INTERNSHIP – Engineer in Implementation of an open-source RISC-V GP-GPU on FPGA (M/F) - 6 months

20 Minutes ago • All levels

Graphics Engineer

Job Description

This 6-month internship at Thales Research & Technology and CortAIx Labs focuses on implementing multiple instances of the open-source Vortex GP-GPU (RISC-V compatible) on a latest-generation AMD/Xilinx FPGA, potentially utilizing HBM memory. The intern will analyze the architecture's intrinsic performance, execute various applications including graphics, vision processing, and neural networks, and compare the results with an NVIDIA Jetson Nano GP-GPU. The role contributes to innovative embedded solutions and Thales's commitment to RISC-V and open-source hardware.

Must Have:

Implement multiple instances of the Vortex GP-GPU on an AMD/Xilinx FPGA.
Study intrinsic performance of the architecture (memory bandwidths, simple kernel execution).
Execute different types of applications (graphics, classic vision processing, and neural networks).
Compare obtained performance with an NVIDIA Jetson Nano GP-GPU.
Skills in FPGA and RTL language.
Skills in C++ language.
Skills in embedded systems.
Student in an Engineering School or equivalent, preparing a Master 2 in electronics or embedded computing.

Perks:

Technological expertise.
Attention to employee balance.
Inclusive and benevolent environment.
Recognized societal and environmental commitment (Thales Solidarity, CAC 40 ESG index).
Conventioned internship with gratification determined by study level.
Handi-Engaged company, recognizing all talents and diversity.

Add these skills to join the top 1% applicants for this job

github

cpp

game-texts

fpga

opencl

linux

neural-networks

open-gl

INTERNSHIP – Engineer in Implementation of an open-source RISC-V GP-GPU on FPGA (M/F) - 6 months

Location: Palaiseau, France

Let's build a future of trust together

Thales is a global leader in high technologies specializing in three sectors: Defense & Security, Aeronautics & Space, and Cyber & Digital. It develops products and solutions that contribute to a safer, more environmentally friendly, and more inclusive world. The Group invests nearly 4 billion euros per year in Research & Development, particularly in key innovation areas such as AI, cybersecurity, quantum, cloud technologies, and 6G. Thales has nearly 81,000 employees in 68 countries.

Our commitments, your benefits

Our technological expertise
Our attention to employee balance
An inclusive and benevolent environment
A recognized societal and environmental commitment (Thales Solidarity, CAC 40 ESG index, etc.)

Your daily life

Thales Research & Technology, our research center in hardware fields, and CortAIx Labs, our research center in AI and digital, have the mission to propose breakthrough innovations, maintain and increase technological leadership, and ensure its competitiveness for the group.

Thales is committed to the employment and integration of people with disabilities. As such, our establishment is recognized as a Handi-Welcoming Organization.

The High-Performance Computing Laboratory is interested in developing innovative solutions for embedded applications. It benefits from expertise in both software and hardware, making it capable of meeting the various needs of the Thales Group.

Thales invests in RISC-V based processors and open-source hardware. Thales is a member of RISC-V International, which specifies the RISC-V instruction set, and a founding member of the OpenHW Foundation, which develops and distributes open-source RISC-V processor cores and corresponding software stacks. Within the OpenHW Foundation, Thales is a major contributor to the development of the CVA6 processor core (https://github.com/openhwgroup/cva6), which is a 32/64-bit RISC-V application processor capable of running Linux and targeting both ASICs and FPGAs.

More recently, Georgia Institute of Technology proposed in 2021 an open-source GP-GPU (General-Purpose GPU) compatible with the RISC-V instruction set, named Vortex. This graphics processor, compatible with FPGA targets, is programmed in OpenCL (PoCL environment) and OpenGL. It is fully configurable to adapt to the complexity of the FPGAs used.

Thus, your objective will be to implement several instances of the Vortex GP-GPU on a latest-generation AMD/Xilinx FPGA target, using HBM memory if possible. You will then need to study the intrinsic performance of the architecture (memory bandwidths, execution of simple computation kernels), then execute different types of applications (i.e., graphics, classic vision processing, and neural networks). A comparison of the obtained performance will finally be carried out with an NVIDIA Jetson Nano GP-GPU.