August 19, 20205 yr https://www.computerbase.de/2020-08/flight-simulator-benchmark-test/?amp=1 Well, there you go, it's not "our system". Basically, it's FSX tech - CPU bottlenecked, bad GPU usage, wildly varying loads depending on plane. Asobos descision to go with DX11, what a fail. Up to 74% FPS loss by plane - good night PMDG. - Currently giving X-Plane 12.10 a spin on Shadow PC. 10 years with X-Plane now, since 10.20
August 19, 20205 yr Not surprised knowing that they use freaking HTML to render 2D gauges ... think of the CPU overhead to interpret the HTML into C++/directX.. what a waste
August 19, 20205 yr Very interesting article. Thank you Colonel X. Of course they have a lot of work on their hands to make this The flight simulator. Cheers Pierre I9 14900K 5.5 64gb ram 6800 RTX5090 Asus Strix Gaming E
August 19, 20205 yr I intend to use MSFS with the Reverb G2. VR support is promised this year, but now I definitely will need to wait for DX12 too. P3D45, 8700K, RTX3080Ti, 32 GB, HDD 3 + 6 TB, SSD 0.5 TB Warthog HOTAS, Honeycomb Bravo, MFG pedals, Reverb G2
August 19, 20205 yr Except if you actually look at the benchmark data I'd suggest that the frames are pretty good with capable hardware. Also great feedback on these forums from many many people regarding the performance. But that wouldn't suit your narrative of absolute misery so no doubt that'll be ignored.
August 19, 20205 yr Whilst some of it is interesting, there are a lot of incorrect assumptions on the part of the writers of that article, which demonstrate some very shaky analytical logic to say the least. For example: 'It turns out that the more complex the aircraft, the lower the GPU utilization. That is hardly surprising. At the same time, however, the CPU usage drops. This confirms the assumption that either the API in the Flight Simulator is limiting or the engine itself causes problems.' Nope, this does not confirm that at all, and to suggest it does is not even remotely scientific. To draw a firm conclusion, you would need to know all parameters calculations-wise, for each of the four models chosen for the assessment (Cessna 152, TBM 930, Boeing 747-8, Airbus A320 Neo). But they don't know these, nor do they make any attempt to find out either, they simply assume 'bigger plane must mean more calculations'. So in this regard, the only thing it confirms, is their confirmation bias making them draw a conclusion they are looking to draw, as a result of not knowing what what frig they are on about. Edited August 19, 20205 yr by Chock Alan Bradbury Check out my youtube flight sim videos: Here
August 19, 20205 yr in every alpha/beta release I tested with my 9900k I never, not even once managed to peg any CPU core to 99% load. GPU? yes but that was with ultra settings and 4K, I would achieve the same result with any modern game. Compared to P3D which is a Core0 dramaqueen, MSFS is more easy on the CPU from my opinion. Many things need to be corrected with MSFS, but claiming it to be CPU hog à la P3D is not accurate. Edited August 19, 20205 yr by SAS443 EASA PPL SEPL + NQ / CB-IR in progress MSFS24 | X-Plane 12
August 19, 20205 yr 13 minutes ago, cepact said: Not surprised knowing that they use freaking HTML to render 2D gauges ... think of the CPU overhead to interpret the HTML into C++/directX.. what a waste Actually, the tech they've used to directly blit html/js/css to graphics surfaces is incredibly clever, and incredibly performant. You would be extremely hard pressed to match the possible polish and speed by hand in D3D (especially for a panel as complex as the Garmin), and definitely way, way, way faster to develop. If you look in dev mode, the majority of the time the GPU is the bottleneck. Occasionally the CPU will be in certain situations.
August 19, 20205 yr It really is amazing to watch all this on the 2nd day of release. If any Flight Sim in the history of Flight Sim looked or felt half as good as MSFS does right now don't think anyone would be here right now, we would all be on that Sim. Can't wait to see what this coming year brings for MSFS. Edited August 19, 20205 yr by SirDan i7-8600k @ 3.70 GHz 16.0 GB Ram OS Win 10-64 bit Geforce GTX 1070
August 19, 20205 yr I definitely see more positive notes about performance than bad. I ignored Colonel X, cant read that bashing anymore. 9900k@5GHz, 32GB@3200 RAM, Strix RTX 2080ti, VKB Gladiator mkii, VKB rudder pedals
August 19, 20205 yr Author Bashing lol. People really have an issue with cold, hard facts nowadays. - Currently giving X-Plane 12.10 a spin on Shadow PC. 10 years with X-Plane now, since 10.20
August 19, 20205 yr Interesting article. Seems to confirm what I've been seeing and reading in that you can get decent detail and smoothness without breaking the bank with hardware. Useful for someone like myself who has to basically buy a new pc to be able to run it as its the first time I've seen a direct Intel / AMD comparison so I can start to get an idea of which way to go. Thanks for posting it for us.
August 19, 20205 yr 5 minutes ago, Colonel X said: cold, hard facts Facts? Really? That 'analysis' conclusion I drew attention to above, was the scientific equivalent of standing by the roadside for 30 seconds, seeing a yellow and red car go past then concluding 'this comprehensively confirms cars can only be either red or yellow.' It's not only incorrect, it's laughable. Edited August 19, 20205 yr by Chock Alan Bradbury Check out my youtube flight sim videos: Here
August 19, 20205 yr 4 minutes ago, Colonel X said: Bashing lol. People really have an issue with cold, hard facts nowadays. From your posts is appears that you're not missing any opportunity to disparage the new sim. What do you gain by doing so and why do you do it?
Archived
This topic is now archived and is closed to further replies.