Cutting Edge
Innovative approaches to advance in progress
Optimizing Technologies
As the name says our technological outlook is dedicated mostly to an optimization of the microchip creation process. But it goes far beyond development of fast hardware and effective software. It’s all about methods and approaches leading to a suitable outcome.
System Level Design
This step of SoC development essentially is dedicated to searching for a proper architecture. After the architecture is selected, it always runs on specific emulators together with special system software to manage the future microchip.
And as such emulation provides new data for architecture to be tuned so the system level design becomes an iterative process. Each emulation leads to architecture modifications and it also causes emulators and system software to be modified respectively.
Any tiny reduction of cost or time makes the whole process significantly cheaper and faster.
Architecture selection due to defined benchmarks and projected application.
SoC modeling and embedded software compilation based on selected architecture and determined source code respectively.
Software run simulation on abstract SoC model.
Design algorithm
Our advance
Optimizing technologies has a set of unique tools and technologies that accelerates system level design. High-level hardware description languages (PPDL), optimizing compilers (UTL, SMCC) and algorithms of the profiling-based SoC’s parameters modification (number of cores, cache sizes, bus width, etc.) dramatically decrease cost and time of the process.
These technologies allow us to develop highly optimized SoCs for a wide range of applications with the best time2market value.
Emulation of heterogeneous SoC's based on QEMU appears to be faster due to the use of the binary translation.
Trace subsystem allows a wide range of application execution analyses on specific SoC. And traces make it possible to debug and profile applications, to undertake a dynamic software analysis and define the bottle-necks of the hardware.
SoC modeling via QEMU-TLM
Advantages
Quickness
Quicker modeling due to using of QEMU.
Effort
Suitable for modeling of multicore heterogeneous architectures with accelerators and peripherals.
Coherence
Unified modulated time system.
Other technologies applied
Mathematical algorithms
• complex systems modelling
• predictive data analytics
• advisory and expert decision support systems, intelligent transport systems
AI and machine learning
• robotics and assistive devices
• human-machine interaction interfaces
• computer vision systems
• 3D laser scanning and information modeling
• decision-making and support systems