Advanced Micro Devices Corporation

Globalfoundries Inc.

Ati Technologies Ulc

Oswin E Housty

Kevin M Brandl

Gerald Robert Talbot

David M Lynch

Naveen Davanam

Andelon X Tra

Shawn Searles

Tahsin Askar

Edoardo Prete

Thomas H Hamilton

Kathirkamanathan Nadarajah

Tan Peng

Sergey Blotsky

Hary Devapriyan Mahesan

Bernucho S Krishna

Harold H Bautista

A system and method for use in dynamic random-access memory (DRAM) comprising entering into a self-refresh mode of operation, exiting the self-refresh mode of operation in response to commands from a self-refresh state machine memory operation (MOP) array, and updating a device state of the DRAM for a target power management state in response to commands from the MOP array.

Memory context restore, reduction of boot time of a system on a chip by reducing double data rate memory training

Methods for reducing boot time of a system-on-a-chip (SOC) by reducing double data rate (DDR) memory training and memory context restore. Dynamic random access memory (DRAM) controller and DDR physical interface (PHY) settings are stored into a non-volatile memory and the DRAM controller and DDR PHY are powered down. On system resume, a basic input/output system restores the DRAM controller and DDR PHY settings from non-volatile memory, and finalizes the DRAM controller and DDR PHY settings for operation with the SOC. Reducing the boot time of the SOC by reducing DDR training includes setting DRAMs into self-refresh mode, and programing a self-refresh state machine memory operation (MOP) array to exit self-refresh mode and update any DRAM device state for the target power management state. The DRAM device is reset, and the self-refresh state machine MOP array reinitializes the DRAM device state for the target power management state.

Systems, apparatuses, and methods for performing secure system memory training are disclosed. In one embodiment, a system includes a boot media, a security processor with a first memory, a system memory, and one or more main processors coupled to the system memory. The security processor is configured to retrieve first data from the boot media and store and authenticate the first data in the first memory. The first data includes a first set of instructions which are executable to retrieve, from the boot media, a configuration block with system memory training parameters. The security processor also executes a second set of instructions to initialize and train the system memory using the training parameters. After training the system memory, the security processor retrieves, authenticates, and stores boot code in the system memory and releases the one or more main processors from reset to execute the boot code.

Secure system memory training

A method is provided for performing memory operations in response to instructions to perform a double data rate (DDR) memory reference voltage training in the voltage domain by a processing device and determining a DDR memory reference voltage and a DDR memory delay time based upon the memory operation. Computer readable storage media are also provided. A circuit is provided that includes a communication interface portion coupled to a memory and to a processing device. The circuit also includes a circuit portion, coupled to the communication interface portion that has a hardware state machine or an algorithm. The state machine or algorithm provides instructions to the processing device to perform a double data rate (DDR) reference voltage training in the voltage domain.

DDR 2D Vref training

A method is provided for sampling a data strobe signal of a memory cycle and determining a receiver enable phase based upon the data strobe signal. The method also includes performing a memory write cycle and a subsequent read cycle and training a read data strobe cycle at a one-quarter memory clock periodic offset. The method also includes determining a correct receiver enable delay in response to a successful read data strobe training cycle. Computer readable storage media are also provided. An apparatus is provided that includes a communication interface portion that is coupled to a memory portion and to a processing device. The apparatus also includes a first circuit portion, coupled to the communication interface portion. The first circuit portion monitors memory cycles on the communication interface portion, determines a receiver enable cycle phase and train a receiver enable cycle without using receiver enable seed.

DDR receiver enable cycle training

In a system having a plurality of processing nodes, a control node divides a task into a plurality of sub-tasks, and assigns the sub-tasks to one or more additional processing nodes which execute the assigned sub-tasks and return the results to the control node, thereby enabling a plurality of processing nodes to efficiently and quickly perform memory initialization and test of all assigned sub-tasks.

Distributed multi-core memory initialization

Techniques are disclosed relating to resource contention resolution in a pre-memory environment. Prior to system memory being accessible, a resource control processing element controls access to a hardware resource by a plurality of processing elements by granting received requests from the processing elements for access to the resource. The resource control processing element may prioritize requests based on a determined amount of utilization of the hardware resource by individual ones of the processing elements. In one embodiment, processing elements request for information from a bus controller (e.g., an SMBus controller) that is usable to initialize system memory. The resource control processing element may respond to the requests by retrieving the requested information from the controller and providing that information to the processing element or by retrieving the requested information from a cache and providing that information to the processing element.

Pre-memory resource contention resolution

During a boot process of a data processing device having a master bootstrap processor device and multiple slave processor devices, memory associated with the master bootstrap processor is not accessible. Accordingly, the master bootstrap processor communicates configuration information to a slave processor by writing configuration information to a register associated with the slave processor. The slave processor communicates an acknowledgment to the master bootstrap processor in response to reading the configuration information.

Multiprocessor communication device and methods thereof

A method includes establishing a first link between a first processor device and a first memory module at a first time. A second link is established between a second processor device and a second memory module at a second time. In response to receiving a first event indicator, a third link is established between the first processor device and the second memory module at a third time, the third time after the first time and the second time.

Memory control device and methods thereof

A method of booting a multi-processor data processing device includes establishing a link between a first processor and a memory. The link is monitored to determine if, in response to a request from the processor, expected initialization data is communicated between the memory and the first processor. If unexpected data is detected on the link, the link is severed and a new link established between a second processor and the memory to allow the second processor to initiate the boot process. This ensures that, in the event of an error in the boot process at the first processor, the device can complete the boot process, thereby reducing device downtime.

Bootstrap device and methods thereof

Timing delays in a double data rate (DDR) dynamic random access memory (DRAM) controller () are trained. A left edge of passing receive enable delay values is determined (). A final value of a receive data strobe delay value and a final value of a transmit data delay value are trained (). A right edge of passing receive enable delay values is determined using a working value of the receive data strobe delay (); and a final receive enable delay value intermediate between the left edge of passing receive enable delay values and the right edge of passing receive enable delay values is set ().

Method for training dynamic random access memory (DRAM) controller timing delays

Methods and apparatus are provided for use in testing a memory () in a multiprocessor computer system (). The multiprocessor computer system () has a plurality of processing nodes (-) coupled in an array wherein each processing node is coupled to at least one other processing node, and a memory () distributed among the plurality of processing nodes (-). A configuration of the array is determined. An initial configuration of the memory () is also determined. The memory () is tested over the array according to the initial configuration to identify a bad memory element. The initial configuration is modified to form a revised configuration that excludes the bad memory element.

Growing community of inventors

Austin, TX, United States of America

Oswin E Housty