Apple Inc.

Mellanox Technologies, Ltd.

Omri Flint

Achmed Rumi Zahir

Lior Zimet

Inder M Sodhi

Liran Fishel

Ami Schwartzman

Benny Koren

Itamar Rabenstein

Ido Bukspan

Amiad Marelli

Oded Wertheim

Dror Aharoni

Elad Harush

Li Rosenbaum

Techniques are disclosed that pertain to synchronizing power states between integrated circuit dies. A system includes an integrated circuit that includes a plurality of integrated circuit dies coupled together. A particular integrated circuit die may include a primary power manager circuit and one or more remaining integrated circuit dies include respective secondary power manager circuits. The primary power manager circuit is configured to issue a transition request to the secondary power manager circuits to transition their integrated circuit dies from a first power state to a second power state. A given secondary power manager circuit is configured to receive the transition request, transition its integrated circuit die to the second power state, and issue an acknowledgement to the primary power manager circuit that its integrated circuit die has been transitioned to the second power state. Techniques are further disclosed relating to managing latency tolerance events within a multi-die integrated circuit.

Blocking latency events in multi-die architecture

A System on Chip (SoC) includes a processor, a parity generation circuit, and a dispatcher circuit. The processor is configured to produce store instructions for storing data blocks in a Non-Volatile-Memory (NVM). The parity generation circuit is configured to calculate parity blocks over the data blocks in accordance with a redundant storage scheme, to send the parity blocks to the NVM, and to produce completion notifications with respect to the parity blocks. The dispatcher circuit is configured to dispatch the store instructions to the NVM. The processor is further configured to send one or more parity-barrier instructions that specify synchronization barriers over the store instructions and the parity, and the dispatcher circuit is configured to dispatch the store instructions to the NVM in compliance with the parity-barrier instructions and the completion notifications.

Dispatcher for improved-performance non-volatile memory store operations

Synchronizing power state changes between multiple dies

Multi-die power synchronization

A method in a network element includes processing input packets using a set of two or more functions that are defined over parameters of the input packets. Each function in the set produces respective interim actions applied to the input packets and the entire set produces respective end-to-end actions applied to the input packets. An end-to-end mapping, which maps the parameters of at least some of the input packets directly to the corresponding end-to-end actions, is cached in the network element. The end-to-end mapping is queried with the parameters of a new input packet. Upon finding the parameters of the new input packet in the end-to-end mapping, an end-to-end action mapped to the found parameters is applied to the new input packet, without processing the new input packet using the set of functions.

End-to-end cache for network elements

Growing community of inventors

Ramat Hasharon, Israel

Omri Flint