Intel Corporation

Other

Stanley Steve Kulick

Randy B Osborne

James W Alexander

Thomas J Holman

Thomas P Thomas

Sin Sim Tan

Xiuting C Man

Erin D Francom

Philip Abraham

Mark A Heap

Rajesh S Pamujula

Siddhartha Chhabra

Uday R Savagaonkar

Christopher P Mozak

Men Long

Linda J Rankin

Alpa T Narendra Trivedi

Jeffrey L Miller

Suneeta Sah

Ramadass Nagarajan

Bryan L Spry

Erik Hallnor

Zhongying Zhang

James M Shehadi

Carlos Ornelas

Edgar Borrayo

Amir Ali Radjai

Chitra Natarajan

Jason Bessette

Varin Udompanyanan

Bezan Kapadia

Hrishikesh S Pai

Michael Cekleov

Methods, systems, and apparatuses relating to package on package memory refresh and self-refresh rate management are described. In one embodiment, an apparatus includes a processor die, a dynamic memory die mounted to and overlapping the processor die, a first thermal sensor of the processor die disposed adjacent to a first hot spot from a first type of workload and a second thermal sensor of the processor die disposed adjacent to a second hot spot from a second type of workload, and a hardware control circuit of the processor die to cause a refresh of a capacitor of the dynamic memory die when either of an output of the first thermal sensor exceeds a first threshold value and an output of the second thermal sensor exceeds a second threshold value.

Apparatuses, methods, and systems for package on package memory refresh and self-refresh rate management

A clock domain crossing can occur earlier in time by detection of when a data signal will coincide with a TSV (time slot valid) signal but the valid signal associated with the data signal will not coincide with a TSV. In response to such a detection, the domain crossing circuit can send the valid signal early, resulting in a valid signal sent on an earlier TSV and the data signal sent on a TSV. In one embodiment, such a system can cause a data signal to be received in a slower clock domain on a first edge of the slower clock signal after the data is queued in the faster clock domain. The sending of the early valid indication can reduce latency in transferring data between clock domains.

Preserving deterministic early valid across a clock domain crossing

A processor is described that includes one or more processing cores. The processor includes a memory controller to interface with a system memory having a protected region and a non protected region. The processor includes a protection engine to protect against active and passive attacks. The processor includes an encryption/decryption engine to protect against passive attacks. The protection engine includes bridge circuitry coupled between the memory controller and the one or more processing cores. The bridge circuitry is also coupled to the protection engine and the encryption/decryption engine. The bridge circuitry is to route first requests directed to the protected region to the protection engine and to route second requests directed to the non protected region to the encryption/decryption engine.

Bridging circuitry between a memory controller and request agents in a system having multiple system memory protection schemes

A processing device and method for cache control including tracking updates to the line state of a cache superline are described. In response to a request pertaining to a superline, a cache controller of the processing device can perform one or more read-modify-write (RMW) operations to (a) a line state vector of a line state array and (b) a counter of the line state array. Based on a determination that one or more requests to the superline have completed, the line state vector from the line state array can be written to a tag array. The cache controller can track pending line state updates to a superline outside of the tag array, and a line state update can occur in the cache controller, rather than awaiting completion of all outstanding operations on a superline. Updates to multiple line states can be maintained simultaneously, and up-to-date ECCs computed.

Method and apparatus for cache line state update in sectored cache with line state tracker

Techniques and apparatuses for clock crossing. A reset circuit on a first die generates a forwarded FIFO reset signal synchronous to a reference clock that identifies a single edge. A clock generation circuit on the first die generates the reference clock signal. Control circuitry on the first die generates a forwarded signal, synchronous to the forwarded clock that identifies a forwarded clock edge with fixed timing relationship to the forwarded clock edge a transmit PLL locks to the single reference edge. A phase locked loop (PLL) on a second die is coupled to receive the reference clock signal, the PLL to generate a local clock signal. A circular FIFO with a write pointer advanced by the forwarded clock and a read pointer advanced by the local clock.

Deterministic clock crossing

Initialization in multiple clock domains. A first die having a master initialization component generates initialization commands. A local initialization agent on the first die is coupled to receive the initialization commands. The local initialization agent manages initialization of one or more components on the first die. A remote initialization agent on a second die is coupled to receive the initialization commands. The remote initialization agent manages initialization of one or more components on the second die. The master initialization component receives acknowledgement messages from the local initialization agent and the remote initialization agent to manage conflicts and dependencies between the local initialization agent and the remote initialization agent and synchronizes events in multiple clock domains that share a reference clock signal by signaling in the reference clock domain.

Sideband initialization

Mechanisms for clock gating. A clock generation circuit provides a clock signal over a clock signal distribution network within an integrated circuit package. Gating elements within the clock signal distribution network disable the clock signal to one or more portions of the clock signal distribution network. A digital locked loop (DLL) maintains settings without tracking when the clock signal is disabled.

Mechanisms for clock gating

Methods and apparatus to improve throughput and efficiency in memory devices are described. In one embodiment, a memory controller may include scheduler logic to issue read or write requests to a memory device in an optimal fashion, e.g., to maximize bandwidth and/or reduce latency. Other embodiments are also disclosed and claimed.

Scalable schedulers for memory controllers

An apparatus and system for controlling traffic on an on-chip network. Embodiments of the apparatus comprise single-ended transmission circuitry and single-ended receiving circuitry on a first chip for coupling with a second chip, the transmission circuitry having impedance matching and lacking equalization, the receiving circuitry lacking equalization, the transmission circuitry and the receiving circuitry having statically configurable features and organized in clusters, wherein the clusters have the same physical layer circuitry design for different configurations of the configurable features, the configurable features including half-duplex mode and full-duplex mode, wherein the first chip and the second chip are on the same package, and wherein a plurality of conductive lines for coupling the first chip with the second chip are matched.

On-package input/output clustered interface having full and half-duplex modes

According to one embodiment a computer system is disclosed. The computer system includes memory. The memory includes two or more rows, where each row has a plurality of memory devices. The computer system also includes a chipset. The chipset includes a detection/correction circuit to detect single and double symbol errors and correct single symbol errors for each memory row, and markers to maintain a log of errors within each memory row.

Mechanism for adjacent-symbol error correction and detection

An apparatus and method are disclosed. In one embodiment, the apparatus trains a memory link using a signal alignment unit. The signal alignment unit aligns a read data strobe signal that is transmitted on the link with the center of a read data eye transmitted on the link. Next, the signal alignment unit aligns a receive enable signal that is transmitted on the link with the absolute time that data returns the data lines of the link a column address strobe signal is sent to the memory coupled to the link. Next, the signal alignment unit aligns a write data strobe signal transmitted on the link with the link's clock signal. Finally, the signal alignment unit aligns the center of the write data eye transmitted on the link with the write data strobe transmitted on the link.

Memory link training

A forwarding device compares a first address from a first coherent input/output (I/O) transaction with an address from at least one processor-issued transaction to determine if an address conflict exists. The forwarding device completes a first processor-issued transaction of the at least one processor-issued transaction if the address conflict exists and rejects the first coherent I/O transaction. The forwarding device holds remaining processor transactions of the at least one processor-issued transaction that have an address conflict with the first address of the first coherent I/O transaction. The forwarding device transmits the first coherent I/O transaction to an external I/O device, waits for the first coherent I/O transaction to return from the external I/O device, and completes the first coherent I/O transaction. The forwarding device releases the remaining processor transactions once the first coherent I/O transaction has been completed.

Mechanism to guarantee forward progress for incoming coherent input/output (I/O) transactions for caching I/O agent on address conflict with processor transactions

Growing community of inventors

Portland, OR, United States of America

Stanley Steve Kulick