PhD Position F/M Compilation of a DSL based on vectorial circuit to SIMD optimized code

The  Inria University of Lille centre, created in 2008, employs 360 people  including 305 scientists in 15 research teams. Recognised for its strong  involvement in the socio-economic development of the Hauts-De-France  region, the Inria University of Lille centre pursues a close  relationship with large companies and SMEs. By promoting synergies  between researchers and industrialists, Inria participates in the  transfer of skills and expertise in digital technologies and provides  access to the best European and international research for the benefit  of innovation and companies, particularly in the region.

For more  than 10 years, the Inria University of Lille centre has been located at  the heart of Lille's university and scientific ecosystem, as well as at  the heart of Frenchtech, with a technology showroom based on Avenue de  Bretagne in Lille, on the EuraTechnologies site of economic excellence  dedicated to information and communication technologies (ICT).

Within the framework of a partnership public with LCIS within the scope of ANR SxC.

Streaming data processing is a crucial approach that focuses on traversing data to extract pertinent
information. Applications ranges from network packet manipulation to analysing DNA. Modern data-
processing tools heavily depend on efficient implementations that harness hardware acceleration to
achieve high performance. This acceleration can sometimes be achieved through automatic compilation,
but frequently demands expert developers to craft optimizations by hand.
One critical facet of this optimization process involves SIMD optimization, where data is packed into
chunks and processed with minimal branching in the code, often using bit vector operations. These
optimizations are at the core of numerous well-known software applications, such as regular expression
matching in tools like ripgrep, JSON parsing in libraries like SimdJSON, and even fundamental operations
like string encoding and decoding (Unicode parsing). Developing these optimizations requires a broad
skill set and is a testament to the expertise of programmers worldwide.

Designing VIR: an intermediate representation of vectorial programs.
During this PhD, we will explore the design and implementation of VIR, an intermediate representation
of vectorial programs heavily influenced by synchronous programming, high-performance compilation
of array languages and vectorial circuits.
The end goal is to have a machine and optimization friendly formal representation of computation
relying heavily on SIMD accelerations.
Some cases studies have already been performed in the context of various experiences, coming from
early-stages internships in which the premises of complete toolchain to evaluate simd solutions has been
designed; or a complete project, vizitig, which proposes simd implementations of programs analyzing
DNA strings.

The core of our specialized language will benefit from these experiences.
In the context of this PhD, the student will:
• make an extensive bibliography of existing approaches for compiling programs into vectorial
code, focusing on intermediate representations to represent parallelism at different abstraction
levels;
• study the specificities of different vectorial targets, especially from the circuit complexity point
of view;
• make different propositions as VIR as intermediaire representation;
• contribute to the compilation stack inside the project, focusing on back-ends.

• Subsidized meals
• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking and flexible organization of working hours
• Professional equipment available (videoconferencing, loan of computer equipment, etc.)
• Social, cultural and sports events and activities
• Access to vocational training
• Social security coverage