How computers shaped 55 years of Škoda design

Exactly 55 years ago, in autumn 1969, several large boxes containing an IBM 360/30 mainframe computer arrived at the Škoda plant in Mladá Boleslav. This was Škoda’s first foray into IT and the start of a major transformation affecting every aspect of car production. How has IT changed at Škoda since then?

24. 9. 2024 Škoda World

What systems were used for car construction at Škoda when you joined in 1993?

At that time we used Silicon Graphics workstations running on Unix. The ICEM DDM system was used as a design platform. In 1993 we had our first workstations, and at that time the construction was switching to them. Later we had dozens of them.

So the workstations were used to design components, run calculations on materials, or even design entire cars?

This was the direct of travel, yes. The limitations and possibilities were determined by how quickly the construction department could be equipped with the workstations, the performance of the machines and the capabilities of the ICEM system, i.e. to what extent it was possible to make a 3D model and then generate the necessary diagrams from it.

In the 1990s the Škoda Octavia was largely designed and modelled on computers.

Is it possible to say which model was the first whose construction or design was predominantly created on a computer?

At the time I joined, the Felicia was being developed from the Favorit, which had already been developed with the help of computers. After the Felicia, the Octavia started to be built in the 1990s, and that car was already largely designed and modelled on computers.

The IBM 360 computer and one of its first operators in 1970

What other areas, apart from car design and construction, were powerful computers and workstations used for?

Mainly for materials calculations and simulations of impact and deformation or vibration and noise, for example. In those days, the possibilities were trivial compared to today, but the software already existed. It was mainly used for subsequent calculations to compare how well the test results matched the assumptions.

In the beginning was the IBM 360

Launched in April 1964, the IBM 360/30 (or more precisely, IBM System/360 Model 30) that arrived at the Škoda factory in the autumn of 1969 was a basic model. It was discontinued in the summer of 1970, which is perhaps why the American export commission COCOM did not oppose its sale to a country behind the Iron Curtain in 1969. Together with the Model 40, the Model 30 became the best-selling machine in the 360 series, with prices starting at $133,000 at the time of launch (equivalent to approximately CZK 31 million in today’s prices).

Did the change from local systems to centralised systems come with the opening of the new C21 data centre in 2001?

Around that time, colleagues from technical development came to us and said they had planned a new system for technical computing from IBM. It was a system with twelve or twenty processors. The next turning point was in 2004, when the design department said they wanted a new, more powerful machine. The system they chose was an Intel Itanium architecture system from SGI. This was an SMP system with less than a hundred processors and a large memory that all the processor cores had access to. One factor in this was that aerodynamic know-how had been expanding in the automotive industry since the late 1990s, so these systems were to be used for aerodynamic calculations as well.

The Škoda data centre today

And then the first cluster system was introduced at Škoda?

We acquired the first SGI cluster system in 2007. The main difference between a cluster system and an SMP system with shared memory is that in a cluster each processor can only see its own operating memory. A cluster is a large number of servers with very fast processors and a big enough memory connected by a fast network. This marked the start of an era that continues to this day, where year after year the hunger for computing capacity has become greater as new versions of software have come out, offering new capabilities and allowing computational models to become more complex.

How did the technical development of cars and the related calculations change in the cluster era?

As the capacity and computational capabilities of clusters grew, the technical development department started to get to a position where their importance in car development kept increasing, because they could already model and compute results very accurately. The key value for the airflow around a car, for example, is the air resistance coefficient C_d, which then influences fuel consumption and emissions. Today, they can calculate this coefficient so accurately, that it deviates from the subsequent wind tunnel measurement by at most one percent. Computers can also perform calculations for the space inside the car – heating, air conditioning, window defrosting and more. All this is the result of growing computing power and improving software, as well as methodology, which means procedures that ensure that everything is easily repeatable.

More about the history of IT at Škoda Auto here.

Punched labels appeared in music boxes and mechanical amusements in the second half of the 19th century.

The size of the cards was derived from the size of the dollar bill so that boxes, safes, and other banking equipment common at the time could be used to store punched cards.

The original cards had just 24 columns, which may have been enough for the 1890 US census, but wasn’t enough for other data records. IBM eventually achieved success with an 80-column card. So if you come across some websites or applications today that have an 80-character-wide comment box, that’s a legacy of the punched card.

In 1970 alone, the computer station at the Škoda factory consumed nine million punched cards for planning, management, record keeping, payroll and inventory management. At a weight of 2.5 g per card, this amounted to 22.5 tonnes of cards.