Intel: Neuromorphic Heterogeneous Computing and Programming Heterogeneous Architectures with SYCL and oneAPI Libraries and Frameworks.

Intel URC intends to sponsor proposals lasting 6-12 months focused on the development of curriculum modules consisting of one to two lecture series with laboratory or exercise component.

Specifically, this solicitation calls for the development of curriculum modules targeting undergraduate education with each module consisting of one to two lecture series with a laboratory or exercise component. Example topics of interest may include the following:

• Introduction to the heterogeneous parallel programming features of SYCL (e.g., unified shared memory, kernels, accessors, buffers)
• Migration of existing heterogeneous parallel code to SYCL
• oneAPI and SYCL for computer science, computational science, electrical and computer engineering
• Understanding the architectural interfaces and attributes of a heterogeneous compute system comprising of CPUs, GPUs, FPGAs, domain-specific accelerators, and the capacity, latency, and bandwidth attributes of the private and shared cache hierarchies and the memory architecture of the overall system
• Methodologies to perform early analysis of application algorithms, and their inherent instruction-, data-, and thread-level parallelism to enable efficient partitioning of the application among heterogeneous compute engines
• Heterogeneous parallel performance optimization (e.g., understanding the relative performance and power advantages of CPU, GPU, and FPGA architectures; determining the best architecture for different sections of a given workload; performance optimization, particularly techniques to reduce data movement between host and device memories; and using Intel® Advisor and Intel® VTune to perform iterative optimizations)
• AI training and inference using oneAPI libraries and frameworks (e.g., heterogeneity, asynchronous execution, optimization, and deployment; AI workload acceleration with the Intel® Distribution of OpenVINO™ toolkit)

Funding: up to $50K

https://www.ra.trdf.co.il/files/Intel Mindshare Curriculum Program RFP_2022_heterogeneous computing.pdf

https://www.ra.trdf.co.il/files/Mindshare_Curriculum_Program _RFP_2022_heterogeneous_computing.pdf