Intel Corporation

Roni Rosner

Michael Behar

Zigi Walter

Moshe Maor

Oren Agam

Ronen Gabbai

Abraham Mendelson

Ronny Ronen

Israel Diamand

Oz Shitrit

Yoav Almog

Ravi Venkatesan

Shlomi Shua

Hong Wang

Milind Baburao Girkar

Ohad Falik

Alexander Gendler

Wei Ey Li

John Thomas Shen

Perry H Wang

Xinmin Tian

Gerolf F Hoflehner

Satish Narayanasamy

Henrietta Bezbroz

Shlomi Alkalay

Naftali Schwartz

Micha G Moffie

Daniel LaVery

Methods, apparatus, systems and articles of manufacture are disclosed to configure heterogenous components in an accelerator. An example apparatus includes a graph compiler to identify a workload node in a workload and generate a selector for the workload node, and the selector to identify an input condition and an output condition of a compute building block, wherein the graph compiler is to, in response to obtaining the identified input condition and output condition from the selector, map the workload node to the compute building block.

Methods and apparatus to configure heterogenous components in an accelerator

Methods, apparatus, systems and articles of manufacture are disclosed that enable out-of-order pipelined execution of static mapping of a workload to one or more computational building blocks of an accelerator. An example apparatus includes an interface to load a first number of credits into memory; a comparator to compare the first number of credits to a threshold number of credits associated with memory availability in a buffer; and a dispatcher to, when the first number of credits meets the threshold number of credits, select a workload node of the workload to be executed at a first one of the one or more computational building blocks.

Methods and apparatus to enable out-of-order pipelined execution of static mapping of a workload

Disclosed embodiments relate to a shared read request (SRR) using a common request tracker (CRT) as a temporary cache. In one example, a multi-core system includes a memory and a memory controller to receive a SRR from a core when a Leader core is not yet identified, allocate a CRT entry and store the SRR therein, mark it as a Leader, send a read request to a memory address indicated by the SRR, and when read data returns from the memory, store the read data in the CRT entry, send the read data to the Leader core, and await receipt, unless already received, of another SRR from a Follower core, the other SRR having a same address as the SRR, then, send the read data to the Follower core, and deallocate the CRT entry.

Shared read&#x2014;using a request tracker as a temporary read cache

Methods and apparatus to implement multiple inference compute engines are disclosed herein. A disclosed example apparatus includes a first inference compute engine, a second inference compute engine, and an accelerator on coherent fabric to couple the first inference compute engine and the second inference compute engine to a converged coherency fabric of a system-on-chip, the accelerator on coherent fabric to arbitrate requests from the first inference compute engine and the second inference compute engine to utilize a single in-die interconnect port.

Methods and apparatus to implement multiple inference compute engines

An optimization unit to search for two or more candidate instructions in an instruction trace and to merge the two or more candidate instructions into a single instruction with multiple data (SIMD) according to a depth of a trace dependency and a common operation code of the two or more candidate instructions.

Method and apparatus to vectorize multiple input instructions

Embodiments of the present invention provide a method, apparatus and system which may include splitting a dependency chain into a set of reduced-width dependency chains; mapping one or more dependency chains onto one or more clustered dependency chain processors, wherein an issue-width of one or more of the clusters is adapted to accommodate a size of the dependency chains; and/or processing in parallel a plurality of dependency chains of a trace. Other embodiments are described and claimed.

System, method and apparatus for dependency chain processing

A cache management logistics controls a transfer of a trace. A first cache couples to the cache management logistics to evict the trace based on a replacement mechanism. A second cache couples to the cache management logistics to receive the trace based on a number of accesses to the trace.

Trace cache filtering

In an embodiment, a method includes receiving a binary of a program code. The binary is based on a first instruction set architecture. The method also includes translating the binary, wherein the translated binary is based on a combination of the first instruction set architecture and a second instruction set architecture.

Controlling compatibility levels of binary translations between instruction set architectures

A distributed system is provided for apportioning an instruction stream into multiple segments for processing in multiple parallel processing units, and for merging the processed segments into a single processed instruction stream having the same sequential relative order as the original instruction stream. Tags may be attached to each segment after apportioning to indicate the order in which the various segments are to be merged. In one embodiment, the end of each segment includes a tag indicating the unit to which the next instruction in the original instruction sequence is directed.

Reordering serial data in a system with parallel processing flows

Growing community of inventors

Binyamina, Israel

Roni Rosner