How F1's level playing field rules have impacted computer chip choices

Where once those outfits with the most money could simply write a big fat cheque to spend their way to bigger and better developments, now that is no longer the case.

Primarily, the arrival of a cost cap in F1 means that teams are strictly limited in what they can spend.

For top manufacturer outfits, it is a battle to keep within the current $140 million spending limit that has been imposed.

But, on top of the financial ceiling, another key change has been the introduction of aerodynamic testing restrictions.

Teams have been limited on how much they can run windtunnels, and how much CFD work they can do, to try to ensure that all teams have similar opportunities.

Beyond that, there is also a sliding scale handicap system in place – where the more successful you are in the world championship then the more restricted you are in what you can do in the windtunnel or with your computer simulations.

The combined impact of all these rules should deliver opportunity to the smaller outfits who have been outspent in recent years, but they have also prompted the need for a completely different approach from the top teams.

Now it is not a question of throwing money at a problem to try to solve it; it is about seeking out the biggest bang for buck possible.

Lewis Hamilton, Mercedes W13

Photo by: Carl Bingham / Motorsport Images

This impact is mostly noticeable to fans in the way that teams are limiting their car upgrades this year.

Everyone is being more strategic in how and when they bring developments, rather than bolting on everything new they find.

But digging deeper into areas where teams have made some changes, in a bid to drive up performance amid the constraints of the cost cap and the aero testing restrictions, it has been fascinating to learn how it has even triggered a deep analysis among teams about their computer chips.

A report published today has shown that a change to AMD computer processors at Mercedes at the start of 2020 has helped it deliver since then an impressive 20% performance improvement on the CFD that was used to model and test aerodynamic flow on its F1 car.

That resulted in the team being able to cut its CFD workload in half, and was a much bigger gain than the standard one or two percent improvements usually seen with each chip evolution.

It came as the result of a move to AMD's second generation EPYC processors, which is felt delivered a better price/performance balance than the systems it had used for the previous three-and-a-half years.

The swap to the new processors was triggered as the result of both the looming cost cap and a change for 2021 in a way that the FIA limited what teams could do for aerodynamic testing.

Simon Williams, Head of aero development software at Mercedes, said that the inclusion of teams' CFD within the cost cap meant some big decisions had to be made - which is why it went for the second generation EPYC processors.

"We had new regulations coming in and were refreshing our systems," he said. "Performance was the key driver of the decision making.

"We looked at AMD and the competitors. The CFD solve was the critical factor when running benchmarks. We needed to get this right, because we're going to be on this hardware for three years.

"The other element was that we have the hardware on premises, and if it takes up the whole data centre then that wasn't going to be an option. EPYC was the front runner in delivering the performance and space saving we needed."

Williams says the push was partly motivated by the way that the FIA opened up an avenue for teams to lift their CFD game from 2021.

"Wind tunnel hours and computational fluid dynamics were previously both regulated to a common number, so you could bias towards wind tunnel or CFD," he explained.

"The FIA decoupled that for 2021. The next thing they did was provide an uplift on the amount of compute we could have.

"The third thing was prorating it based on where your performance is in the championship, so there's a sliding scale that biases the capacity you have to develop aerodynamics.

"We're trying to make the most of that, as well as just the raw compute. There's also a regulation of how many geometries we can run in a certain period, which usually spans eight weeks. We're trying to maximize everything we can do in that period to get the most out of our CFD."

AMD EPYC 7003 Series Processor

Photo by: AMD

With around 1800 new geometry simulations allowed per eight-week period in 2021, Mercedes needed to know that its processors were making the most of each run.

"It's about trying to maximise the work that the CFD solve can do per clock cycle," he added.

Williams says the 20% performance gain has been impressive, but equally the lack of downtime has been an important element too.

"The new system is allowing us to focus our effort on aerodynamic performance," he adds.

"Reliability has been great too. If we miss even a few hours that puts us on the back foot, so it's important that these systems are robust and reliable. This is something that EPYC has delivered.

"The time scales are so short from initial idea, to CFD, to testing in the tunnel, and then into the car. It could be measured in weeks. The fact that we've been able to deliver consistently over the year has been key."