Intento Design strongly believes in the rise of the cognitive computing era, beyond standard artificial intelligence, leveraging intelligent machines that empower humans in their daily critical missions
Intento Design is a growing French startup that is building its growth on a disruptive innovation developed by its founder, Dr Ramy Iskander. Based on 25 years of academic research and a portfolio of four patents, Intento Design raised a total of €2.61m from public and private investors in 2015 and 2017 to develop its software solution ID-Xplore v2.0 (TRL8-TRL9) for the automated design and porting of analog blocks. Intento Design aims to extend its offer to design complete analog systems and to facilitate system integration, all using human-inspired cognitive computers.
Today, Intento Design helps designers of analog microelectronic components to substantially gain in productivity and profit margins, reduce production efforts, delays and costs, and maximise component quality and robustness.
Though an increasingly large part of our everyday experiences start in digital world, each of these experiences actually happen in an extraordinarily continuous and physical analog world. Maintaining the digital world (software, apps, data, operating systems, microprocessors, memories, and so on) and the analog world (electronic circuits and devices, micro-mechanical systems, sensors, optical devices, etc.) with the same pace of evolution becomes more critical than ever and directly influences individuals, industries, economies and societies all over the world.
Since the 1980s, the digital sector has experienced a full evolution due to the market introduction of mature software digital design tools. This transformation did not occur in the analog world, however, and analog designers are still using simulation tools for design purposes that are some 40 years old. If such a transformation occurs, the impact on everyone will be substantial.
Today, analog design is impacted by:
- The worldwide explosion of the portion of the population who use smartphones, tablets, smart surfaces, and so on;
- Mobility data-driven activities;
- The rapid evolution of all industrial sectors like aerospace, defense, medical, transport, etc.;
- The rapid evolution of linked sectors such as wireless communications; and
- The rapid emergence of novel artificial intelligence paradigms for the Internet of Things (IoT).
Furthermore, designing analog devices and components ‘is an art as much as a science’ (McKinsey 2011), as it takes 5-10 years to train competitive analog designers, something which is considered to be the main barrier to entry in the analog sector. All these facts push and oblige Analog Design Industry (ADI) actors, such as design houses, integrated device manufacturers (IDMs), analogue IP providers, foundries, system integrators, and SoC (System on Chip ) fabless, to:
- Design and migrate faster with new technologies without losing market competitiveness: Because of rapid product renewal cycles, the design of analog systems is becoming more complex due to the emergence of advanced devices such as FinFET and FDSOI used in integrated systems suitable for Internet of Everything (IOE) and the IOT. The increase in the complexity and density of such devices adds more design challenges for designers and today’s design tools, which require considerable costs and efforts to be adapted for these new devices.
- Produce error-free and correct-by-construction analog designs: While the design of digital systems needs few design iterations, due to a good level of maturity of design tools, analog systems need much more design iterations. The analog design industry (ADI) acknowledges a strong business need to eliminate or at least limit the additional costs coming originally from design errors. Using commercially available design tools, ADI actors do not have the capacities to alleviate actual design costs and consequently increase their profit margins.
Designing complex analog systems with no errors is a major challenge and a technological bottleneck for all actors in the field of semiconductor. To meet this market needs, Intento Design proposes three disruptive innovations:
This is a disruptive software which allows the systematic and automatic identification of all types of design errors arising from under-specification or over-specification and to prevent their propagation from design steps to realisation steps, hence avoiding design costs coming from deficient masks (errors costs $5m (~€4.32m) per mask for FinFET, and $1m per mask for standard processes). In addition, it guarantees the error-free and correct-by-construction functional designs of analog blocks.
ID-Xplore does not change designers’ habits or the industrial design processes that were commercially implemented several years ago. It is conceived, rather, as a standalone tool that is plugged into the most-used design and simulation environments such as Cadence Design Systems (80% market share). Moreover, ID-Xplore captures the designer’s knowledge for using it upfront in the design cycle so as to guide the designer in finding out the most appropriate design.
ID-Xplore is commercialised by Intento Design (version v2.0, released in January 2018). It covers customers’ needs related to the design phase of the analog blocks level. In accordance with Intento Design’s defined strategy, ID-Xplore is intended to completely automate system design, simulation, and integration.
Intento Design works in a strong partnership with STMicroelectronics to customise its software to design using ST FDSOI 28nm and migrate in GF FDSOI nodes.
ID-Xplore is the answer to semiconductors companies designing analog and mixed signal systems that are becoming increasingly critical. Consequently, ID-Xplore addresses a huge market opportunity to drive businesses in delivering chips on time, at a minimal cost, and to fully benefit from the latest semiconductor technologies available. Consequently, because electronic devices are now found almost everywhere, utilising ID-Xplore means that there is great potential (in the mid-term) to impact European and global industrial sectors such as aerospace, defence, medical, transport, telecommunications, consumer electronics, and many others.
This disruptive software is a direct exploitation of the foreground results of the AUTOMICS project, which has been co-ordinated by Iskander for the last five years (starting in July 2012). This patented technology allows for the modelling of the lateral propagation of minority and majority carriers inside a piece of doped substrate. This model is automatically extracted for a complete chip and simulated in a spice format in a just a few minutes. The extracted netlist is an interconnection of three types of parasitic components:
- Intento Diode;
- Intento Resistor; and
- Intento Homojunction.
These models have been designed by the Intento Design team to extensively analyse Smart Power integrated circuits (ICs), particularly in electrical vehicles (e.g. 300 sensors) for the automotive market. The software predicts complex substrate coupling effects such as latchup, high injection, temperature dependence, and so on. Once predicted by spice simulation, the designer can immune the smart power IC against these effects, then re-extract and re-simulate again to check that the phenomenon has disappeared.
The traditional approach to immune Smart Power ICs was to design and fabricate several versions and then measure these effects. This approach is very costly due to mask fabrication. With ID-Substrate, the chip fabrication is replaced by standard simulation. This results in much cheaper, reliable, durable and robust chips.
ID-Substrate will have a positive impact on the European competitiveness of the automotive market which is worth more than €750bn per year.
Intento Design has signed a corporate agreement with ams AG, a major designer and manufacturer of sensor solutions, to develop and deploy its software to the market by the first quarter of 2019.
Cognitive and Autonomous Perceptrons™:
Recently, Intento Design’s R&D team (led by Iskander) succeeded in developing the first Perceptron model that does not need any form of machine learning. Instead of training the Perceptron on raw data sets originating from a circuit simulation or other sources of data, the new type of Perceptron uses graph theory to formally model the circuit knowledge domain, replacing training with raw data.
Intento Design describes the new perceptron as ‘cognitive and autonomous’. It is cognitive as it integrates the human knowledge of the circuit as a graph, and it is autonomous since it is capable of quickly solving a problem related to the circuit knowledge described by the graph. Fig. 1 illustrates the approach adopted by the Intento R&D team to construct the new type of perceptron for a given circuit.
The Intento R&D team have evidenced the high capacity and resolution speed achieved by the new perceptron in solving complex circuit problems as perceived by a human being.
Intento design strongly believes in the rise of the cognitive computing era, beyond standard artificial intelligence, leveraging intelligent machines that empower humans in their daily critical missions.
Founder and CEO
+33 6 24 18 88 93