Abstract: Embodiments are generally directed apparatuses, methods, techniques and so forth to select two or more processing units of the plurality of processing units to process a workload, and configure a circuit switch to link the two or more processing units to process the workload, the two or more processing units each linked to each other via paths of communication and the circuit switch.

Abstract: Embodiments are generally directed apparatuses, methods, techniques and so forth determine an access level of operation based on an indication received via one or more network links from a pod management controller, and enable or disable a firmware update capability for a firmware device based on the access level of operation, the firmware update capability to change firmware for the firmware device. Embodiments may also include determining one or more configuration settings of a plurality of configuration settings to enable for configuration based on the access level of operation, and enable configuration of the one or more configuration settings.

Abstract: Examples may include sleds for a rack in a data center including physical compute resources and memory for the physical compute resources. The memory can be disaggregated, or organized into first level and second level memory. A first sled can comprise the physical compute resources and a first set of physical memory resources while a second sled can comprise a second set of physical memory resources. The first set of physical memory resources can be coupled to the physical compute resources via a local interface while the second set of physical memory resources can be coupled to the physical compute resources via a fabric.

Abstract: Examples may include techniques to allocate physical accelerator resources from pools of accelerator resources. In particular, virtual computing devices can be composed from physical resources and physical accelerator resources dynamically allocated to the virtual computing devices. The present disclosure provides that physical accelerator resources can be dynamically allocated, or composed, to a virtual computing device despite not being physically coupled to other components in the virtual device.

Abstract: Embodiments are generally directed to apparatuses, method, techniques, and so forth including a memory coupled to processing circuitry, wherein the memory stores a firmware interface table and the firmware interface table comprises an entry to identify a non-enumerable resource. Embodiments include accessing the firmware interface table to identify the non-enumerable resource.

Abstract: A processor includes a front end including a decoder to decode an instruction, a scheduler to assign execution of the instruction to a core, and a core to execute the instruction. The instruction specifies that interrupts such as corrected machine check interrupts are to be selectively suppressed. The processor further includes an error handling unit including logic to determine that an interrupt caused by an error is to be created and that an error consumer has requested interrupt notification. The error handling unit further includes logic to, based on the instruction specifying that interrupts are to be selectively suppressed, send the interrupt to a producer that issued the instruction rather than the error consumer.

Abstract: Technologies for providing dynamically managed quality of service in a distributed storage system include an apparatus having a processor. The processor is to determine capabilities of one or more compute devices of the distributed storage system. The processor is also to obtain an indicator of a target quality of service to be provided by the distributed storage system, determine target performance metrics associated with the target quality of service, determine target configuration settings for the one or more compute devices of the distributed storage system to provide the target quality of service, configure the one or more compute devices with the target configuration settings, determine whether a present performance of the distributed storage system satisfies the target quality of service, and reconfigure the one or more compute devices in response to a determination that the target quality of service is not satisfied. Other embodiments are described and claimed.

Abstract: Particular embodiments described herein provide for an electronic device that can be configured to store data in a secure domain in a cloud network, create encryption keys, where each encryption key is to provide a different type of access to the data, and store the encryption keys in a secure domain key store in the cloud network. In an example, each encryption key provides access to a different version of the data. In another example, a counter engine stores the location of each version of the data in the cloud network.

Abstract: A processor of an aspect includes a decode unit to decode a read from memory instruction. The read from memory instruction is to indicate a source memory operand and a destination storage location. The processor also includes an execution unit coupled with the decode unit. The execution unit, in response to the read from memory instruction, is to read data from the source memory operand, store an indication of defective data in an architecturally visible storage location, when the data is defective, and complete execution of the read from memory instruction without causing an exceptional condition, when the data is defective. Other processors, methods, systems, and instructions are disclosed.

Abstract: A processor includes a logic to determine an error condition reported in an error bank. The error bank is communicatively coupled to the processor and is associated with logical processors of the processor. The processor includes another logic to generate an interrupt indicating the error condition. The processor includes yet another logic to selectively send the interrupt to a single one of the logical processors associated with the error bank.

Abstract: A non-volatile random access memory (NVRAM) is used in a computer system to enhance support to sleep states. The computer system includes a processor, a non-volatile random access memory (NVRAM) that is byte-rewritable and byte-erasable, and power management (PM) module. A dynamic random access memory (DRAM) provides a portion of system address space. The PM module intercepts a request initiated by an operating system for entry into a sleep state, copies data from the DRAM to the NVRAM, maps the portion of the system address space from the DRAM to the NVRAM, and turns off the DRAM when transitioning into the sleep state. Upon occurrence of a wake event, the PM module returns control to the operating system such that the computer system resumes working state operations without the operating system knowing that the portion of the system address space has been mapped to the NVRAM.

Abstract: Provided are a method, system, computer readable storage medium, and switch for configuring a switch to assign partitions in storage devices to compute nodes. A management controller configures the switch to dynamically allocate partitions of at least one of the storage devices to the compute nodes based on a workload at the compute node.

Abstract: Embodiments are generally directed to high capacity energy backed memory with off device storage. A memory device includes a circuit board; multiple memory chips that are installed on the circuit board; a controller to provide for backing up contents of the memory chips when a power loss condition is detected; a connection to a backup energy source; and a connection to a backup data storage that is separate from the memory device.

Abstract: An apparatus is described that includes memory controller logic circuitry to interface with a memory side cache of a multi-level system memory. The memory controller logic circuitry includes error tracking circuitry to track errors of cache line slots in the memory side cache. The memory controller logic circuitry also comprises faulty list circuitry to store identifiers of faulty cache line slots that are deemed to be excessively error prone. The memory controller logic circuitry is to declare a miss in the memory side cache for requests that map to cache line slots identified in the faulty list.

Abstract: Methods and apparatus for highly available rack management in Rack Scale environment. Rack Management Modules (RMMs) are configured to manage power and thermal zones in a rack including a plurality of pooled system drawers, wherein each pooled system drawer is associated with a respective power zone including power sensors and power control devices and a respective thermal zone including thermal sensors and thermal devices. During operation, one of the RMMs is implemented as a master RMM, and the other is implemented as a slave RMM. The master RMM is used to monitor the power and thermal zones. State information is periodically synchronized between the master RMM and the slave RMM. The RMMs are further configured to perform a fail-over operation in connection with a failed or failing RMM, where after the fail-over operation the slave becomes the new master RMM and the previous master RMM becomes the new slave.

Abstract: A mechanism for PCIe cable topology discovery in a Rack Scale Architecture (RSA) and associated methods, apparatus, and systems. Pooled system drawers installed in rack are interconnected via multiple PCIe cables coupled to PCIe ports on the pooled system drawers. The PCIe ports are associated with host ports connections between server nodes and host ports in respective pooled system drawers are automatically detected, with corresponding PCIe connection information being automatically generated and aggregated to determine the PCIe cable topology for the rack. In one aspect, PCIe devices are emulated for each host port in a pooled storage drawer including pooled PCIe storage devices. Server nodes in a pooled compute drawer send PCIe configuration messages over the PCIe cables, with returned reply messages generated by the emulated PCIe devices identifying the host ports. Information pertaining to the host ports, pooled system drawers, and server nodes is used to determine the PCIe cable topology.

Abstract: Mechanisms to enable management controllers to learn the control plane hierarchy in data center environments. The data center is configured in a physical hierarchy including multiple pods, racks, trays, and sleds and associated switches. Management controllers at various levels in a control plane hierarchy and associated with switches in the physical hierarchy are configured to add their IP addresses to DHCP (Dynamic Host Control Protocol) responses that are generated by a DCHP server in response to DCHP requests for IP address requests initiated by DHCP clients including manageability controllers, compute nodes and storage nodes in the data center. As the DCHP response traverses each of multiple switches along a forwarding path from the DCHP server to the DHCP client, an IP address of the manageability controller associated with the switch is inserted. Upon receipt at the DHCP client, the inserted IP addresses are extracted and used to automate learning of the control plane hierarchy.

Abstract: Examples may include a basic input/output system (BIOS) for a computing platform communicating with a controller for a non-volatile dual in-line memory module (NVDIMM). Communication between the BIOS and the controller may include a request for the controller to scan and identify error locations in non-volatile memory at the NVDIMM. The non-volatile memory may be capable of providing persistent memory for the NVDIMM.

Abstract: A system and method are described for integrating a memory and storage hierarchy including a non-volatile memory tier within a computer system. In one embodiment, PCMS memory devices are used as one tier in the hierarchy, sometimes referred to as “far memory.” Higher performance memory devices such as DRAM placed in front of the far memory and are used to mask some of the performance limitations of the far memory. These higher performance memory devices are referred to as “near memory.

Abstract: A method is described that includes deciding to enter a lower power state, and, shutting down a memory channel in a computer system in response where thereafter other memory channels in the computer system remain active so that computer remains operative while the memory channel is shutdown.