US 2023/0401059 A1 · METHOD AND DEVICE FOR VARIABLE PRECISION COMPUTING

Abstract

The present disclosure relates to a floating-point computation circuit comprising: an internal memory (104, 114) storing one or more floating-point values in a first format; status registers (124) defining a plurality of floating-point number format types associated with corresponding identifiers, each format type indicating at least a maximum size (BIS, MBB); and a load and store unit (108, 118) for loading floating-point values from and storing floating-point values to an external memory (120, 122), the load and store unit (108, 118) being configured: to receive, in relation with a first store operation, a first floating-point value from the internal memory (104, 114) and a first of said identifiers; and to convert the first floating-point value from the first format to a first external memory format having a maximum size (BIS, MBB) defined by the floating-point number format type designated by the first identifier.

Claims17

Dependent claims

A floating-point computation circuit comprising:

an internal memory storing one or more floating-point values in a first format;
status registers defining a plurality of floating-point number format types associated with corresponding identifiers, each format type indicating at least a maximum size;
and
a load and store unit for loading floating-point values from an external memory to the internal memory and storing floating-point values from the internal memory to the external memory, the load and store unit being configured:
to receive, in relation with a first store operation, a first floating-point value from the internal memory and a first of said identifiers;
and
to convert the first floating-point value from the first format to a first external memory format having a maximum size defined by the floating-point number format type designated by the first identifier.

4 more elements

A method of floating-point computation comprising:

storing, by an internal memory of a floating-point computation device, one or more floating-point values in a first format;
loading, by a load and store unit of a floating-point computation device, floating-point values from an external memory to the internal memory, and storing, by the load and store unit, a first floating-point value from the internal memory to the external memory, wherein the load and store unit is configured to perform said storing by:
receiving, in relation with a first store operation, the first floating-point value from the internal memory and a first identifier;
obtaining, from status registers defining a plurality of floating-point number format types associated with corresponding identifiers, at least a maximum size associated with the first identifier;
and
converting the first floating-point value from the first format to an external memory format having a maximum size defined by the floating-point number format type designated by the first identifier.

3 more elements

Dependencies are inferred from the claim text.

Method and device for variable precision computing

Life of this patent

Method and device for variable precision computing

Life of this patent

Abstract

Claims17