Jump to content

Recommended Posts

Posted

Dear ED team,

It would be great to have a kind of benchmarking Flight/Mission/Track or how to call it.

Some default situations, which shows and tests all the different aspects of the performance. (Flight though canyon, clouds, over the water, very high, etc..)

The output of it should not be a simple number, but also a verbose report, including system specs, active scripts, etc.

1. This could a easy ways for you, ED, to get reliable data from the users, if there are any performance of visualization issues with new versions.

At the moment the data from users is a wild mix of numbers, tables, unhelpful statements, trackfiles, rough system specs and screenhots.

2. Also for the players it would be very nice way to compare different graphical settings.

3. The perfect extension of such a benchmark could be to be able to automatically run a set of different settings. So that after an hour or so, the player can see the different results at once.

 

I assume this was suggested somewhere before, but the current Open Beta discussion demonstrates the value of such a feature - for the players and ED.

 

Regards

Limaro

  • Thanks 1

Windows 11 Pro 64Bit - AMD Ryzen 5800X3D - AMD 6900 XT - 64 GB RAM

VPC MongoosT-50CM2 Throttle, VPC WarBRD Base, VPC MongoosT-50CM2 Grip

Posted

I think the AI makes reproducible tracks impossible. 

i9-14900KS | ASUS ROG MAXIMUS Z790 HERO | 64GB DDR5 5600MHz | iCUE H150i Liquid CPU Cooler | ASUS TUF GeForce RTX 4090 OC | Windows 11 Home | 2TB Samsung 980 PRO NVMe | Corsair RM1000x | LG 48GQ900-B 4K OLED Monitor | CH Fighterstick | Ch Pro Throttle | CH Pro Pedals | TrackIR 5

Posted
2 hours ago, SharpeXB said:

I think the AI makes reproducible tracks impossible. 

It doesn't.

It's just that there are two different track recording systems and the one most commonly used and most easily available does not actually record events as they happen. It's also not the AI that makes it different — it's the non-deterministic nature of a number of (world and unit) simulation steps.

❧ ❧ Inside you are two wolves. One cannot land; the other shoots friendlies. You are a Goon. ❧ ❧

Posted (edited)

Given the fact that the replay tracks don’t even work I’m not sure how a benchmark track would. But it would be great to have both. 
The trouble with a benchmark track in a sim like this is that it runs live AI and flight model physics which may not be running during a replay. It’s easy to check for yourself, simply run a mission live with Fraps and then do another FPS test on the track. The trouble is you can’t get a reliable replay since it’s broken so there’s no way to test this currently. In other sims I would get different results from live vs track. 

Edited by SharpeXB

i9-14900KS | ASUS ROG MAXIMUS Z790 HERO | 64GB DDR5 5600MHz | iCUE H150i Liquid CPU Cooler | ASUS TUF GeForce RTX 4090 OC | Windows 11 Home | 2TB Samsung 980 PRO NVMe | Corsair RM1000x | LG 48GQ900-B 4K OLED Monitor | CH Fighterstick | Ch Pro Throttle | CH Pro Pedals | TrackIR 5

Posted
6 hours ago, SharpeXB said:

Given the fact that the replay tracks don’t even work I’m not sure how a benchmark track would.

That is not a fact.
That is just your experience with one particular type of replays — and ones that involve far more complex matters than is needed for this, at that.

In particular, it's something that can actually be done without even involving any of the pseudo-random traits that plague SP tracks and thus something that even with the current system can be made to reproduce reliable and repeatable outcomes. The tricky bit isn't any of that, but the detailed log and summary of the performance stats during and at the end of the benchmark. Here, too, the data is readily available — just not recorded and packaged.
 

❧ ❧ Inside you are two wolves. One cannot land; the other shoots friendlies. You are a Goon. ❧ ❧

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...