let's take a track that is 2500 m long with 12 m wide road, that is 30 000 sq meters. At 10 x 10 cm grid resolution, that would mean 3 million "cells". Let's say you squeeze all data tightly and simulate 4 values at 8 bit resolution, gives you 12 MB, which actually would be a nice chunk of memory Now with your 1 GB statement, we could basically update the full track 1.4 times per second. And you could play trickery like update only those regions touched by cars at higher frequency etc. I don't want to undermine their tech in any way, but like @Christian Göpfert said computers are kinda fast If anything it should be highlighted that yes in 2017 one should actually expect games/sims to use that power and it's great they do
yepp.. but livetrack is just a little part of the whole thing. "LT uses about 1gb a minute of raw data being updated..." it was about that the replays only can use a simplified and compressed version of the livetrack-data that covers 30x30m in the replay, that's why the replays can't show the track exactly how it looked during the race. It's a bit complicated how it works but they would need to sacrifice 50% LT fidelity for better matching replays (that's how I understood it) due to limited memory budgets for livetrack. But "we feel it's better to spend it on as much fidelity as possible" Imagine a 10 minute replay would need 10GB more space only to show the track exactly like it was in the race. So a complicated situation, and they had to make a decision. (the right one if you ask me)
i don't really know if it's 10x10cm or 8x8cm or 1x1cm. And it's for sure more than 4 values ^^ it's all speculation. It is what it is, how it all came together to be so much data is beyond us.
wait.. 4 values a 8 bit makes 32 bit per update. 3 million cells so it would be 96,000,000 bit so 96MB of data per update for only 4 values for a track that's only 2500m long (livetrack doesn't include grass afaik, there's another thing called livegrass) this means it could update the data only every 6 seconds. that's a bit slow. Also there is the question if every data is updated at the same rate etc... things we will never know.
your graphics card has a bandwidth of 320 GB/second or 19 TB/minute in the original post you quoted I made a mistake indeed, but your 96 MB is bits, not bytes. so you can divide by 8 again. You can think of a HD screen as 2 million 4 byte values. it's certainly awesome what they pulled off technically in total and they move a lot of bytes every frame across all systems. As I never worked on the network side of things, I wonder how they can keep the simulations stable across multiplayer etc. that is magic to me
yeah true.. Mbits and MBytes.. sometimes confuse this. So 12 MByte and it could run with 1Hz update rate...
ich erwarte nicht viel von sms aber nur eine simulation. es ist halt die frage ob sie es können. I do not expect much of sms but only one simulation. It is just the question whether they can it.
was ist besser als pcars 1?? ,schnee, regen, fahrphysik. What is better than pcars 1 ?? , Snow, rain, driving physics.
seriously? I'll let you find out for yourself. Or maybe not, depending if you buy it or not. I don't really care.
My godddddd, my ears are bleeding... Is that really the sound of the Nissan GTR GT3? Sounds like the engine is full of gravel I love more the Vacuum sound
Hm, sounds strange, tight. I see the AI cars driving very "unsettled". And i hoped for more damage effect. We will see what it brings. On the clips here it doesnt look much different to PC1.
I don't know... The Nissan's sounds are pretty weak. And I see no evidence of an "rFactor 2 with better FFB" in these videos so far. I see a slightly better PCARS1.
That GTR does not sound right, I'll give you that. But I wonder: how can you see FFB in a video ? Just curious...
Ok, I changed my mind, maybe some of these guys are playing with significantly smoother FFB effects than what I'm used to, because I've just watched a few other videos on YT and the steering wheel movement looks fine in most of them. Don't tell me you can't see the FFB effects in action, the slight movements on the virtual wheel are still visible when the car goes over bumps and the FFB shakes the wheel for example.
Never been impressed by Pcars audio, it's almost like the audio guy makes them up as he goes to what he thinks they should sound like. Overall so far I've seen nothing sound or graphics wise throughout the web that shows Pcars 2 is kickass or a massive improvement, the AI still look incapable of going round corners without stopping, and still seem to move all over the place, physics or FFB well until you'll drive it your never know, I can see alot of 2hr test drives happening on Steam.
I will tell you Testers from WMD have been complaining about the virtual wheel not matching the movement of the real wheel. That might explain why you don't see certain movements. Now, I did not follow the thread about it, since I always disable the virtual wheel while my real wheel take its place in the cockpit, so I can't tell you if that has been changed or not. In the meanwhile, have a look (and listen) at this video featuring John Sabol (from Inside Sim Racing) trying out pCars 2. He seems to be rather impressed with the changes from pCars1. But then again, the usual haters will chime in to tell he is biased because he got a free Project Cars 2 T-shirt...right ?