter F. |author-link=Walter F. Tichy |date=September 2016 |title=Unums 2.0: An Interview with John L. Gustafson |url=http://ubiquity.acm.org/article.cfm?id=3001758 |website=Ubiquity.ACM.org |access-date=2017-01-30 |quote=I started out calling them "unums 2.0," which seemed to be as good a name for the concept as any, but it is really not a "latest release" so much as it is an alternative.}}

{{cite journal |title=Beating Floating Point at its Own Game: Posit Arithmetic |author-first1=John Leroy |author-last1=Gustafson |author-link1=John Leroy Gustafson |author-first2=Isaac |author-last2=Yonemoto |date=2017 |journal=Supercomputing Frontiers and Innovations |volume=4 |number=2 |publisher=Publishing Center of South Ural State University, Chelyabinsk, Russia |doi=10.14529/jsfi170206 |url=http://superfri.org/superfri/article/download/137/232 |access-date=2017-11-04 |url-status=live |archive-url=https://web.archive.org/web/20171104234856/http://superfri.org/superfri/article/download/137/232 |archive-date=2017-11-04 |doi-access=free}}

{{cite web |title=Posit Arithmetic |author-first=John Leroy |author-last=Gustafson |author-link=John Leroy Gustafson |date=2017-10-10 |url=https://posithub.org/docs/Posits4.pdf |access-date=2017-11-04 |url-status=live |archive-url=https://web.archive.org/web/20171105001022/https://posithub.org/docs/Posits4.pdf |archive-date=2017-11-05}}

John L. Gustafson and I. Yonemoto. (February 2017) Beyond Floating Point: Next Generation Computer Arithmetic. [Online]. Available: https://www.youtube.com/watch?v=aP0Y1uAA-2Y

{{cite conference |last1=Lindstrom |first1=Peter |last2=Lloyd |first2=Scott |last3=Hittinger |first3=Jeffrey |title=Universal Coding of the Reals: Alternatives to IEEE Floating Point |conference=Conference for Next Generation Arithmetic |date=March 2018 |publisher=ACM |article-number=5 |doi=10.1145/3190339.3190344}}

S. Chung, "Provably Correct Posit Arithmetic with Fixed-Point Big Integer." ACM, 2018.

J. Chen, Z. Al-Ars, and H. Hofstee, "A Matrix-Multiply Unit for Posits in Reconfigurable Logic Using (Open)CAPI." ACM, 2018.

Z. Lehoczky, A. Szabo, and B. Farkas, "High-level .NET Software Implementations of Unum Type I and Posit with Simultaneous FPGA Implementation Using Hastlayer." ACM, 2018.

Rohit Chaurasiya, John Gustafson, Rahul Shrestha, Jonathan Neudorfer, Sangeeth Nambiar, Kaustav Niyogi, Farhad Merchant, Rainer Leupers, "[https://posithub.org/docs/iccd_submission_v1.pdf Parameterized Posit Arithmetic Hardware Generator.]" ICCD 2018: 334-341.

S. Langroudi, T. Pandit, and D. Kudithipudi, "Deep Learning Inference on Embedded Devices: Fixed-Point vs Posit". In Energy Efficiency Machine Learning and Cognitive Computing for Embedded Applications (EMC), 2018. [Online]. Available: https://sites.google.com/view/asplos-emc2/program

{{cite journal |author-first1=Ulrich W. |author-last1=Kulisch |author-link1=Ulrich W. Kulisch |author-first2=Willard L. |author-last2=Miranker |author-link2=Willard L. Miranker |title=The Arithmetic of the Digital Computer: A New Approach |publisher=SIAM |journal=SIAM Rev. |volume=28 |number=1 |pages=1–40 |date=March 1986 |doi=10.1137/1028001}}

{{cite web |url=https://github.com/ityonemo/Unum2 |title=GitHub - ityonemo/Unum2: Pivot Unums. |website=GitHub |date=2019-04-29}}

{{cite web |url=https://github.com/stillwater-sc/universal |title=GitHub - stillwater-sc/Universal: Universal Number Arithmetic |website=GitHub |date=2019-06-16}}

{{Cite web|url=https://gitlab.com/cerlane/SoftPosit|title=Cerlane Leong / SoftPosit · GitLab|website=GitLab}}

{{Cite web|url=http://www.jhauser.us/arithmetic/SoftFloat.html|title=Berkeley SoftFloat|website=www.jhauser.us}}

{{cite web |url=https://www.nextplatform.com/2019/07/08/new-approach-could-sink-floating-point-computation/ |title=New Approach Could Sink Floating Point Computation |author-last=Feldman |author-first=Michael |date=2019-07-08 |website=www.nextplatform.com |language=en-US |access-date=2019-07-09}}

{{cite web |url=https://www.vice.com/en/article/a-new-number-format-for-computers-could-nuke-approximation-errors-for-good/ |title=A New Number Format for Computers Could Nuke Approximation Errors for Good |author-last=Byrne |author-first=Michael |date=2016-04-24 |website=Vice |language=en-US |access-date=2019-07-09}}

{{Cite web |title=Standard for Posit Arithmetic (2022) |author=Posit Working Group |date=2022-03-02 |url=https://posithub.org/docs/posit_standard-2.pdf |url-status=live |archive-url=https://web.archive.org/web/20220926071621/https://posithub.org/docs/posit_standard-2.pdf |archive-date=2022-09-26 |access-date=2022-12-21}}

{{Cite journal|author=David Mallasén |author2=Alberto A. Del Barrio |author3=Manuel Prieto-Matias |date=2024 |title=Big-PERCIVAL: Exploring the Native Use of 64-Bit Posit Arithmetic in Scientific Computing|journal=IEEE Transactions on Computers |volume=73 |issue=6 |pages=1472–1485 |doi=10.1109/TC.2024.3377890 |arxiv=2305.06946 }}

}}

• {{cite web |author-first=John L. |author-last=Gustafson |author-link=John Gustafson (scientist) |title=Right-Sizing Precision: Unleashed Computing: The need to right-size precision to save energy, bandwidth, storage, and electrical power |date=March 2013 |url=http://www.johngustafson.net/presentations/Right-SizingPrecision1.pdf |access-date=2016-06-06 |url-status=live |archive-url=https://web.archive.org/web/20160606203112/http://www.johngustafson.net/presentations/Right-SizingPrecision1.pdf |archive-date=2016-06-06}}
• {{cite web |title=Slidecast: John Gustafson Explains Energy Efficient Unum Computing |date=2015-03-02 |author-first=Rich |author-last=Brueckner |publisher=Inside HPC |url=http://insidehpc.com/2015/03/slidecast-john-gustafson-explains-energy-efficient-unum-computing/ |work=The Rich Report |access-date=2016-06-10 |url-status=live |archive-url=https://web.archive.org/web/20160710220847/http://insidehpc.com/2015/03/slidecast-john-gustafson-explains-energy-efficient-unum-computing/ |archive-date=2016-07-10}}
• {{cite web |author-first=John L. |author-last=Gustafson |author-link=John Gustafson (scientist) |title=The end of numerical error |date=2015 |url=http://arith22.gforge.inria.fr/slides/06-gustafson.pdf |access-date=2016-06-06 |url-status=live |archive-url=https://web.archive.org/web/20160606203155/http://arith22.gforge.inria.fr/slides/06-gustafson.pdf |archive-date=2016-06-06}}
• {{cite web |author-first=John L. |author-last=Gustafson |author-link=John Gustafson (scientist) |title=A Radical Approach to Computation with Real Numbers – Unums version 2.0 |orig-date=2016-02-22 |date=2016-06-03 |url=http://www.johngustafson.net/presentations/Unums2.0.pptx |type=PPT |access-date=2016-07-10 |url-status=live |archive-url=https://web.archive.org/web/20160710194003/http://www.johngustafson.net/presentations/Unums2.0.pptx |archive-date=2016-07-10}} (NB. PDFs come without notes: [http://www.johngustafson.net/presentations/Unums2.0.pdf] [https://web.archive.org/web/20160710210430/http://www.johngustafson.net/presentations/Unums2.0.pdf])
• {{cite web |author-first=John L. |author-last=Gustafson |author-link=John Gustafson (scientist) |title=An Energy-Efficient and massively parallel approach to valid numerics |location=OCRAR Seminar |date=2016-06-06 |url=http://www.johngustafson.net/presentations/UnumArithmetic-ICRARseminar.pptx |type=PPT |access-date=2016-07-10 |url-status=live |archive-url=https://web.archive.org/web/20160710201955/http://www.johngustafson.net/presentations/UnumArithmetic-ICRARseminar.pptx |archive-date=2016-07-10}} [http://www.johngustafson.net/presentations/UnumArithmetic-ICRARseminar.pdf] [https://web.archive.org/web/20160710202036/http://www.johngustafson.net/presentations/UnumArithmetic-ICRARseminar.pdf]
• {{cite web |author-first=John L. |author-last=Gustafson |author-link=John Gustafson (scientist) |title=A Radical Approach to Computation with Real Numbers |date=2016 |publisher=SuperFri.org |url=http://www.johngustafson.net/pubs/RadicalApproach.pdf |access-date=2016-07-10 |url-status=live |archive-url=https://web.archive.org/web/20160710221518/http://www.johngustafson.net/pubs/RadicalApproach.pdf |archive-date=2016-07-10}}
• {{cite web |author-first=Ulrich W. |author-last=Kulisch |author-link=Ulrich W. Kulisch |title=Up-to-date Interval Arithmetic from closed intervals to connected sets of real numbers |date=2015 |type=preprint |publisher=Institut für Angewandte und Numerische Mathematik – Karlsruhe Institute of Technology (KIT), Germany |id=ID 15/02 |url=http://www.math.kit.edu/iwrmm/seite/preprints/media/preprint%20nr.%2015-02.pdf |access-date=2016-07-12 |url-status=live |archive-url=https://web.archive.org/web/20160712121728/http://www.math.kit.edu/iwrmm/seite/preprints/media/preprint%20nr.%2015-02.pdf |archive-date=2016-07-12}}
• {{cite web |author-first=Thomas |author-last=Risse |title=Unum – an expedient extension of IEEE 754 |publisher=Institute of Informatics & Automation (IIA), Faculty EEE & CS, Bremen University of Applied Sciences, Germany |date=2016-03-10 |type=presentation |location=London South Bank University (LSBU), UK |url=http://www.weblearn.hs-bremen.de/risse/papers/LSBU/LSBU_Unum.pdf |access-date=2016-07-12 |url-status=live |archive-url=https://web.archive.org/web/20160712122236/http://www.weblearn.hs-bremen.de/risse/papers/LSBU/LSBU_Unum.pdf |archive-date=2016-07-12}}
• {{cite web |title=Prof. W. Kahan's Comments on SORN Arithmetic |date=2016-07-15 |author-first=William M. |author-last=Kahan |author-link=William Kahan |url=https://people.eecs.berkeley.edu/~wkahan/SORNers.pdf |access-date=2016-08-01 |url-status=live |archive-url=https://web.archive.org/web/20160801162751/https://people.eecs.berkeley.edu/~wkahan/SORNers.pdf |archive-date=2016-08-01}}
• {{cite thesis |author-last=Hunhold |author-first=Laslo |date=2016-11-08 |publisher=Universität zu Köln, Mathematisches Institut |type=Bachelor thesis |title=The Unum Number Format: Mathematical Foundations, Implementation and Comparison to IEEE 754 Floating-Point Numbers |url=http://frign.de/publications/2016-11-08-the_unum_number_format.pdf |access-date=2016-10-23 |url-status=live |archive-url=https://web.archive.org/web/20170107142232/http://frign.de/publications/2016-11-08-the_unum_number_format.pdf |archive-date=2017-01-07 |arxiv=1701.00722v1}}
• {{cite book |author-first=Pat H. |author-last=Sterbenz |title=Floating-Point Computation |date=1974-05-01 |edition=1st |series=Prentice-Hall Series in Automatic Computation |publisher=Prentice Hall |location=Englewood Cliffs, New Jersey, USA |isbn=0-13-322495-3}}
• {{cite web |author-first=Skip |author-last=Cave |title=J Programming Language Implementation of 3-bit, 4-bit, 8-bit and 16-bit Precision Unums |date=2016-08-17 |url=https://groups.google.com/d/msg/unum-computing/6A2V7CdHp8A/bLc8KuXABQAJ |access-date=2017-05-03}} (Roger Stokes' download link: [http://www.learningj.com/unumslab.zip])
• {{cite thesis |author-last=Ingole |author-first=Deepak |date=2017-09-28 |publisher=Slovak University of Technology in Bratislava, Slovakia |type=PhD Thesis |title=Embedded Implementation of Explicit Model Predictive Control |url=http://www.kirp.chtf.stuba.sk/assets/publication_info.php?id_pub=1865}}

• {{cite web |title=Conference for Next Generation Arithmetic (CoNGA) |date=2017 |url=https://posithub.org |access-date=2017-11-04 |url-status=live |archive-url=https://web.archive.org/web/20171104235925/https://posithub.org/ |archive-date=2017-11-04}}
• {{cite web |title=SoftPosit |date=2018 |url=https://gitlab.com/cerlane/SoftPosit |access-date=2018-06-13}}
• {{cite web |title=Community Source Code Contribution |date=2018 |url=https://posithub.org/khub_community |access-date=2018-06-13}}
• {{cite web |url=https://www.johndcook.com/blog/2018/04/11/anatomy-of-a-posit-number/ |title=Anatomy of a posit number |date=2018-04-11 |access-date=2019-08-09}}

Category:Floating point types

Category:Articles with example C code