Probably some sort of rolling road able to reproduce various track scenarios in a static, controlled environment.
Probably some sort of rolling road able to reproduce various track scenarios in a static, controlled environment.
It is interesting. How they can simulate all tracks and especially turns ?"And at Red Bull in the factory, the VTC super test stand is available day and night. The entire car and motor can then drive virtually on any race track in the world. The test stand is directly connected to the simulator."
https://redbullracing.redbull.com/artic ... orner-2019The final word 🗣
Christian looks back at the season and forward to next year .
Honeymoon with Honda
It’s no secret that in recent seasons the relationship between Renault and Red Bull deteriorated, so it’s vindicating that Honda have made good on their promises. “This is the first year for a long time where what was promised has been delivered. It’s a different type of partnership, it’s a true partnership and you can see what it means to Honda when they get a result: the emotion, the pride, the satisfaction throughout the whole business” added Christian.
2020 vision
"We need to come out of the blocks competitively next year, Mercedes are still the benchmark going into next year so we need to try and push them from the first race in order to take the challenge to them. There is a real buzz and anticipation in the factory at the moment, so we need to capitalise on that and build on the positive momentum from 2019.”
I think engine must be installed to the mule car, gearbox and wheels which allow to simulate loads on pu, to be able to said that "we can test it with 330 kph. Otherwise it makes nonesense.HPD wrote: ↑20 Dec 2019, 16:07The VTT is the Virtual Test Track, right?
I always thought that the VTT worked with a Simulator + Test Bench.
We know that RB and Honda are a couple of minutes away in Milton Keynes.
Won't they use the VTT at the RB factory connected to the Honda test bench? It seems to me something quite normal and easy to do.
https://i.ibb.co/tHNFvVp/simulador.png
True.etusch wrote: ↑20 Dec 2019, 19:22I think engine must be installed to the mule car, gearbox and wheels which allow to simulate loads on pu, to be able to said that "we can test it with 330 kph. Otherwise it makes nonesense.HPD wrote: ↑20 Dec 2019, 16:07The VTT is the Virtual Test Track, right?
I always thought that the VTT worked with a Simulator + Test Bench.
We know that RB and Honda are a couple of minutes away in Milton Keynes.
Won't they use the VTT at the RB factory connected to the Honda test bench? It seems to me something quite normal and easy to do.
https://i.ibb.co/tHNFvVp/simulador.png
Only AMD has perfected 7nm. Intel is stuck at "barely" 10nm. Only Qualcomm has 7nm outside AMD, and it's not fully scalable yet. Actually, it's why those 2 anchor my stock portfolio..godlameroso wrote: ↑17 Dec 2019, 22:13It's really interesting how interlinked development is. The VTT is fed data from the actual racing car on track, which then feeds the simulator model, which leads to chassis developments, which again feed the VTT. The simulator model is used in the 'driver in the loop' simulator, which trains the racing driver, which again feeds into and helps develop the simulation model, which is tied up in the previously mentioned loop.
That cycle is limited by the teams processing powers. It's going to be really interesting given how computation is now taking another big step forward in the near future. The large chip makers(only 2 consumer grade silicon chip manufacturers on earth) are now perfecting their 7nm process nodes for consumer grade hardware. This is half the size of current conventional process nodes (14nm). Transistor density and computational power is going to keep increasing, to the point the current teraflop limits will seem downright draconian based on the parallel processing revolution that's about to take place.
AMD doesn't have a fab. TSMC is developing the manufacturing, i.e. the 7nm process. And they are producing the chips for AMD, Apple, Nvidia, ... . Qualcomm chips are produced at Samsung foundries (in 7nm) nowadays. And Intel doesn't really have anything high performance on 10nm yet. And their 10nm on mobile has a pretty big frequency problem anyway.Zynerji wrote: ↑20 Dec 2019, 22:54Only AMD has perfected 7nm. Intel is stuck at "barely" 10nm. Only Qualcomm has 7nm outside AMD, and it's not fully scalable yet. Actually, it's why those 2 anchor my stock portfolio..godlameroso wrote: ↑17 Dec 2019, 22:13It's really interesting how interlinked development is. The VTT is fed data from the actual racing car on track, which then feeds the simulator model, which leads to chassis developments, which again feed the VTT. The simulator model is used in the 'driver in the loop' simulator, which trains the racing driver, which again feeds into and helps develop the simulation model, which is tied up in the previously mentioned loop.
That cycle is limited by the teams processing powers. It's going to be really interesting given how computation is now taking another big step forward in the near future. The large chip makers(only 2 consumer grade silicon chip manufacturers on earth) are now perfecting their 7nm process nodes for consumer grade hardware. This is half the size of current conventional process nodes (14nm). Transistor density and computational power is going to keep increasing, to the point the current teraflop limits will seem downright draconian based on the parallel processing revolution that's about to take place.
When you can stick 4 of these in a single rack server, you can quickly reach crazy horsepower for simulation.
https://www.newegg.com/p/N82E1681911358 ... t0QAvD_BwE
You forgot to mention Nvidia has announced a 7nm and they already are leaders in AI distributed computation. Would be great to see Nvidia involvement.Zynerji wrote: ↑20 Dec 2019, 22:54Only AMD has perfected 7nm. Intel is stuck at "barely" 10nm. Only Qualcomm has 7nm outside AMD, and it's not fully scalable yet. Actually, it's why those 2 anchor my stock portfolio..godlameroso wrote: ↑17 Dec 2019, 22:13It's really interesting how interlinked development is. The VTT is fed data from the actual racing car on track, which then feeds the simulator model, which leads to chassis developments, which again feed the VTT. The simulator model is used in the 'driver in the loop' simulator, which trains the racing driver, which again feeds into and helps develop the simulation model, which is tied up in the previously mentioned loop.
That cycle is limited by the teams processing powers. It's going to be really interesting given how computation is now taking another big step forward in the near future. The large chip makers(only 2 consumer grade silicon chip manufacturers on earth) are now perfecting their 7nm process nodes for consumer grade hardware. This is half the size of current conventional process nodes (14nm). Transistor density and computational power is going to keep increasing, to the point the current teraflop limits will seem downright draconian based on the parallel processing revolution that's about to take place.
When you can stick 4 of these in a single rack server, you can quickly reach crazy horsepower for simulation.
https://www.newegg.com/p/N82E1681911358 ... t0QAvD_BwE