Khaled Hamidouche

ATOMIC MEMORY OPERATIONS

MULTI-HOST REMOTE MEMORY ACCESS

HARDWARE ASSISTED ATOMICS SUPPORT IN DATACENTER FABRICS

PARALLEL INTEGRATED COLLECTIVE COMMUNICATION AND MATRIX MULTIPLICATION OPERATIONS

IN-SWITCH EMBEDDING BAG POOLING

Non-Blocking Parallel Bulk Memory Operations

Low Latency Offloading of Collectives over a Switch

Multi-Tree Reduction with Execution Skew

Processing Element-Centric All-to-All Communication

DISTRIBUTED CACHING POLICY FOR LARGE-SCALE DEEP LEARNING TRAINING DATA PRE-PROCESSING

Communication of Data for a Model Between Nodes in an Electronic Device

Efficient memory-semantic networking using scoped memory models

GPU NETWORKING USING AN INTEGRATED COMMAND PROCESSOR

Efficient memory-semantic networking using scoped memory models

Network command coalescing on GPUs

Systems and methods for reducing instruction code memory footprint for multiple processes executed at a coprocessor

Optimized and scalable sparse triangular linear systems on networks of accelerators

NETWORK-RELATED PERFORMANCE FOR GPUS

Network packet templating for GPU-initiated communication

Optimized asynchronous training of neural networks using a distributed parameter server with eager updates

GPU networking using an integrated command processor