Back to results
US 12,443,398 B2Granted
Kernel fusion for machine learning
AssigneeNVIDIA Corporation
InventorsAndrew Kerr, Mike Murphy, Mostafa Hagog, Julien Demouth, John Tran
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and techniques are presented to compile code. In at least one embodiment, one or more compilers are to compile one or more compiled portions of code with one or more intermediate representations of one or more portions of code.
Claims28
One or more processors, comprising:
- processing circuitry to:
- compile one or more intermediate representation (IR) host code software programs comprising one or more user-specified functions;
- separately compile one or more object code device software programs comprising one or more call sites that correspond to the one or more user-specified functions;
- and
- compile the one or more IR host code software programs together with the one or more object code device software programs to generate a single object code software program.
2 more elements
A system comprising:
- one or more processors comprising processing circuitry to:
- compile one or more intermediate representation (IR) host code software programs comprising one or more user-specified functions;
- separately compile one or more object code device software programs comprising one or more call sites that correspond to the one or more user-specified functions;
- compile the one or more IR host code software programs together with the one or more object code device software programs to generate a single object code software program;
- and
- memory to store the single object code software program.
3 more elements
A non-transitory machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
- perform one or more compilers to:
- compile one or more intermediate representation (IR) host code software programs comprising one or more user-specified functions;
- separately compile one or more object code device software programs comprising one or more call sites that correspond to the one or more user-specified functions;
- and
- compile the one or more IR host code software programs together with the one or more object code device software programs to generate a single object code software program.
2 more elements
A method, comprising:
- performing one or more compilers to:
- compile one or more intermediate representation (IR) host code software programs comprising one or more user-specified functions to be performed by one or more graphics processors;
- separately compile one or more object code device software programs comprising one or more call sites that correspond to the one or more user-specified functions;
- and
- compile the one or more IR host code software programs together with the one or more object code device software programs to generate a single object code software program.
2 more elements
Dependencies are inferred from the claim text.