Stutter, which has been occurring frequently since SU13, has been drastically reduced.
As usual, after takeoff, I first orbited 60,000 ft above Tokyo to stabilize the drawing, but even in the MOD area where the influence of the communication environment is supposed to be weak, the drawing spread quickly and without any hitches. I was able to see and experience the effects. (This is strange. I don’t think that changing from 7G/s to 28G/s will reduce the impact of optical lines.In addition, it appears that it is not only the official photo scenery in the innermost city center that is rapidly expanding, but also the MOD drawing area of local device storage…)
I once again feel a mystery as to what the rolling cache is doing.
Contains MODs for the Shinjuku and Marunouchi areas based on PLATEAU (Public 3D Building Reproduction Project) data. Each of these has a capacity of several GB, and regardless of communication, the data is already accommodated in M.2.
Nevertheless, there are improvements.These exist separately at a distance of several kilometers, and up until now, when going back and forth, you would have suffered from severe FPS drops and sound noise, but at the general drawing limit (200 knots, there were no problems, and at 400 knots, the FPS drop was only slow for a few seconds.
I didn’t have any problems, so I took a 500km trip from Tokyo to Kyoto all at once. 100ft, 30 minutes.
There are no bald mountains in the forest drawing.
On the way, there was a big city called Nagoya where I installed a heavy MOD, but the drop was within the ideal range.
I think it’s an amazing effect.
I would like to be able to quickly draw a satellite city district in western Tokyo, so I would also like to be able to deploy a medium-sized manual cache to a RAM disk.
There are many student towns and cities where creators live, and the JR Chuo Line is a straight track that is over 10 kilometers long, and it would be difficult for us to comfortably draw at 900 knots. I would like to be able to draw at high speed, which is a pleasant feeling.
Manual cache seems to be able to create low and medium precision without any problem.
On the other hand, if you try to create a manual cache that includes high-precision cells, MSFS will CTD in a few seconds even for just one cell.
The cause may be that you are using enhancement using GOOGLEMAP.
However, I hesitate to remove this, and with the current improvement seen, it may not be necessary.
Hey there!!! Well. after some deliberation, I decided to try out 2.02 bios. I found I had incorrectly flashed on BIOS version 2.00. After that or bios version 2.02 you are supposed to power it down after flashing for 10 seconds before entering any settings or entering windows. Guess what. My OC’s are effortless in this bios. My FCLK is up at 2100 and I’m going to push on to try and hit 2200, maybe more. I’m using less voltages. Been slowly dropping my vsoc too. I did take a slight performance hit from upgrading to WIN11 23H3 however. That fix that’s all over the net, is snake oil. I tried it. Once core isolation was enabled I gained some latency on the memory, and L3 Cache in AIDA64
If you have the 7950x3d, make sure that you manually tell Xbox Game Bar that FS2020 is a “game”. It doesn’t automatically identify it as such and won’t push it to CCD0. The chip is such a beast that I forgot to do this for a few days and the sim still ran fine on CCD1.
and how does one do such a thing, I’ve quickly developed a hate for GameBar as all I seem to be able to do is open it and as soon as I touch anything else it closes.
Windows key + G, Settings, “Remember this is a game”.
I’m assuming I’ve completed this as the tag isnt available.
but then I click on the game and the toolbar goes away?
This all seems not to be functional?
Nope - that’s a feature! Don’t touch the screen except wirh the area of the game bar app.
While running FS2020, open the game bar, click on the gear icon, under the “general” tab, make sure that “remember this is a game” is checked.
Like I said, it isn’t by default and your 7950x3d will not run it on the vcache CCD.
@Salem978 Its because you have 7800X3D. That’s why you have no entry. This thread is in regards to the 7950X3D.
I also just learned that if FS loses focus (you are working on a second screen) then FS will run on CCD1 (without the vcache). As soon as you click back to FS it moves back over to CCD0. This is seamless and there is no studder at all during the transition, but framerate changes dramatically.
@TitanMiller2081 Yes but what also will change FPS dramatically is your driver settings. If you have an NVIDIA card make sure you have the setting “Background Max Frame Rate” to OFF.
No idea, but wondering what you are trying to test or benchmark! Y-Cruncher tests memory, not the performance of your CPU.
other than make sure the memory controller IMC is working okay and you SIMMS are fine.
If your trying to bench the CPU use something like Cinebench.
Basically the memory and memory controller together. A few months back @MrTonySM and I were talking about y-crunching so I figured I’d share.
This… is amazing actually. Very nice to see, great work!!!
HAHA!!! It’s been a while and time to update on this thread. Things are changing. New AGESA releases, and different ways of doing things. I can tell ya, what we’re told and what can happen is different. I run now with my power plan at Optimal (Feed it the juice) I have my PBO set to drive in voltages at the motherboard’s limits. I have a x6 scalar, and this thing is boosting hard. It still stays in the high 60’s to 71C Max in AVX512 CPU stress. (I’ve seen this thing upwards of 140+Watts now)
Here is the kicker. I’ve begun my journey into ECLK overclocking on my 7950x3d. These chips… all x3d can run at its maximum boost all the time (Bye-bye 1% lows to the point a 7800x3d will crush any intel chip at anything). I am not helping anyone do this (It will turn off the turbo, and boost algorithm). I will point in the right direction. Yes, a 7800X3d can run all cores at 5050mhz and higher, all the time (With the help of ECLK). The same can be done with the 7900x3d, and the 7950x3d. It takes a bios adjustment, editing in regedit, and also giving 2 commands in Powershell. This isn’t all. I’ve already been able to boot into windows with my CCD0 at 5.4GHZ. It’s not stable yet, and I am going to push for around 5.6GHZ on CCD0 (If possible with my set-up). You verry much have to juice the x3d to do it. Positive in the curve optimizer, and also adjusting the PBO voltages manually or setting “Motherboard Limits”. I will report back with some benchmarks once I stabilize.
EDIT: I might add, none of this is to be done without caution. And you very much need a motherboard with an external clock generator. Otherwise, you’d lock in your x3d to 4.2GHZ all core without the turbo and boosting algorithm (Of course it can be reversed)
You lost me at….
Actually no, you lost me way before that!!
Sorry brutha. I tried to keep it not too technical. Basically you can shut off the boost and run these things at a stable clock speed like intel chips. No fluctuating frequencies. It makes the 1%lows (Stutters) go bye-bye for good (At least until software catches up) It’s complicated. This is going to be harder than OC’ing memory. But they do run like that too. It just takes some work. And for the other part. I am actually Overclocking my 7950x3d. The thing you’re not supposed to be able to do