About this Research Topic
The golden age of computer architecture represents a time of great innovation in materials, devices, architectures, and systems in general. Exciting physical effects like spintronics, new non-volatile energy-efficient memories, widespread reconfiguration, and novel in-memory computing paradigms (e.g., in SRAM, DRAM or ReRAM) are just a few examples of promising directions to continue increasing the computing power and efficiency of systems. This is particularly important in the age of digitalization, when computing has become essential for many societal and scientific activities. Computing systems are made accessible to humans through programming models, and ultimately through programming languages. Although systems have long moved away from the classical simplified von Neumann computing model, widespread programming languages like C/C++ and Fortran are still reminiscent of that model, while incremental additions for parallelism are achieved through C++ threads or futures. Novel computing and programming abstractions are thus required alongside compilers and tools to better cater for emerging systems.
The ever-increasing complexity of emerging computing systems keeps widening the programming productivity gap; i.e., it is increasingly more challenging to efficiently program for a multi-core, a heterogeneous system-on-chip, and a heterogeneous distributed system than it is to program for a single core processor. In addition to this, there is a semantic gap between high-performance programming languages like C/C++ and Fortran, and domain experts such as biologists and geneticists, etc., who increasingly depend on computing. A similar semantic gap exists between our current tooling, for instance in compiler intermediate models, which are too fine-grained to target coarse-grained emerging accelerators, whether in the form of neuromorphic accelerators or in-memory computing engines, etc. Closing these gaps requires research into novel abstractions, domain-specific languages, intermediate languages, formal semantics, rewrite systems, semantic preserving optimization, and abstractions and cost models of emerging hardware architectures.
In the context of this Research Topic, we are interested in original contributions to the following fields:
- Programming and compiler frameworks
- Domain-specific languages and high-level compilers for emerging systems
- Intermediate languages and lowering methodologies
- Formal semantics and rewrite systems
- High-level optimization algorithms and AI/ML-aided compilers
- Models of computation for emerging computing paradigms, e.g., in-memory computing and emerging accelerators
- Abstractions and cost models of emerging hardware and computing paradigms
- Cross-layer support: interplay with runtime systems, hardware resource managers, and memory controllers
The ever-increasing complexity of emerging computing systems keeps widening the programming productivity gap; i.e., it is increasingly more challenging to efficiently program for a multi-core, a heterogeneous system-on-chip, and a heterogeneous distributed system than it is to program for a single core processor. In addition to this, there is a semantic gap between high-performance programming languages like C/C++ and Fortran, and domain experts such as biologists and geneticists, etc., who increasingly depend on computing. A similar semantic gap exists between our current tooling, for instance in compiler intermediate models, which are too fine-grained to target coarse-grained emerging accelerators, whether in the form of neuromorphic accelerators or in-memory computing engines, etc. Closing these gaps requires research into novel abstractions, domain-specific languages, intermediate languages, formal semantics, rewrite systems, semantic preserving optimization, and abstractions and cost models of emerging hardware architectures.
In the context of this Research Topic, we are interested in original contributions to the following fields:
- Programming and compiler frameworks
- Domain-specific languages and high-level compilers for emerging systems
- Intermediate languages and lowering methodologies
- Formal semantics and rewrite systems
- High-level optimization algorithms and AI/ML-aided compilers
- Models of computation for emerging computing paradigms, e.g., in-memory computing and emerging accelerators
- Abstractions and cost models of emerging hardware and computing paradigms
- Cross-layer support: interplay with runtime systems, hardware resource managers, and memory controllers
Keywords: Programming Languages, Formal Semantics, Domain-specific Languages, Compilers, Progressive Lowering, Code Optimization, Runtime Management, Emerging Computer Architectures, Reconfigurable Computing, In-memory Computing, Non-von Neumann Computing
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
