Hewlett Packard Enterprise Development LP

Chinmay Ghosh

Mashood Abdulla Kodavanji

Sharad Singhal

Soumitra Chatterjee

Mohan Parthasarathy

Gautham Bhat Kumbla

Clarete Riana Crasta

Syed Ismail Faizan Barmawer

Ramya Ahobala Rao

Mashood Abdulla K

Porno Shome

In some examples, a system combines modified data tracking structures in a plurality of computer nodes into a combined tracking data structure, where a modified data tracking structure includes indicators of modified data portions in a network-attached memory. The system stores the combined tracking data structure at the network-attached memory. As part of an incremental data backup operation, the system uses the combined tracking data structure to provide the modified data portions from the network-attached memory to a backup storage system.

Incremental data backup using a combined tracking data structure

Some examples relate to providing a fabric-attached memory (FAM) for applications using message passing procedure. In an example, a remotely accessible memory creation function of a message passing procedure is modified to include a reference to a region of memory in a FAM. A remotely accessible memory data structure representing a remotely accessible memory is created through the remotely accessible memory creation function. When an application calls a message passing function of the message passing procedure, a determination is made whether the remotely accessible memory data structure in the message passing function includes a reference to the region of memory in the FAM. In response to a determination that the remotely accessible memory data structure includes a reference to the region of memory in the FAM, the message passing function call is routed to a FAM message passing function corresponding to the message passing function.

Fabric-attached memory for applications using message passing procedure

A first process executing on a first HPC compute node of an HPC cluster sends a write command, including a global identifier for a memory window that is globally accessible to processes executing in the HPC cluster, to an HPC memory node of the HPC cluster instructing the HPC memory node to write first data to the memory window allocated at the HPC memory node. The first data is sent by the first process to a memory window allocated on the HPC memory node. The first data is written by the HPC memory node to the memory window that includes randomly accessed, addressable memory locations.

Shared memory in high performance computing (HPC) messaging systems

Unused memory registrations in a remote memory system are proactively identified and freed. A memory server maintains a list of resource tracking entries for its outstanding memory registrations. The memory server continually monitors its list of resource tracking entities. When the memory server finds an unused memory registration in its list, it unregisters the corresponding shared memory region. The monitoring may be performed in the background of the memory server by, e.g., a monitor thread. As a result, unused memory registrations may be quickly released.

Memory registration for remote direct memory access

In accordance with example implementations, a process includes receiving, by a connector that is associated with a compute node and is associated with a fabric-attached memory (FAM), an application programming interface (API) called to perform an operation that is associated with a hierarchical data format (HDF) object of an HDF file. The API call includes a HDF object identifier, which corresponds to the HDF object. The process includes, responsive to the request, based on the HDF object identifier, accessing, by the connector, mapping information that is stored in the FAM; and using, by the connector, the mapping information to identify a FAM descriptor corresponding to a first data item that is stored in the FAM and corresponds to the HDF object. The process includes, responsive to the request, serving, by the connector, the API call responsive to the identification of the FAM descriptor.

Serving application programming interface calls directed to hierarchical data format files stored in fabric-attached memories

Repeating patterns are identified in a matrix. Based on the identification of the repeating patterns, instructions are generated, which are executable by processing cores of a dot product engine to allocate analog multiplication crossbars of the dot product engine to perform multiplication of the matrix with a vector.

Crossbar allocation for matrix-vector multiplications

According to examples, an apparatus may include a processor and a non-transitory computer readable medium having instructions that when executed by the processor, may cause the processor to partition a matrix of elements into a plurality of sub-matrices of elements. Each sub-matrix of the plurality of sub-matrices may include elements from a set of columns of the matrix of elements that includes a nonzero element. The processor may also assign elements of the plurality of sub-matrices to a plurality of crossbar devices to maximize a number of nonzero elements of the matrix of elements assigned to the crossbar devices.

Partition matrices into sub-matrices that include nonzero elements

Systems and methods are provided for sparse matrix vector multiplication with a matrix vector multiplication unit. The method includes partitioning a sparse matrix of entries into a plurality of sub-matrices; mapping each of the sub-matrices to one of a plurality of respective matrix vector multiplication engines; partitioning an input vector into a plurality of sub-vectors; computing, via each matrix vector multiplication engine, a plurality of intermediate result vectors each resulting from a multiplication of one of the sub-matrices and one of the sub-vectors; for each set of rows of the sparse matrix, adding elementwise the intermediate result vectors to produce a plurality of result sub-vectors; and concatenating the result sub-vectors to form a result vector.

Growing community of inventors

Bengaluru, India

Chinmay Ghosh