December 29, 20241 yr I am a huge 2024 word not allowed. But performance is worse in 2024 than 2020. I saw it on my first flight. I saw it in the 3 day alpha pre release. I saw it in every flight in 2D or VR in 2024. Still, it is smooth and stable and I have few complaints. But fps is down on my system, every flight. (I also have 2020 installed so that I can compare). But performance is still very good, even in VR with Pimax Crystal Light. 5800X3D, RTX4070, 600 Watt, one or two 1440p 32" screens, 64 GB RAM, 4 TB PCle 3 NVMe, Warthog throttle, VKB NXT EVO stick, Honeycomb Alpha yoke, CH quad, 3 Logitech panels, 2 StreamDecks, Desktop Aviator Trim Panel. Crystal Light VR.
December 29, 20241 yr With 64Gb Ram, Fenix at same airports as in 2020, load times are good, FPS is similar, then suddenly tanks to single figures. Can be caused by clicking on second monitor. Sometimes restored by clicking on first monitor, sometimes not. Switching out of and back to TAA can sometimes restore performance. End result is not worth flying Fenix in 2024, especially in VR. On the other hand, flying the balloon or Joby in VR is great fun and performance equivalent to 2020. Petraeus
December 29, 20241 yr 20 minutes ago, Petraeus said: With 64Gb Ram, Fenix at same airports as in 2020, load times are good, FPS is similar, then suddenly tanks to single figures. Can be caused by clicking on second monitor. Sometimes restored by clicking on first monitor, sometimes not. Switching out of and back to TAA can sometimes restore performance. End result is not worth flying Fenix in 2024, especially in VR. On the other hand, flying the balloon or Joby in VR is great fun and performance equivalent to 2020. Are you using frame generation? That will disengage when you move to a second monitor / that's in focus [MSI MPG X870E Carbon | 9800X3D (PBO +200Mhz / -20 Offset) | Corsair 64GB DDR5 (Custom Timings) | RTX 4090 Founders Edition (Undervolted) | WD SNX 850X 4TB + 4TB | Antec Flux Pro]
December 29, 20241 yr Just now, Sethos said: Are you using frame generation? That will disengage when you move to a second monitor / that's in focus No, I'm not. Petraeus
December 29, 20241 yr On 12/28/2024 at 3:49 AM, Cpt_Piett said: Visuals are definitely better in terms of the lighting model, raytraced shadows etc. Despite this, I'm getting better performance. So I'm guessing they've optimized multithreading, making the sim less mainthread-limited. But then there's the issue of the RAM/VRAM leak? Is this still an issue? I'm not up to date on the official forums. Running some tests now - ILS approach rwy 13L KJFK over Manhattan with FSHud. Unfortunately, I can't get CapFrameX to do any 2024 captures... So all I've got is the RTSS overlay which luckily includes a frame time graph and real-time avg FPS + avg 1% lows and 0.2% lows. Another aspect: SMT / HT on or off? With AMD SMT on, all threads are being used quite well: Any idea about using HT on/off for FS2024 ? It seemed FS2020 works beter without HT. Any thoughts ? MSI Tomahawk Z790, I7-13700K, DDR5 6000mhz, MSI 4090, 3x SSD 980 PRO, Corsair 360 Liguid CPU cooler, Corsair H1200V2 power.
December 29, 20241 yr Author 1 hour ago, altenae said: Any idea about using HT on/off for FS2024 ? It seemed FS2020 works beter without HT. Any thoughts ? 2020 was slightly better without HT. I’m not sure about 2024 yet. It does seem to use all the threads though. 7950X3D | RTX 4090 | 64GB DDR5
December 29, 20241 yr I did some comparisons this morning between memory usage in 2020 vs 2024. I have 32g of ram, and a 3060ti with 8G of RAM. As a control, I loaded using the ini a321 in '24, and the ini v2 neo in 2020. I loaded into Flytampa's KBOS. I chose this instead of a default airport becasue I wanted a 3rd party airport that is compatible with 2020 and 2024 without requiring separate versions like KPSP. In both sims, GPU usage was about 86% No difference between them. RAM usage, boy howdy. FS2020: 58-60% usage. FS2024: 86-90% usage. This likely explains why things like VA ACARS are really throwing 2024 for a loop on my system. I don't have a lot of headroom.
December 29, 20241 yr Can I answer twice? In 2D at 4K, I'm getting excellent performance in 2024, better than 2020 I believe. In VR, it's considerably worse than 2020. [email protected] - ROG Strix Z790-E - 2X16Gb G.Skill Trident DDR5 6400 CL32 - MSI RTX 4090 Suprim X - WD SN850X 2 TB M.2 - XPG S70 Blade 2 TB M.2 - MSI A1000G PCIE5 1000 W 80+ Gold PSU - Liam Li 011 Dynamic Razer case - 58" Panasonic TC-58AX800U 4K - Pico 4 VR HMD - WinWing HOTAS Orion2 MAX - ProFlight Pedals - TrackIR 5 - W11 Pro (Passmark:12574, CPU:63110-Single:4785, GPU:50688)
December 29, 20241 yr I presume that in VR, the resolution is higher and so the difference between 2020 and 2024 performance is more obvious. I see lower fps in both 2D and VR with 2024. But not enough to be a problem. 2024 is still plenty good in VR for me. Pimax CL, 4070, 5800X3D. If I had some sort of Intel cpu instead of an X3D, then VR probably would not be good. 5800X3D, RTX4070, 600 Watt, one or two 1440p 32" screens, 64 GB RAM, 4 TB PCle 3 NVMe, Warthog throttle, VKB NXT EVO stick, Honeycomb Alpha yoke, CH quad, 3 Logitech panels, 2 StreamDecks, Desktop Aviator Trim Panel. Crystal Light VR.
December 30, 20241 yr Author 12 hours ago, BrammyH said: RAM usage, boy howdy. FS2020: 58-60% usage. FS2024: 86-90% usage. That’s a huge difference. In one flight my RAM use was so high that windows was becoming sluggish. But that flight was an anomaly. 7950X3D | RTX 4090 | 64GB DDR5
December 30, 20241 yr Author After upgrading to the 566.45 hotfix driver (from 566.36), I did a performance comparison, which is also somewhat relevant to performance in general (at least with my hardware). Test: Fenix A319 flying Southend - London City - Heathrow at 3000ft. TLOD 120. FPS locked to 65 using RTSS. With both drivers, I got quite a few "hard stutters" over London CBD. This is somewhat surprising, as I had enough headroom. After the two tests, I flew Heathrow to London City, again at 3000ft, with TLOD 200 and FPS unlocked. I got FPS 75-80. On to the actual numbers (using CapFrameX): 566.36 566.45 1% low avg FPS: 40.1 36.2 FTV* below 2ms: 92.4% 94.0% *frame time variance 566.36: 566.45: Edited December 30, 20241 yr by Cpt_Piett 7950X3D | RTX 4090 | 64GB DDR5
December 30, 20241 yr Author I've mentioned before that IMO performance is better in 2024 vs 2020. Now I need to rethink this. It seems that performance is better (at least in terms of FPS) in situations with large 3rd party airport, complex airliner, ATC addon and AI traffic. But outside of those situations - I'm not sure if the performance is better. Back to more testing I suppose! 7950X3D | RTX 4090 | 64GB DDR5
December 30, 20241 yr 13 minutes ago, Cpt_Piett said: I've mentioned before that IMO performance is better in 2024 vs 2020. Now I need to rethink this. It seems that performance is better (at least in terms of FPS) in situations with large 3rd party airport, complex airliner, ATC addon and AI traffic. But outside of those situations - I'm not sure if the performance is better. Back to more testing I suppose! Same here, while FPS are significantly higher in heavy scenarios, they do not increase that much outside those. While i have more than 100FPS (FG on) in rural areas in 2020, I never get above 100FPS in 2024, even in remote areas. Might be due to the increased load by the vastly improved terrain and vegetation in 2024? Who knows... Greetings, Chris AMD Ryzen 7 9800X3D, 2x32GB DDR5 6000MT/s RAM, MSI RTX 4090 Ventus 3X, Windows 11 Home, MSFS2024
December 30, 20241 yr And yet again we have an anomaly. I often get 140/150 in more secluded areas..... We are all connected..... To each other, biologically...... To the Earth, chemically...... To the rest of the Universe atomically. Devons rig Intel Core i5 13600K @ 5.1GHz / G.SKILL Trident Z5 RGB Series Ram 64GB / GIGABYTE GeForce RTX 4070 Ti GAMING OC 12G Graphics Card / Sound Blaster Z / Meta Quest 2 VR Headset / Klipsch® Promedia 2.1 Computer Speakers / ASUS ROG SWIFT PG279Q ‑ 27" IPS LED Monitor ‑ QHD / 1x Samsung SSD 850 EVO 500GB / 2x Samsung SSD 860 EVO 1TB / 1x Samsung - 970 EVO Plus 2TB NVMe / 1x Samsung 980 NVMe 1TB / 2 other regular hd's with up to 10 terabyte capacity / Windows 11 Pro 64-bit / Gigabyte Z790 Aorus Elite AX Motherboard LGA 1700 DDR5
December 30, 20241 yr 10 hours ago, Cpt_Piett said: That’s a huge difference. In one flight my RAM use was so high that windows was becoming sluggish. But that flight was an anomaly. I’ve thought about throwing more ram at the problem, but until we get more 3rd party planes into 24 I am holding off.
Create an account or sign in to comment