Intel Corporation

Dell Products Inc.

Sujoy Sen

Susanne M Balle

Francesc Guim Bernat

Anil Vasudevan

Linden Cornett

Evan Custodio

Rahul Khanna

Reshma Lal

Pradeep M Pappachan

Robert O Sharp

David B Minturn

Selvakumar Panneer

Parthasarathy Sarangam

Luis S Kida

Soham Jayesh Desai

Paul H Dormitzer

Slawomir Putyrski

Henry M Mitchel

Thomas E Willis

Kenneth C Creta

Durgesh Srivastava

William T Futral

Sivakumar Radhakrishnan

Bassam N Coury

Mohan J Kumar

Li Zhao

Yadong Li

Donald K Newell

Dave B Minturn

Karthik Kumar

Hemal Vinodchandra Shah

Srihari Makineni

Anshuman Thakur

Lidia Warnes

Akhilesh Kumar

Steven R King

Mrittika Ganguli

Narayan Ranganathan

Arun Raghunath

Phanindra Mannava

Kshitij Arun Doshi

Rajesh M Sankaran

Ravishankar Iyer

Murugasamy K Nachimuthu

Xiao Liang Hu

James B Crossland

Ziye Yang

Frank L Berry

Nimrod Diamant

Gang Cao

Robert G Southworth

D Michael Bell

Jose Niell

Gregory Cummings

Joe Grecco

Scott David Peterson

Anjaneya Reddy Chagam Reddy

Partha Sarangam

Gary Tsao

Ravi Iyer

Marcelo Cintra

Yi Zou

Joseph J Grecco

Tushar Gohad

Francois Dugast

Suresh S Chittor

Kiran A Padwekar

Ram Huggahalli

Ching-Tsun Chou

Prafulla Deuskar

David C Lee

Donald L Faw

Dennis Bell

Rajee S Ram

Srinand Venkatesan

Andalib Khan

Cunming Liang

Huan Zhou

Changpeng Liu

Sandeep Agarwal

Deepu Syam Sreedhar M

Neerav Parikh

Siva Balasubramanian

Ronald L Dammann

Shaopeng He

Roberto Penaranda Cebrian

Siddhartha Kumar Panda

Chong Han

Malek Musleh

Greg D Cummings

Sugesh Chandran

Kunal Sablok

Haitao Kang

Pedro Yebenes Segura

Krishna Kumar P K

Anupama Kurpad

Allister Alemania

Gene Wu

Methods and apparatus to minimize hot/cold page detection overhead on running workloads. A page meta data structure is populated with meta data associated with memory pages in one or more far memory tier. In conjunction with one or more processes accessing memory pages to perform workloads, the page meta data structure is updated to reflect accesses to the memory pages. The page meta data is used to determine which pages are 'hot' pages and which pages are “cold” pages, wherein hot pages are memory pages with relatively higher access frequencies and cold pages are memory pages with relatively lower access frequencies. Variations on the approach including filtering meta data updates on pages in memory regions of interest and applying a filter(s) to trigger meta data updates based on (a) condition(s). A callback function may also be triggered to be executed synchronously with memory page accesses.

Method to minimize hot/cold page detection overhead on running workloads

Examples described herein relate to providing an interface to an operating system (OS) to create different memory pool classes to allocate to one or more processes and allocate a memory pool class with a process of the one or more processes. In some examples, a memory pool class of the different memory pool classes defines a mixture of memory devices in at least one memory pool available for access by the one or more processes. In some examples, memory devices are associated with multiple memory pool classes to provide multiple different categories of memory resource capabilities.

Memory pool management

An apparatus to facilitate disaggregated computing for a distributed confidential computing environment is disclosed. The apparatus includes one or more processors to: provide a remote GPU middleware layer to act as a proxy for an application stack on a client platform that is separate from the remote server platform, wherein the remote GPU middleware layer comprises is to expose an abstraction of the remote GPU to userspace components of a remote GPU stack, the userspace components running on the client machine; communicate with a kernel mode driver of the one or more processors to cause the host memory to be allocated for data structures used to communicate commands between the client and the remote GPU; and invoke the kernel mode driver to submit a workload generated by the application stack, the workload submitted for processing by the remote GPU using the data structures allocated in the host memory.

Disaggregated computing for distributed confidential computing environment

Examples described herein include one or more processors; a network interface; and a direct memory access (DMA) engine communicatively coupled to the one or more processors. In some examples, the DMA engine is to receive a DMA data access request and based on an address in the DMA data access request corresponding to a remote memory device, the DMA engine is to cause the network interface to generate at least one packet for transmission to the remote memory device. In some examples, if the source address corresponds to a local memory device and the destination address corresponds to a remote memory device, the DMA engine is to cause the network interface to generate at least one packet for transmission to the remote memory device.

Direct memory access (DMA) engine with network interface capabilities

Methods and apparatus for end-to-end data plane offloading for distributed storage using protocol hardware and Protocol Independent Switch Architecture (PISA) devices. Hardware-based data plane forwarding is implemented in compute and storage switches that comprise smart server switches running software executing in a kernel and user space. The compute switch is coupled to one or more compute servers/nodes and the storage server is coupled to one or more storage servers or storage arrays. The hardware-based data plane forwarding facilitates an end-to-end data plane between the computer server(s) and storage server(s)/array(s) that is offloaded to hardware. In one example the software comprises Ceph components used to implement control plane operations in connection with hardware offloaded data plane operations, and storage traffic employs the NVMe-oF protocol and the kernels include NVMe-oF modules. In one aspect the hardware-based data plane forwarding is implemented using programmable P4switch chips. In one aspect the storage and server switches are Top of Rack (ToR) switches.

End-to-end data plane offloading for distributed storage using protocol hardware and PISA devices

Technologies for dividing work across one or more accelerator devices include a compute device. The compute device is to determine a configuration of each of multiple accelerator devices of the compute device, receive a job to be accelerated from a requester device remote from the compute device, and divide the job into multiple tasks for a parallelization of the multiple tasks among the one or more accelerator devices, as a function of a job analysis of the job and the configuration of each accelerator device. The compute engine is further to schedule the tasks to the one or more accelerator devices based on the job analysis and execute the tasks on the one or more accelerator devices for the parallelization of the multiple tasks to obtain an output of the job.

Technologies for dividing work across accelerator devices

An apparatus to facilitate disaggregated computing for a distributed confidential computing environment is disclosed. The apparatus includes a graphics processing unit (GPU) to: provide a virtual GPU monitor (VGM) to interface over a network with a middleware layer of a client platform, the VGM to interface with the middleware layer using a message passing interface; configure and expose, by the VGM, virtual functions (VFs) of the GPU to the middleware layer of the client platform; intercept, by the VGM, request messages directed to the GPU from the middleware layer, the request messages corresponding to VFs of the GPU to be utilized by the client platform; and generate, by the VGM, a response to the request messages for the middleware client.

An apparatus to facilitate disaggregated computing for a distributed confidential computing environment is disclosed. The apparatus includes one or more processors to facilitate receiving a manifest corresponding to graph nodes representing regions of memory of a remote client machine, the graph nodes corresponding to a command buffer and to associated data structures and kernels of the command buffer used to initialize a hardware accelerator and execute the kernels, and the manifest indicating a destination memory location of each of the graph nodes and dependencies of each of the graph nodes; identifying, based on the manifest, the command buffer and the associated data structures to copy to the host memory; identifying, based on the manifest, the kernels to copy to local memory of the hardware accelerator; and patching addresses in the command buffer copied to the host memory with updated addresses of corresponding locations in the host memory.

Examples described herein relate to a switch, when operational, that is configured to receive in a packet an indicator of number of remaining bytes in a flow and to selectively send a congestion message based on a fullness level of a buffer and indication of remainder of the flow. In some examples, the indicator is received in an Internet Protocol version 4 consistent Options header field or Internet Protocol version 6 consistent Flow label field. In some examples, the congestion message comprises one or more of: an Explicit Congestion Control Notification (ECN), priority-based flow control (PFC), and/or in-band telemetry (INT). In some examples, to selectively send a congestion message to a transmitter based on a fullness level of a buffer that stored the packet and the number of remaining bytes in flow, the switch is to determine whether the buffer is large enough to store the remaining bytes in the flow.

Selective congestion notification by a network interface device

Examples described herein and includes at least one processor and a direct memory access (DMA) device. In some examples, the DMA device is to: access a command from a memory region allocated to receive commands for execution by the DMA device, wherein the command is to access content from a local memory device or remote memory node. In some examples, the DMA device is to: determine if the content is stored in a local memory device or a remote memory node based on a configuration that indicates whether a source address refers to a memory address associated with the local memory device or the remote memory node and whether a destination address refers to a memory address associated with the local memory device or the remote memory node. In some examples, the DMA device is to: copy the content from a local memory device or copy the content to the local memory device using a memory interface.

Page-based remote memory access using system memory interface network device

Examples described herein includes a network interface controller comprising a memory interface and a network interface, the network interface controller configurable to provide access to local memory and remote memory to a requester, wherein the network interface controller is configured with an amount of memory of different memory access speeds for allocation to one or more requesters. In some examples, the network interface controller is to grant or deny a memory allocation request from a requester based on a configuration of an amount of memory for different memory access speeds for allocation to the requester. In some examples, the network interface controller is to grant or deny a memory access request from a requester based on a configuration of memory allocated to the requester. In some examples, the network interface controller is to regulate quality of service of memory access requests from requesters.

Shared memory

Technologies for dynamically managing resources in disaggregated accelerators include an accelerator. The accelerator includes acceleration circuitry with multiple logic portions, each capable of executing a different workload. Additionally, the accelerator includes communication circuitry to receive a workload to be executed by a logic portion of the accelerator and a dynamic resource allocation logic unit to identify a resource utilization threshold associated with one or more shared resources of the accelerator to be used by a logic portion in the execution of the workload, limit, as a function of the resource utilization threshold, the utilization of the one or more shared resources by the logic portion as the logic portion executes the workload, and subsequently adjust the resource utilization threshold as the workload is executed. Other embodiments are also described and claimed.

Technologies for dynamically managing resources in disaggregated accelerators

Examples described herein relate to a network interface that includes an initiator device to determine a storage node associated with an access command based on an association between an address in the command and a storage node. The network interface can include a redirector to update the association based on messages from one or more remote storage nodes. The association can be based on a look-up table associating a namespace identifier with prefix string and object size. In some examples, the access command is compatible with NVMe over Fabrics. The initiator device can determine a remote direct memory access (RDMA) queue-pair (QP) lookup for use to perform the access command.

Offload of storage node scale-out management to a smart network interface controller

Examples described herein includes an apparatus comprising: a network interface configured to: receive a request to copy data from a local memory to a remote memory; based on a configuration that the network interface is to manage a cache store the data into the cache and record that the data is stored in the cache. In some examples, store the data in the cache comprises store most recently evicted data from the local memory into the cache. In some examples, the network interface is to store data evicted from the local memory that is not stored into the cache into one or more remote memories.

Network interface controller with eviction cache

Growing community of inventors

Portland, OR, United States of America

Sujoy Sen