Advanced Micro Devices Corporation

Samsung Electronics Co., Ltd.

Globalfoundries Inc.

Other

Advanced Silicon Technologies LLC

Mitchell K Alsup

Geoffrey S Strongin

Mark Douglas Hummel

Andrew William Lueck

Michael Haertel

Benjamin T Sander

Brian D McMinn

Gregory William Smaus

James K Pickett

David Conrad Tannenbaum

Roger Dwain Isaac

Boris Beylin

Michael A Filippo

Srinivasan Iyer

Maxim Lukyanov

Alexander Grosul

William Alexander Hughes

Michael C Shebanow

Jerry D Moench

Teik-Chung Tan

Krishnan V Ramani

David Edward Kroesche

Ramsey W Haddad

Yang Jiao

Joe Jones

Rama S Gopal

Christopher James Goodman

Derek Lentz

SungSoo Park

Michael Filippo

A transcendental calculation unit includes a configuration table storing a set of constants and provide a selected one of the constants, a power series multiplier that iteratively develops a power series, a coefficient series multiplier and accumulator that develops an accumulated product of the power series and the constant, and a round and normalize stage that rounds the accumulated product and normalizes rounded product.

Transcendental calculation unit apparatus and method

A Transcendental Calculation Unit includes a Configuration Table storing a set of constants and provide a selected one of the constants, a Power Series Multiplier that iteratively develops a power series, a Coefficient Series Multiplier and Accumulator that develops an accumulated product of the power series and the constant, and a Round and Normalize Stage that rounds the accumulated product and normalizes rounded product.

A system and method for binding instructions to a graphical processing unit (GPU) includes a GPU configured to receive bindlessly compiled instructions and interpret the bindlessly compiled instruction at runtime to identify a needed conversion The GPU generates a conversion information based on the bindlessly compiled instruction and needed conversion and converts the bindlessly compiled instruction according to the conversion information to generate a bound format instruction. The GPU may then execute the bound format instruction.

Efficient interface and transport mechanism for binding bindless shader programs to run-time specified graphics pipeline configurations and objects

According to one general aspect, an apparatus may include a network of node circuits and a central arbiter circuit. The network of node circuits is within an integrated circuit, wherein the network includes a plurality of segments. The central arbiter circuit may be configured to schedule a routing of a message between a pair of node circuits in the network, wherein the routing includes a guaranteed latency between the pair of node circuits.

Central arbitration scheme for a highly efficient interconnection topology in a GPU

A graphics processing unit may include a register file memory, a processing element (PE) and a load-store unit (LSU). The register file memory includes a plurality of registers. The PE is coupled to the register file memory and processes at least one thread of a vector of threads of a graphical application. Each thread in the vector of threads are processed in a non-stalling manner. The PE stores data in a first predetermined set of the plurality of registers in the register file memory that has been generated by processing the at least one thread and that is to be routed to a first stallable logic unit that is external to the PE. The LSU is coupled to the register file memory, and the LSU accesses the data in the first predetermined set of the plurality of registers and routes to the first stallable logic unit.

System and method for maintaining data in a low-power structure

A method for improving power, performance, area (PPA) for mixed precision computations in a processing environment. The method includes determining a braiding factor as a number of units of work encoded into a physical thread. A value of the braiding factor is determined based on a mix of precision requirements presented for individual units of work. Units of work are classified as instructions for applied code transformation based on associated precision requirements for the processing environment. Instruction inputs from specified registers are packed together into a destination register according to the determined value of the braiding factor. The packed instructions presented in vector form are executed with an instruction set architecture configured for executing packed instructions of different precisions.

Architecture and execution for efficient mixed precision computations in single instruction multiple data/thread (SIMD/T) devices

A method for computing in a thread-based environment provides manipulating an execution mask to enable and disable threads when executing multiple conditional function clauses for process instructions. Execution lanes are controlled based on execution participation for the process instructions for reducing resource consumption. Execution of particular one or more schedulable structures that include multiple process instructions are skipped based on the execution mask and activating instructions.

Control flow in a thread-based environment without branching

A method for executing instructions in a thread processing environment includes determining a multiple requirements that must be satisfied and resources that must be available for executing multiple instructions. The multiple instructions are encapsulated into a schedulable structure. A header is configured for the schedulable structure with information including the determined multiple requirements and resources. The schedulable structure is schedule for executing each of the multiple instructions using the information.

Trace-based instruction execution processing

A method for graphics processing includes generating one or more transcendental instructions in a graphics processing unit (GPU). Micro-code is formed for processing the one or more transcendental instructions in the GPU. The micro-code is processed using an iterative process including cubic interpolation and an evaluation of a cubic polynomial.

Micro-coded transcendental instruction execution

Various embodiments of methods and systems for implementing a microprocessor that includes a trace cache and attempts to transition fetching from instruction cache to trace cache only on label boundaries are disclosed. In one embodiment, a microprocessor may include an instruction cache, a branch prediction unit, and a trace cache. The prefetch unit may fetch instructions from the instruction cache until the branch prediction unit outputs a predicted target address for a branch instruction. When the branch prediction unit outputs a predicted target address, the prefetch unit may check for an entry matching the predicted target address in the trace cache. If a match is found, the prefetch unit may fetch one or more traces from the trace cache in lieu of fetching instructions from the instruction cache.

Transitioning from instruction cache to trace cache on label boundaries

In one embodiment, a system comprises one or more input/output (I/O) devices; an I/O memory management unit (IOMMU) coupled to receive memory requests sourced by the I/O devices and configured to provide address translation for the memory requests; and a virtual machine monitor (VMM) configured to manage one or more virtual machines on the system, wherein the VMM is configured to virtualize the IOMMU, providing one or more virtual IOMMUs for use by one or more virtual machines.

Virtualizing an IOMMU

In an embodiment, an input/output (I/O) memory management unit (IOMMU) comprises at least one memory configured to store translation data; and control logic coupled to the memory and configured to translate an I/O device-generated memory request using the translation data. The translation data corresponds to one or more device table entries in a device table stored in a memory system of a computer system that includes the IOMMU, wherein the device table entry for a given request is selected by an identifier corresponding to the I/O device that generates the request. The translation data further corresponds to one or more I/O page tables, wherein the selected device table entry for the given request includes a pointer to a set of I/O page tables to be used to translate the given request.

Direct memory access (DMA) address translation in an input/output memory management unit (IOMMU)

Various embodiments of methods and systems for implementing a set of microcode operations corresponding to a microcoded instruction as a microcode subroutine are disclosed. In one embodiment, a microprocessor includes a dispatch unit configured to dispatch operations and a scheduler coupled to the dispatch unit and configured to schedule dispatched operations for execution. In response to receiving a microcoded instruction, the dispatch unit is configured to dispatch a microcode subroutine call operation that specifies a tag identifying a microcode subroutine.

System and method of implementing microcode operations as subroutines

Address translation for input/output (I/O) devices and interrupt remapping for I/O devices in an I/O memory management unit (IOMMU)

A method for providing proactive synchronization in a computer system includes a processor requesting exclusive access to a given memory resource. The request may include one or more addresses associated with the given memory resource. The method also includes comparing each of the addresses in the request to each address in a plurality of sets of addresses. Each address in the sets of addresses may correspond to a respective memory resource to which a requestor has exclusive access. In addition, in response to any address of the one or more addresses matching any address in the plurality of sets of addresses, the method includes returning a count value associated with the set including the matching address. The count value may be indicative of the number of requestors contending for the matching address. Software may utilize this count value to proactively choose an item with lower contention probabilities in subsequent attempts.

Growing community of inventors

Austin, TX, United States of America

Mitchell K Alsup