About: Bfloat16 floating-point format

Facets (new session)
Description
Metadata
Settings
- Rule:
- Inverse Functional Properties:
- "Same As":

About: Bfloat16 floating-point format Goto Sponge NotDistinct Permalink

An Entity of Type : owl:Thing, within Data Space : dbpedia.org associated with source document(s)
QRcode icon

http://dbpedia.org/describe/?url=http%3A%2F%2Fdbpedia.org%2Fresource%2FBfloat16_floating-point_format

The bfloat16 (Brain Floating Point) floating-point format is a computer number format occupying 16 bits in computer memory; it represents a wide dynamic range of numeric values by using a floating radix point. This format is a truncated (16-bit) version of the 32-bit IEEE 754 single-precision floating-point format (binary32) with the intent of accelerating machine learning and near-sensor computing. It preserves the approximate dynamic range of 32-bit floating-point numbers by retaining 8 exponent bits, but supports only an 8-bit precision rather than the 24-bit significand of the binary32 format. More so than single-precision 32-bit floating-point numbers, bfloat16 numbers are unsuitable for integer calculations, but this is not their intended use. Bfloat16 is used to reduce the storage r

Attributes	Values
rdf:type	Thing
rdfs:label	Bfloat16 (cs) Bfloat16 (de) Bfloat16 floating-point format (en)
rdfs:comment	V oboru informatiky je bfloat16 (brain floating point) označení konkrétního způsobu reprezentace čísel v počítači pomocí pohyblivé řádové řárky. Jedná se o formát založený na dvojkové soustavě, kde vyjadřuje , dalších 8 bitů vyjadřuje exponent a posledních 7 bitů vyjadřuje . Jedná se v podstatě o variantu dvaatřicetibitového datového typu single definovaného standardem IEEE 754. Byl zaveden zejména pro podporu strojového učení. (cs) bfloat16 (brain floating point with 16 bits) ist die Bezeichnung für ein Gleitkommaformat in Computersystemen. Es handelt sich um ein binäres Datenformat mit einem Bit für das Vorzeichen, 8 Bits für den Exponenten und 7 Bits für die Mantisse. Es handelt sich also um eine in der Mantisse gekürzte Version des IEEE 754 single-Datentyps. bfloat16 wird insbesondere in Systemen für maschinelles Lernen eingesetzt, wie beispielsweise TPUs, sowie bestimmten Intel-Xeon-Prozessoren und Intel FPGAs. (de) The bfloat16 (Brain Floating Point) floating-point format is a computer number format occupying 16 bits in computer memory; it represents a wide dynamic range of numeric values by using a floating radix point. This format is a truncated (16-bit) version of the 32-bit IEEE 754 single-precision floating-point format (binary32) with the intent of accelerating machine learning and near-sensor computing. It preserves the approximate dynamic range of 32-bit floating-point numbers by retaining 8 exponent bits, but supports only an 8-bit precision rather than the 24-bit significand of the binary32 format. More so than single-precision 32-bit floating-point numbers, bfloat16 numbers are unsuitable for integer calculations, but this is not their intended use. Bfloat16 is used to reduce the storage r (en)
dcterms:subject	Floating point types Binary arithmetic
Wikipage page ID	57499027 (xsd:integer)
Wikipage revision ID	1124029922 (xsd:integer)
Link from a Wikipage to another Wikipage	Dynamic range Intelligent sensor 0 16-bit Single-precision floating-point format NaN Machine learning Computer memory Computer number format Hardware acceleration Subnormal number CUDA Type conversion Minifloat AMD ARM architecture AVX-512 Floating point types FPGA Google Brain Floating point Primitive data type Binary arithmetic Half-precision floating-point format Hexadecimal TensorFlow Tensor processing unit AI accelerator Binary number Mixed-precision arithmetic Infinity OpenCL Single precision Nervana Systems Exponent bias IEEE 754 ISO/IEC 10967 Offset binary Xeon Sign bit Significand −0 Exponent Precision (arithmetic)
sameAs	Bfloat16 floating-point format Bfloat16 floating-point format Bfloat16 floating-point format Bfloat16 floating-point format
dbp:wikiPageUsesTemplate	dbt:Confuse dbt:Legend dbt:Lowercase_title dbt:Reflist dbt:Short_description dbt:Data_types dbt:Floating-point
has abstract	V oboru informatiky je bfloat16 (brain floating point) označení konkrétního způsobu reprezentace čísel v počítači pomocí pohyblivé řádové řárky. Jedná se o formát založený na dvojkové soustavě, kde vyjadřuje , dalších 8 bitů vyjadřuje exponent a posledních 7 bitů vyjadřuje . Jedná se v podstatě o variantu dvaatřicetibitového datového typu single definovaného standardem IEEE 754. Byl zaveden zejména pro podporu strojového učení. (cs) bfloat16 (brain floating point with 16 bits) ist die Bezeichnung für ein Gleitkommaformat in Computersystemen. Es handelt sich um ein binäres Datenformat mit einem Bit für das Vorzeichen, 8 Bits für den Exponenten und 7 Bits für die Mantisse. Es handelt sich also um eine in der Mantisse gekürzte Version des IEEE 754 single-Datentyps. bfloat16 wird insbesondere in Systemen für maschinelles Lernen eingesetzt, wie beispielsweise TPUs, sowie bestimmten Intel-Xeon-Prozessoren und Intel FPGAs. (de) The bfloat16 (Brain Floating Point) floating-point format is a computer number format occupying 16 bits in computer memory; it represents a wide dynamic range of numeric values by using a floating radix point. This format is a truncated (16-bit) version of the 32-bit IEEE 754 single-precision floating-point format (binary32) with the intent of accelerating machine learning and near-sensor computing. It preserves the approximate dynamic range of 32-bit floating-point numbers by retaining 8 exponent bits, but supports only an 8-bit precision rather than the 24-bit significand of the binary32 format. More so than single-precision 32-bit floating-point numbers, bfloat16 numbers are unsuitable for integer calculations, but this is not their intended use. Bfloat16 is used to reduce the storage requirements and increase the calculation speed of machine learning algorithms. The bfloat16 format was developed by Google Brain, an artificial intelligence research group at Google. The bfloat16 format is utilized in Intel AI processors, such as Nervana NNP-L1000, Xeon processors (AVX-512 BF16 extensions), and Intel FPGAs, Google Cloud TPUs, and TensorFlow. ARMv8.6-A, AMD ROCm, and CUDA also support the bfloat16 format. On these platforms, bfloat16 may also be used in mixed-precision arithmetic, where bfloat16 numbers may be operated on and expanded to wider data types. (en)
prov:wasDerivedFrom	wikipedia-en:Bfloat16_floating-point_format?oldid=1124029922&ns=0
page length (characters) of wiki page	29531 (xsd:nonNegativeInteger)
foaf:isPrimaryTopicOf	wikipedia-en:Bfloat16_floating-point_format
is Link from a Wikipage to another Wikipage of	DL Boost Power10 Minifloat AVX-512 Advanced Vector Extensions Ampere (microarchitecture) Floating-point arithmetic List of Intel CPU microarchitectures Half-precision floating-point format AArch64 AI accelerator Advanced Matrix Extensions Mixed-precision arithmetic CPUID IEEE 754 BF16 Bf16 Bfloat16 Brain floating-point format
is Wikipage redirect of	BF16 Bf16 Bfloat16 Brain floating-point format
is foaf:primaryTopic of	wikipedia-en:Bfloat16_floating-point_format

Faceted Search & Find service v1.17_git139 as of Feb 29 2024

Alternative Linked Data Documents: ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 08.03.3330 as of Mar 19 2024, on Linux (x86_64-generic-linux-glibc212), Single-Server Edition (378 GB total memory, 50 GB memory in use)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software