19 minutes ago, Lotharen said:

go for the M.2 NVME. They are more expensive but much faster that SATA models.

Speed differences that you'll never be able to tell in a blind test on games or during normal use of the computer, you won't even be able to tell SATA3 from NVME.

1 minute ago, EmaRacing said:

Why Intel though?
AMD's price\performance in multicore is unbeatable ATM.

I think that's more a matter of "taste" than anything. I like Intel and will stay with them unless the other team is significantly better in something. Having a lot of cores has been AMD's selling point recently, but MSFS basically tops out at 6 physical cores.

3 hours ago, Colonel X said:

So I am running the good ol' 4790K, clocked at 4.6Ghz. The new 10700K seems to be the ideal CPU for MSFS (the 10900K has more cores, but MSFS doesn't utilize them).

 

How many cores does MSFS use then? I know that it uses all four of mine equally.

5 minutes ago, Kilo60 said:

Isn’t Ryzen better at multi core applications rather than an Intel CPU which has better single core performance to which MSFS would utilize better?

Yes, but that performance gap will almost certainly shrink even further with Zen 3.

Also, in the grand scheme, it doesn't really make such a big difference. You get, what, 5 FPS more or something when you go for Intel, and at a prohibitive price, with much higher power consumption and cooling requirements.

Wouldn't be worth it for me, as MSFS isn't the only thing I use my PC for.

1 minute ago, desbean said:

How many cores does MSFS use then? I know that it uses all four of mine equally.

It will benefit from up to 6 cores (12 threads). Anything more won't do anything right now, so even the 8 core 10700K will be somewhat underutilized.

57 minutes ago, UAL4life said:

I guess I could google it, but for the sake of discussion, will Ryzen have improved clock speeds instead of cores?

Improved IPC, clocks will only be very slightly higher. But improving the architecture is a superior solution to boosting clock speeds, because this will disproportionally increase power consumption as well. Comet Lake consumes lots of power because of that reason.

10 minutes ago, Kilo60 said:

Isn’t Ryzen better at multi core applications rather than an Intel CPU which has better single core performance to which MSFS would utilize better?

With the same amount of cores at the same clocks no, as Intel cores perform stronger in games overall. But with AMD you can get the same amount of cores for less money.

4 minutes ago, EmaRacing said:

Speed differences that you'll never be able to tell in a blind test on games or during normal use of the computer, you won't even be able to tell SATA3 from NVME.

Things could change with DirectStorage and implementations like RTX IO though (which will only be compatible with NVME drives). A few months ago I would recommend SATA SSDs, but now budget NVME drives (like the Western Digital Blue SN550 and Kingston A2000) are so close to SATA prices that it makes sense to step up to one.

Edited September 16, 20205 yr by ChaoticBeauty

3 hours ago, Colonel X said:

So I am running the good ol' 4790K, clocked at 4.6Ghz. The new 10700K seems to be the ideal CPU for MSFS (the 10900K has more cores, but MSFS doesn't utilize them).

I've been looking at https://www.cpubenchmark.net/singleThread.html for single core performance, the single most important factor in flight simulation. 

4790K @ 4.0 = 2.47 Performance points.

10700K @ 3.8 = 3.08 Performance points.

That's a performance increase by 25% (not considering my overclock, or the 10700K's turbo speeds up to 5+Ghz), or in other words, if you get 25FPS at a situation in MSFS, the new CPU would run that with 31FPS. While that represents just the increase you need to kill the dips below 30FPS (I lock at 30FPS), and additionally will tackle the stutters caused by core overload, it's still, by all means NOT MUCH, considering the 6 years between the CPU's, and the roughly 700-900 bucks needed for the CPU, Mobo and RAM.

Seeing that I am basically locked at 30, and I only get dips at, say, LAX in "bad positions" (such as 07L looking at the entire airport with traffic) to maybe 27, I really wonder if it's worth the spending (and work). Am I missing something here? Will the increase be larger because the 10700K will run at turbo speeds (probably). Sure, once we get more complex add-ons, dips will increase, but overall, it's kind of depressing that even the fastest CPU can't lift your performance more than a mere 25%.

It's the primary reason why I'm still on an i7 2600K. If I can't get at least double the fps, it's not worth a 1K investment.

3 minutes ago, EmaRacing said:

Speed differences that you'll never be able to tell in a blind test on games or during normal use of the computer, you won't even be able to tell SATA3 from NVME.

Some of the quest for speed in computer building reminds me of those car audio competitions where they fill a minivan with giant speakers and amps for a system that's tuned to put out a single tone at the highest volume possible until the speakers blow (which happens very quickly) - a volume which would melt your brain if you sat in the car when it went off. Obviously, 99% of the power going through those things is absolutely pointless, but people do it anyway. 😉

 

I'm only running at 1080p, but I'm pleasantly surprised not to feel the need to upgrade my 2 year old build. MSFS has definitely broken the old reality where you had to wait several years after the new flight sim came out before anyone was even making hardware that could run it halfway decently.

 

 

10 minutes ago, EmaRacing said:

Speed differences that you'll never be able to tell in a blind test on games or during normal use of the computer, you won't even be able to tell SATA3 from NVME.

BullS&$%  I play wow and I notice a HUGE difference in load times when logging in, changing zones, ect compared to when I had a SATA3 SSD. So totally untrue.

9 minutes ago, Der Zeitgeist said:

with much higher power consumption and cooling requirements

That's caused by a dated design being pushed to its limits in a desperate attempt to not look too bad (on paper) compared to their competitor.
I built the first AMD System in 20 years because of this. 190€ for my 3600 felt like stealing.

Just now, EmaRacing said:

I built the first AMD System in 20 years because of this. 190€ for my 3600 felt like stealing.

Yes, going for Intel in 2020 is simply unwise.

6 minutes ago, Lotharen said:

BullS&$%  I play wow and I notice a HUGE difference in load times when logging in, changing zones, ect compared to when I had a SATA3 SSD. So totally untrue.

Have you tried different technologies on fresh installs? I believe something else was slowing you down.

Yup similar experience to you.  I had a 4790k and been looking at upgrading for a while and the performance gains were pathetic.  In the end I went with a Ryzen 7 3800+ (I got a good price for it from Microcenter) and that was enough for me to upgrade because I doubled my cores. But I also wanted something for some other things I do outside of gaming.  But yeah it was depressingly pathetic to look at year over year.

I went from an over locked 4770k, that stuttered every 5 seconds and was basically unusable even on low settings to a Ryzen 7 on stock settings. I now get well over 30fps on high (only running an rx5600xt) and GPU limited even if I switch to ultra. The difference is night and day, despite the limited performance difference per core on paper.

1 hour ago, Colonel X said:

Pretty sure these benchmarks are conducted with the fitting RAM. After all, these CPU's are all made for a distinct RAM speed.

No doubt, and not that I'm an expert on benchmarking, but I have seem some synthetic benchmarks that mostly measure computationally heavy CPU tasks on relatively small amounts of data from RAM, so RAM speed doesn't affect performance much. And I have seen gaming benchmarks where increase in RAM speed gave a not insignificant performance gain (though it became marginal beyond 3200 Mhz).