Apparatus and method for converting a floating-point value from half precision to single precision
AssigneeIntel Corporation
InventorsRobert Valentine, Mark Charney, Raanan Sade, Elmoustapha Ould-Ahmed-Vall, Jesus Corbal
Life of this patent
Estimating the term…
Abstract
An embodiment of the invention is a processor including execution circuitry to, in response to a decoded instruction, convert a half-precision floating-point value to a single-precision floating-point value and store the single-precision floating-point value in each of the plurality of element locations of a destination register. The processor also includes a decoder and the destination register. The decoder is to decode an instruction to generate the decoded instruction.
Claims20
A processor comprising:
- a decoder to decode a first instruction to generate a first decoded instruction;
- a destination register having a plurality of element locations;
- execution circuitry to execute the first decoded instruction, wherein execution of the first decoded instruction is to include converting a half-precision floating-point value to a single-precision floating-point value and storing the single-precision floating-point value in each of the plurality of element locations of the destination register.
A method comprising:
- decoding a first instruction to generate a first decoded instruction;
- determining that a broadcast indicator is set;
- converting, in response to the first decoded instruction, a half-precision floating-point value to a single-precision floating-point value;
- and
- storing the single-precision floating-point value in each of the plurality of element locations of a destination register.
A machine-readable medium having program code stored thereon which, when executed by a machine, causes the machine to perform the operations of:
- decoding a first instruction to generate a first decoded instruction;
- determining that a broadcast indicator is set;
- converting, in response to the first decoded instruction, a half-precision floating-point value to a single-precision floating-point value;
- and
- storing the single-precision floating-point value in each of the plurality of element locations of a destination register in the machine.
The machine-readable medium of claim 19, wherein the half-precision floating-point value is one of a first set of half-precision floating-point values to be represented by a first vector to be stored in the lowest element of the first source location.
Dependencies are inferred from the claim text.