April 7, 20233 yr Moderator 2 hours ago, Ianrivaldosmith said: This isn’t a classic car, it’s a modern sim. Earlier you said the above despite me saying I am using P3D. 9 minutes ago, Ianrivaldosmith said: That’s the whole point in a forum, to debate and question. Why make a statement if you don’t want people to question it? You seem to be under the impression I'm running MSFS. I'm not. Surely the same rules about the fastest CPU you can afford apply to all sims, not just MSFS. That's been the rule for the 30 years I've been using flight sim. They still apply. Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
April 7, 20233 yr Moderator 29 minutes ago, turbomax said: wouldn't you agree that 30 fps is kind of wasteful, wouldn't 20 fps suffice and our planet could be a cooler place? You practice what you preach I assume. 🙄 19 minutes ago, turbomax said: if you really prefer fewer cores, how about Intel Core-2-Duo-E7000. You clearly can't discuss things in a sensible manner. I won't be wasting my time with any further replies to you. Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
April 7, 20233 yr 59 minutes ago, Ray Proudfoot said: The reason I'm favouring the AMD is it has fewer cores. I don't need the number the i9-13900K has. The reason I changed my 13700K for an AM5 rig is for the 100mb+ of cache. The only scenario on MSFS that I get 60 FPS is on major hubs with 100s of AI Aircraft. I expect this to change with the new CPU. That and also for some strategy games that I play... Edited April 7, 20233 yr by GCBraun PC1: AMD Ryzen 9800X3D | Zotac RTX 5090 SOLID | Asus TUF X670E-Plus | G.SKILL 64GB DDR5 PC 6000 CL30 | 4TB NVMe | Noctua NH-D15 | Asus TUF 1000W Gold | be quiet! Pure Base 500DX | Noctua NH-D15S | LG OLED CX 48" + 2x Acer Nitro XV240YP 24" + 2x 15.6" Touch-screen Panels PC2: AMD Ryzen 7500F | Asrock 7900 GRE Challenger OC | Gigabyte B650I AX | Corsair 32GB DDR5 6000 CL36 | 1TB NVMe | CM Hyper 212 | Corsair 750W Gold | Lian Li TU150 ITX | SAMSUNG Odyssey G9 49" Winctrl Ursa Minor Sidestick + Ursa Minor 32 Throttle & PAC - Thrustmaster Boeing TCA Yoke - Thrustmaster HOTAS Warthog - Honeycomb Bravo Throttle - MFG Crosswind Rudder Pedals - TrackIR - Stream Deck XL + Stream Deck Plus - Winctrl MCDU + 2 MFD's - Meta Quest 3 (VR)
April 7, 20233 yr Moderator 1 hour ago, GCBraun said: The reason I changed my 13700K for an AM5 rig is for the 100mb+ of cache. The only scenario on MSFS that I get 60 FPS is on major hubs with 100s of AI Aircraft. I expect this to change with the new CPU. Comparing the cache on the i9-13900KS to the 7950XD3 the Intel has 512Mb more L1, same L2 and 92Mb less L3. Reading this article L1 would appear to be the fastest and most important. So given that, shouldn’t the Intel be the better choice? https://www.makeuseof.com/tag/what-is-cpu-cache/ Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
April 7, 20233 yr 34 minutes ago, Ray Proudfoot said: Comparing the cache on the i9-13900KS to the 7950XD3 the Intel has 512Mb more L1, same L2 and 92Mb less L3. Reading this article L1 would appear to be the fastest and most important. So given that, shouldn’t the Intel be the better choice? https://www.makeuseof.com/tag/what-is-cpu-cache/ Where are you getting that data? The 13900KS has 80kb of L1 cache per core, 2mb of L2 per core, and 36mb of total shared L3 cache. The 7800X3D has 64kb of L1 cache per core, 1mb of L2 per core, and 96mb of shared L3 cache. So while it's true that Intel has more L1 and L2 cache per core, the L3 cache is the big deal as the latency between the various cache levels is *much* lower than between Cache and RAM - so the more you can store even in L3 cache, the less you need to swap between cache and RAM, boosting performance in certain applications like MSFS and P3D as proven by the benchmarks. Edited April 7, 20233 yr by StAgre 7800X3D - RTX 5080 - 64GB DDR5 - Dan C4-SFX
April 7, 20233 yr Moderator 19 minutes ago, StAgre said: Where are you getting that data? From this site… https://www.cpubenchmark.net/compare/5234vs5160/AMD-Ryzen-9-7950X3D-vs-Intel-i9-13900KS Thanks for clarifying the importance of L3 cache. Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
April 7, 20233 yr 9 minutes ago, Ray Proudfoot said: From this site… https://www.cpubenchmark.net/compare/5234vs5160/AMD-Ryzen-9-7950X3D-vs-Intel-i9-13900KS Thanks for clarifying the importance of L3 cache. But there they show the correct figures: 7950X3D vs 13900K Cache L1: 1,024KB, L2: 16.0MB, L3: 128MB L1: 1,536KB, L2: 16.0MB, L3: 36MB PC1: AMD Ryzen 9800X3D | Zotac RTX 5090 SOLID | Asus TUF X670E-Plus | G.SKILL 64GB DDR5 PC 6000 CL30 | 4TB NVMe | Noctua NH-D15 | Asus TUF 1000W Gold | be quiet! Pure Base 500DX | Noctua NH-D15S | LG OLED CX 48" + 2x Acer Nitro XV240YP 24" + 2x 15.6" Touch-screen Panels PC2: AMD Ryzen 7500F | Asrock 7900 GRE Challenger OC | Gigabyte B650I AX | Corsair 32GB DDR5 6000 CL36 | 1TB NVMe | CM Hyper 212 | Corsair 750W Gold | Lian Li TU150 ITX | SAMSUNG Odyssey G9 49" Winctrl Ursa Minor Sidestick + Ursa Minor 32 Throttle & PAC - Thrustmaster Boeing TCA Yoke - Thrustmaster HOTAS Warthog - Honeycomb Bravo Throttle - MFG Crosswind Rudder Pedals - TrackIR - Stream Deck XL + Stream Deck Plus - Winctrl MCDU + 2 MFD's - Meta Quest 3 (VR)
April 7, 20233 yr Moderator 18 minutes ago, GCBraun said: But there they show the correct figures: 7950X3D vs 13900K Cache L1: 1,024KB, L2: 16.0MB, L3: 128MB L1: 1,536KB, L2: 16.0MB, L3: 36MB The numbers I posted earlier are confirmed with this post. Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
April 7, 20233 yr 4 minutes ago, Ray Proudfoot said: The numbers I posted earlier are confirmed with this post. But this is also correct...the 13900KS has a bit more L1 (measured in kb, not mb), same amount of L2 and much less L3 when compared to the X3D. PC1: AMD Ryzen 9800X3D | Zotac RTX 5090 SOLID | Asus TUF X670E-Plus | G.SKILL 64GB DDR5 PC 6000 CL30 | 4TB NVMe | Noctua NH-D15 | Asus TUF 1000W Gold | be quiet! Pure Base 500DX | Noctua NH-D15S | LG OLED CX 48" + 2x Acer Nitro XV240YP 24" + 2x 15.6" Touch-screen Panels PC2: AMD Ryzen 7500F | Asrock 7900 GRE Challenger OC | Gigabyte B650I AX | Corsair 32GB DDR5 6000 CL36 | 1TB NVMe | CM Hyper 212 | Corsair 750W Gold | Lian Li TU150 ITX | SAMSUNG Odyssey G9 49" Winctrl Ursa Minor Sidestick + Ursa Minor 32 Throttle & PAC - Thrustmaster Boeing TCA Yoke - Thrustmaster HOTAS Warthog - Honeycomb Bravo Throttle - MFG Crosswind Rudder Pedals - TrackIR - Stream Deck XL + Stream Deck Plus - Winctrl MCDU + 2 MFD's - Meta Quest 3 (VR)
April 7, 20233 yr 3 minutes ago, Ray Proudfoot said: The numbers I posted earlier are confirmed with this post. Yeah that makes more sense, the L1 cache is Kilobytes, not Megabytes. Also, L1 and L2 are divided by the cores, so anything running on multiple cores needs to be copied across every core utilizing it - L3 is shared and so the larger size is more useful as there is less duplication of information 7800X3D - RTX 5080 - 64GB DDR5 - Dan C4-SFX
April 7, 20233 yr Moderator 6 minutes ago, GCBraun said: But this is also correct...the 13900KS has a bit more L1 (measured in kb, not mb), same amount of L2 and much less L3 when compared to the X3D. Yes, my mistake in Kb, not Mb. 6 minutes ago, StAgre said: Yeah that makes more sense, the L1 cache is Kilobytes, not Megabytes. Also, L1 and L2 are divided by the cores, so anything running on multiple cores needs to be copied across every core utilizing it - L3 is shared and so the larger size is more useful as there is less duplication of information Ye gods this does get complicated. 😳 But after watching the Dutch lady’s review the AMD Ryzen 7 7800XD3 looks a worthy competitor. Thoughts? https://youtu.be/bgYAVKscg0M Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
April 7, 20233 yr 13 minutes ago, Ray Proudfoot said: Yes, my mistake in Kb, not Mb. Ye gods this does get complicated. 😳 But after watching the Dutch lady’s review the AMD Ryzen 7 7800XD3 looks a worthy competitor. Thoughts? https://youtu.be/bgYAVKscg0M The confusing thing about the 7800/7900/7950X3D is that if you care about the Cache size and gaming performance only, the 7800X3D should perform the best, outperforming the 7900x3d and at least on par with the 7950x3d as the 79xx chips have two compute units, only one of which has the 3D cache. So it's up to the scheduler to properly allocate tasks, which it's not particularly good at. I think if your computer is purely dedicated to gaming and otherwise only used for lighter workloads, there's very little reason to buy the 7900 or 7950. With regards to the 13900KS, the 7800X3D should perform on par in gaming at the worst case, and vastly outperform it in the best case while using significantly less power and costing $250 less. On top of this, B650 motherboards are generally cheaper than Z790 boards as well. However, for tasks that don't utilize the cache or are highly multithreaded, of course the 13900KS will crush the 7800X3d, but these tasks are generally more on the video editing/compute side of the spectrum, and less on the gaming side. Edited April 7, 20233 yr by StAgre 7800X3D - RTX 5080 - 64GB DDR5 - Dan C4-SFX
April 7, 20233 yr Author 31 minutes ago, StAgre said: With regards to the 13900KS, the 7800X3D should perform on par in gaming not SHOULD but DOES in MSFS: Edited April 7, 20233 yr by turbomax AMD 7800X3D, Windows 11, Gigabyte X670 AORUS Elite AX Motherboard, 64GB DDR5 G.SKILL Trident Z5 NEO RGB (AMD Expo), RTX 4090, Samsung 980 PRO M.2 NVMe SSD 2 TB PCIe 4.0, Samsung 980 PRO M.2 NVMe SSD 1 TB PCIe 4.0, 4K resolution 50" TV @60Hz, VR: Pimax Crystal Light + HP Reverb G2 @ 90 Hz, Honeycomb Bravo Throttle Quadrant, be quiet 1000W PSU, Noctua NH-U12S chromax.black air cooler. 60-130 fps. no CPU overclocking. very nice.
April 7, 20233 yr While this graph is quite interesting, I believe most of these mainstream reviewers just don't know how to properly test simulators. They are surely just flying around with a default aircraft at 10k feet over NYC without significant amounts of traffic. Therefore, the difference could be even bigger when using higher end add-ons on major hubs with lots of AI... This efficiency chart is also quite amazing: Edited April 7, 20233 yr by GCBraun PC1: AMD Ryzen 9800X3D | Zotac RTX 5090 SOLID | Asus TUF X670E-Plus | G.SKILL 64GB DDR5 PC 6000 CL30 | 4TB NVMe | Noctua NH-D15 | Asus TUF 1000W Gold | be quiet! Pure Base 500DX | Noctua NH-D15S | LG OLED CX 48" + 2x Acer Nitro XV240YP 24" + 2x 15.6" Touch-screen Panels PC2: AMD Ryzen 7500F | Asrock 7900 GRE Challenger OC | Gigabyte B650I AX | Corsair 32GB DDR5 6000 CL36 | 1TB NVMe | CM Hyper 212 | Corsair 750W Gold | Lian Li TU150 ITX | SAMSUNG Odyssey G9 49" Winctrl Ursa Minor Sidestick + Ursa Minor 32 Throttle & PAC - Thrustmaster Boeing TCA Yoke - Thrustmaster HOTAS Warthog - Honeycomb Bravo Throttle - MFG Crosswind Rudder Pedals - TrackIR - Stream Deck XL + Stream Deck Plus - Winctrl MCDU + 2 MFD's - Meta Quest 3 (VR)
April 7, 20233 yr Moderator 21 minutes ago, StAgre said: I think if your computer is purely dedicated to gaming and otherwise only used for lighter workloads, there's very little reason to buy the 7900 or 7950. Thank you for confirming the same conclusion I came to watching that video. I use my flight sim PC for flying exclusively. And doing backups of course. 8 minutes ago, GCBraun said: While this graph is quite interesting, I believe most of these mainstream reviewers just don't know how to properly test simulators. They are surely just flying around with a default aircraft at 10k feet over NYC without significant amounts of traffic. Therefore, the difference could be even bigger when using higher end add-ons on major hubs with lots of AI... Even if they don’t know how best to test a flight sim you would hope they run the same test each time. It’s always a good idea to check more than one review and one important aspect she didn’t cover is overclocking. Maybe the Intel and 7950XD3 can be overclocked higher than the 7800XD3. Ray (Cheshire, England). System: P3D v5.3HF2, Intel i9-13900K, MSI 4090 GAMING X TRIO 24G, Crucial T700 4Tb M.2 SSD, Asus ROG Maximus Z790 Hero, 32Gb Corsair Vengeance DDR5 6000Mhz RAM, Win 11 Pro 64-bit, BenQ PD3200U 32” UHD monitor, Fulcrum One yoke, Fulcrum Throttle Quadrant. Cheadle Hulme Weather website.
Archived
This topic is now archived and is closed to further replies.