To address a number of queries & comments below. The faster RAM speed (32GB) may well have had a nominal bearing on the final results, but this just reinforces my conclusion. The 32GB showed a marginally higher FPS ( not in the VR test however). So if it did have a positive influence on the outcome, it just highlights that 64GB will not enhance FPS. Remember, and as stated in the video the results are indicative only, so you can make an informed decision if to upgrade ahead of MSFS2024 being released.
I'm using the Reverb G2 and I did notice my memory usage peaked when I was at highly detailed airports (such as FlyTampa's Schiphol), especially when using high fidelity planes, causing lots of stutters. Upgrading from DDR5 32GB to 64GB fixed it for me. It didn't need much more, but those few extra gigabytes did the trick. I peak at around 35GB now.
Cheers Mark, all good information. Curious to see the tests repeated when launched to see if extra ground detail has an impact. Appreciate if you could do a rolling cache comparison too, as I like many, have poor internet speeds. We may need to stick with 2020, which is no bad thing.
I'm happy you did this test. I upgraded from 32GB to 64GB this past Friday. While I'm not noting an increase in overall fps performance, I can say that the overall smoothness is better. I tested mine with exactly the same weather conditions, time of day, and with the same flight path over the LA area into Burbank. I used the DC Designs F-15E, as it typically has a higher than normal hit on performance on my system. We'll see how it goes with MSFS2024. Intel Core i9 10900KF, GeForce RTX 4070 12GB, and 2x 32GB sticks (64GB total) DDR4 3200 DRAM, LG Flex 42 4K OLED at 120Hz, with my fps capped in sim at 60fps.
Smoothness is key here, same for me, I recently swapped out 32gb for 96gb (2x48) and it's smooth as anything, no micro stutters, even when switching tasks into windows and back.
Yor tests with 32GB vs 64GB found higher results on the 32GB which might be due to the difference in RAM speeds. You state that the Teamgroup 32GB memory kit was running at 6800MHZ and the Corsair 64GB memory kit was running at 6400MHZ. Also, the timings on the Teamgroup may have been tighter, but without knowing what your test configuration was for the RAM, it's unclear with your results. It's a known fact that Flightsim loves faster ram.
Always good to have a bit of headroom, Ram isn't that expensive. better to have to much than to little. allows other programs to be run at the same time . if you have a butt kicker or a sim rig they need their programs to run them. Thanks Mark interesting stuff. where you running your motion rig when testing ?
I haven't done a direct comparison after upgrading from 48GB to 96GB as I didn't have a replay file to use and the memory is water-cooled so none trivial to switch back. Both kits were 2 sticks with the 48GB being single sided and running at its rated 7200CL36 speed whilst the 96GB kit is double sided and running at 6800CL32 from its rated 6600CL32. I haven't noticed any difference in performance. 14900KS at 5.9GHz all core with 4090. The extra memory helps if you run productivity as well as Sim.
I am curious what is going to be coming this next year for people like me that can't do VR ? I wonder if tobi eye tracker will come out with something new in 2025 or someone else will ? I've been holding out to see what comes out and the price. 32mg ram is working good for me in MSFS.
Sorry but 5% is not margin of error it is only that, if the equipment is the same. You have not covered the impact of Ram timing, using single channel 32GB on 1: 1 which will provide much better performance with less stutter compared to 64GBs on Dual channel and possibly at 2:1. M$ is right 32GBs of well tuned DDR5 will giver significantly better performance overall if you know what you are doing. Dual channel and 2:1 increase latency (and the risk of stutters) significantly in complex fast moving scenes. C172 is a sleep walk for CPU & graphics try a roll over London landscape at low alt in an F18 and watch the numbers tank.
Whilst I acknowledge ram timing may have had a bearing on the results, it would be nominal at best. Note the VR test recorded like for like in terms of fps. The test aircraft was a Beechcraft Bonanza G36, 2 X G1000 units on board with 4K textures. Appreciate your feedback and I will be doing a A320 test over London (on final for EGLL) for comparison purposes once 2024 is out.
The only thing I can tell is on MSFS 2020, with 32GB RAM, I sometime met CTD while flying. After upgrade to 64GB, almost no CTD. Spec: i9-10900F + RTX4090
Did you use two 32GB sticks, or did you add two 16GB sticks? If you are using Intel (and I think it is true for AMD as well), you would find that two 32GB sticks is faster than four 16GB sticks because the IMC has to switch back and forth between the two channels.
Definitely not much in it 1>2FPS, but it's not a fair comparison as the Ram used is completely different spec and from different manufacturer. The different system speed won't make much difference but will have an effect, different timings on the memory will have a lot more dramatic effect than the overall memory speed and would be nice to know what memory timing both sets of ram were running at in this test, if the 32Gb system has tighter memory timings it will beat the 64Gb in performance plus in this test it also does have the higher running speed. The GPU temp is suspect as well the GPU should get up to "normal running temps" within 10>15 mins after this it should run the same temps all day long unless you are changing the stress level of the GPU but if these are the same as in this test there should be no rise of the seen 6 Deg in system temps. The fact that both 2D and VR showed this distinct change in temps for the 64Gb set up tells me something else has changed to increase the temp of the GPU not just the length of time you had the PC running! I am wondering if something got moved with cables ect during the upgrade that changed the airflow in the case! While both systems seemed to run fine I have seen that many people report that in MSFS and in DCS World that many report no increase in FPS but do seem to get better 1% lows and much less stuttering especially in VR in their systems with 64Gb and higher. Depending on how bad that stuttering is the extra memory might still be worth it to get a smoother flight experience even if you don't get more FPS. Thanks for running this test though and keep up the great work! Fly Safe
Comparing stuff like this should ideally be done with the same number of sticks (I assume two in both cases which is good) but also with the same memory speed and timings, which doesn't seem to be the case here. But anyway, we can conclude 32 GB is enough for the sim, but performance between 32 and 64 gigs should not be compared in this case.
Great information, as always. For people who have a computer that otherwise meets the ideal specs, the cost of an additional 32GB is roughly only about 4% of the cost of their entire computer, so I think most people will do it. That's what I did.
I'm pretty sure Asobo themselves came out and stated that 64GB was only necessary for those who wish to run multiple applications in the background while Flight Simulator is running. If one runs Flight Simulator by itself, 64GB isn't necessary.
Currently 32 GB is more than enough! I also flew the 2024 alpha version and it needed more memory, but 32 GB was still enough. I'm thinking (I also have 64 GB RAM) to put the rolling cache on a 32 GB RAM disk when the 2024 is available! But I'll have to wait a few more days for that. Thanks for your video.👍
Thank you very much, but :-) these 32GB and 64GB screens should show the same elements. As we can see, they are often different and it is difficult to compare, I see that in VR presenttation it is done in good way.
Next time add frametime graph to overlay. Compusemble's comparison showed almost elimination of stutter with 64 GB ram over 32. Avarage FPS won't show this diffrence.
To address a number of queries & comments below. The faster RAM speed (32GB) may well have had a nominal bearing on the final results, but this just reinforces my conclusion. The 32GB showed a marginally higher FPS ( not in the VR test however). So if it did have a positive influence on the outcome, it just highlights that 64GB will not enhance FPS. Remember, and as stated in the video the results are indicative only, so you can make an informed decision if to upgrade ahead of MSFS2024 being released.
Yes ... Informative indeed Mark ..thanks .
Hope all's ok and you enjoy your break .
I'm using the Reverb G2 and I did notice my memory usage peaked when I was at highly detailed airports (such as FlyTampa's Schiphol), especially when using high fidelity planes, causing lots of stutters. Upgrading from DDR5 32GB to 64GB fixed it for me. It didn't need much more, but those few extra gigabytes did the trick. I peak at around 35GB now.
I never cease to be amazed by the clock/bus running at ever-increasing GHz frequencies.
Thank you for the video Mark, very well done, I loved how you managed to get the side by side screen synchronized so well, great job editing! 🙂
Mark, your videos are always extremely relevant and answer the questions that we are all asking. The best. Thank you.
Thank you for saving us money with your professional subjective review!
Very interesting video. Thanks again Mark.
Cheers Mark, all good information. Curious to see the tests repeated when launched to see if extra ground detail has an impact. Appreciate if you could do a rolling cache comparison too, as I like many, have poor internet speeds. We may need to stick with 2020, which is no bad thing.
I'm happy you did this test. I upgraded from 32GB to 64GB this past Friday. While I'm not noting an increase in overall fps performance, I can say that the overall smoothness is better. I tested mine with exactly the same weather conditions, time of day, and with the same flight path over the LA area into Burbank. I used the DC Designs F-15E, as it typically has a higher than normal hit on performance on my system. We'll see how it goes with MSFS2024.
Intel Core i9 10900KF, GeForce RTX 4070 12GB, and 2x 32GB sticks (64GB total) DDR4 3200 DRAM, LG Flex 42 4K OLED at 120Hz, with my fps capped in sim at 60fps.
Smoothness is key here, same for me, I recently swapped out 32gb for 96gb (2x48) and it's smooth as anything, no micro stutters, even when switching tasks into windows and back.
Thanks very much, that helped!
Thanks Mark. You just saved me some money!
Thank you for the video
Yor tests with 32GB vs 64GB found higher results on the 32GB which might be due to the difference in RAM speeds. You state that the Teamgroup 32GB memory kit was running at 6800MHZ and the Corsair 64GB memory kit was running at 6400MHZ. Also, the timings on the Teamgroup may have been tighter, but without knowing what your test configuration was for the RAM, it's unclear with your results. It's a known fact that Flightsim loves faster ram.
Your electricity consumption is higher as well as your temperatures in 64GB Ram. 5C° difference is significant. I think 32 is the sweet spot.
Might be different with faster airplanes that have to load in way more assets over a period of couple of minutes, if not hours?
64gb @6600 ghz brings a stutters free experience in vr (PIMAX Crystal Light) 7800x3d +4090 vs 32 gb @ 6000 ghz. That's my experience.
Always good to have a bit of headroom, Ram isn't that expensive. better to have to much than to little. allows other programs to be run at the same time . if you have a butt kicker or a sim rig they need their programs to run them. Thanks Mark interesting stuff. where you running your motion rig when testing ?
@@julianhall2008 no my static rig.
I haven't done a direct comparison after upgrading from 48GB to 96GB as I didn't have a replay file to use and the memory is water-cooled so none trivial to switch back. Both kits were 2 sticks with the 48GB being single sided and running at its rated 7200CL36 speed whilst the 96GB kit is double sided and running at 6800CL32 from its rated 6600CL32. I haven't noticed any difference in performance. 14900KS at 5.9GHz all core with 4090. The extra memory helps if you run productivity as well as Sim.
I feel more for those who fitted a 4TB ssd in anticipation of a huge install
First thank you for the info sir 👍🏼
I am curious what is going to be coming this next year for people like me that can't do VR ? I wonder if tobi eye tracker will come out with something new in 2025 or someone else will ? I've been holding out to see what comes out and the price. 32mg ram is working good for me in MSFS.
Sorry but 5% is not margin of error it is only that, if the equipment is the same. You have not covered the impact of Ram timing, using single channel 32GB on 1: 1 which will provide much better performance with less stutter compared to 64GBs on Dual channel and possibly at 2:1. M$ is right 32GBs of well tuned DDR5 will giver significantly better performance overall if you know what you are doing. Dual channel and 2:1 increase latency (and the risk of stutters) significantly in complex fast moving scenes. C172 is a sleep walk for CPU & graphics try a roll over London landscape at low alt in an F18 and watch the numbers tank.
Whilst I acknowledge ram timing may have had a bearing on the results, it would be nominal at best. Note the VR test recorded like for like in terms of fps. The test aircraft was a Beechcraft Bonanza G36, 2 X G1000 units on board with 4K textures. Appreciate your feedback and I will be doing a A320 test over London (on final for EGLL) for comparison purposes once 2024 is out.
The only thing I can tell is on MSFS 2020, with 32GB RAM, I sometime met CTD while flying. After upgrade to 64GB, almost no CTD. Spec: i9-10900F + RTX4090
Did you use two 32GB sticks, or did you add two 16GB sticks? If you are using Intel (and I think it is true for AMD as well), you would find that two 32GB sticks is faster than four 16GB sticks because the IMC has to switch back and forth between the two channels.
1:43 shows that he replaced the 16GB sticks with two 32GB sticks
Definitely not much in it 1>2FPS, but it's not a fair comparison as the Ram used is completely different spec and from different manufacturer. The different system speed won't make much difference but will have an effect, different timings on the memory will have a lot more dramatic effect than the overall memory speed and would be nice to know what memory timing both sets of ram were running at in this test, if the 32Gb system has tighter memory timings it will beat the 64Gb in performance plus in this test it also does have the higher running speed.
The GPU temp is suspect as well the GPU should get up to "normal running temps" within 10>15 mins after this it should run the same temps all day long unless you are changing the stress level of the GPU but if these are the same as in this test there should be no rise of the seen 6 Deg in system temps. The fact that both 2D and VR showed this distinct change in temps for the 64Gb set up tells me something else has changed to increase the temp of the GPU not just the length of time you had the PC running! I am wondering if something got moved with cables ect during the upgrade that changed the airflow in the case!
While both systems seemed to run fine I have seen that many people report that in MSFS and in DCS World that many report no increase in FPS but do seem to get better 1% lows and much less stuttering especially in VR in their systems with 64Gb and higher. Depending on how bad that stuttering is the extra memory might still be worth it to get a smoother flight experience even if you don't get more FPS.
Thanks for running this test though and keep up the great work! Fly Safe
Comparing stuff like this should ideally be done with the same number of sticks (I assume two in both cases which is good) but also with the same memory speed and timings, which doesn't seem to be the case here. But anyway, we can conclude 32 GB is enough for the sim, but performance between 32 and 64 gigs should not be compared in this case.
Did you take off from Meigs Field? I had to buy it for MSFS 2020. It was the first place I flew from when I first bought MSFS way back in the 90tys
Meigs is a default airport (hand crafted) following the second USA World Update in MSFS
A good comparaison were to downclock at 6400mhz the 32 gb kit with the same timings of the 64 gb
You didn't show the 1% low. When you have 64GB, the 1% low is 100% greater than a setup with 32GB, considerably reducing Micros Stutters
@@fernandot9366 1% low are shown.
Great information, as always. For people who have a computer that otherwise meets the ideal specs, the cost of an additional 32GB is roughly only about 4% of the cost of their entire computer, so I think most people will do it. That's what I did.
The video is basically complete at 0:46.
In higher Resolution 64 GB makes Sense.
I'm pretty sure Asobo themselves came out and stated that 64GB was only necessary for those who wish to run multiple applications in the background while Flight Simulator is running. If one runs Flight Simulator by itself, 64GB isn't necessary.
👍👍👍
Currently 32 GB is more than enough! I also flew the 2024 alpha version and it needed more memory, but 32 GB was still enough. I'm thinking (I also have 64 GB RAM) to put the rolling cache on a 32 GB RAM disk when the 2024 is available! But I'll have to wait a few more days for that. Thanks for your video.👍
Thank you very much, but :-) these 32GB and 64GB screens should show the same elements. As we can see, they are often different and it is difficult to compare,
I see that in VR presenttation it is done in good way.
I would like to have such VR but I can't afford it at the moment
Next time add frametime graph to overlay. Compusemble's comparison showed almost elimination of stutter with 64 GB ram over 32. Avarage FPS won't show this diffrence.
For once Mark you may actually have saved me money! I’ll tell the wife, she’ll never believe it 😂
@@get_outdoors 🤣
My gaming PC is worth almost $3,000 USD.
ludicrus
32 GB is enough if you dont stream or record in 2020, VR isnt avaible in 2024 so its just guessing and we have to see by the release.
@@JDVVV378 VR is available in 2024 (PC).
@ yes True but 2024 is not released jet and VR wasn’t in the Alpha
To bad you use 2020 instead of 2024... stays an assumption now
👍👍👍