Here’s a video where Scott Stankard, Director of Product Management at AMD is talking about the asymmetric design. Also mentions that Curve Optimizer is enabled on the new X3Ds.
Interesting.
Scott confirms what we were saying, that the cache setup is asymmetric and that the higher clocks will only be attainable on the CCX without the Vcache.
At the same time, it is comforting to hear him hint AMD worked with Microsoft and others during product development to tackle scheduling for the asymmetric design on the higher end CPUs.
AMD has provided an inhouse comparison between 7950X3D and 13900K for some games. The comparison with 7800X3D on those games will be close.
The gap could even widen with MSFS.
The answer is yes:
I expect 7800X3D to be a better choice than 13900K for MSFS.
That’s still what I’m hoping, but I’d take AMD graphs with a grain of salt - not sure we can realistically expect that the gap will be better when running MSFS in VR. But I guess than even a +10-20% increase would justify going for the new x3d cpus since their price is not likely to exceed the 13900K anyway… Still, I think I might wait for the real benchmarks to be sure…
Yes, as you point out there is much more than top end raw FPS performance at play in this choice.
1% / 0.1% FPS, overall game smoothness, CPU pricing, AM5 platform longevity, …
All of these factors point to 7800X3D being a better choice than 13900K for MSFS.
Then again, for workstation related workloads, the higher end AMD CPU will be a great choice as well.
And for a lot budget minded consumers, 5800X3D remains the value king for MSFS. Even lower pricing, DDR4, lots of people get to keep their MB, and very good performance as well.
Well, as far as I’m concerned I was already ready to invest in something as pricey as the 13900K, and now I’m in fact considering whether the 7800x3d might in fact deliver all that I need (purely focused on vr gaming here) as long as it performs at least as well as the 13900K for vr gaming, although I’d still be curious to figure out whether the 7900x3d and 7950x3d would bring anything more for me, but it doesn’t look like it.
Purely VR for MSFS, we need to see actual real life benchmarks, but I would say it will be worth the wait at the very least
The way to prevent this is to set the affinity mask after the “usual startup music” starts playing during startup. (delay the setting for a few seconds)
When starting MSFS, MSFS itself (presumably) checks and retrieves the currently available cores from the OS and starts the thread, causing inconsistencies and failure to start. The timing is the moment when the XBox logo, MSFS logo ends and the MSFS window turns dark and blue.
That’s easy, I do a lot of delayed tasks in my .bat files with timers.
I just realized AMD posted the official product page, and it lists an interesting figure under Connectivity:
System Memory Type
DDR5
Memory Channels
2
Max Memory Speed
2x1R
DDR5-5200
2x2R
DDR5-5200
4x1R
DDR5-3600
4x2R
DDR5-3600
Does it mean DDR5 6000Mhz is a waste of money? Or am I not getting something? Lower speed memory can be had with lower latency and for much cheaper. I need 64Gb, and the price difference for 2x32Gb modules is huge…
This is the only data rate that AMD supports, and each memory manufacturer’s independently validated memory will operate at higher voltages and faster speeds. Thus, there is heat generation.
The standard memory speed is the JEDEC standard 1.1V, but in the case of G.Skill Trident Z5, for example, it is 1.4V (EXPO profile).
Hence, you are on your own if it does not work. However, there are a good number of Ryzen7000 series running at 6000+.
It seems that when configured with 4 memory cards, it is not running as fast as the manufacturer or as a matter of fact. Most stay at 4400-4800.
Thanks. Well, than I’d have to take the pain of getting 2x32Gb DDR5 6000Mhz. Ouch…
Running DDR5 with four sticks is very hard on the memory controller, particularly with AMD due to the infinity clock concept limiting memory frequency. If you need 64 GB, buying a kit with 2x32 GB sticks is more expensive, but will save you potential headaches down the road.
Yes, that’s what I figured as well, based on my research. Well, I do need 64Gb so I’ll have to get 2x32 kit. The bog question is what speed?
Even the I9-13900K only “officially” supports up to 5600 MT/s:
I’m probably going to go the the 6000 MT/s, myself. That is AMD’s confirmed “sweet spot” for these chips.
I’m following here, as I’m in the same boat. I’m not opposed to spending whatever I need to get maximum performance, but what’s going to give me maximum performance?
I just purchased a 4090, but my current i9 9900 is now sucking wind.
So go Intel 13900, or wait for the 7800 or 7950 X3D?
Following.
7800 or 7950 X3D will definitely beat 13900. They are just around the corner: AMD published February 14th as the launch date, but then backtracked and “unlaunched” it. Still, sometime in February they will go on sale. If you just want the best, there’s no question. 7950X3D is the answer. AFAIK Itel 13900KS boosts higher but after just a few dozen seconds is throttling back due to immense heat. It’s an oven, compared to AMD.
If it’s the budget, considering new expensive board and DDR5, depending on everyone’s existing setup AMD way can be much more expensive.
I’m pretty much decided on 7950X3D unless something bad comes up in benchmarks etc.
At this moment the only remaining question is that.
How will 7950X3D compare with 7800X3D in real life?
We know 3DVCache will benefit both.
We also know 7950 clocks much higher - on the cores that are not equipped with 3D Vcache.
We don’t know what these differences will mean for MSFS.
Of the two, 7800X3D will be the better value for MSFS, that’s for sure.
It would be a nice surprise if 7950X3D performs vastly better for MSFS, out of the box.
We can only hope 7950X3D won’t have worse performance for MSFS, than 7800X3D.
But the actual gaming performance of the hybrid CPU will depend on scheduling, and is likely to evolve over time, with OS updates.
Most likely 7800X3D will remain the natural choice for those only interested in MSFS.
7950X3D being better suited for people who need the additional cores for other tasks (video editing, virtualization etc).
13900K…by far the best
tl;dr - not good.
i have an ancient i7 7700K with a great performing MSI 4090 and the cpu is holding it back (not as much as some would think though ). i’m going to buy a 7800X3D as soon as they’re available for purchase.
It’s not that surprising a 5% frequency increase results in a 5% performance improvement for single-thread. It’s clear this will be a billed as superior binned 13900K for overclockers. Most current 13900K can easily be pushed to 6 or even 6.2 Ghz without much effort, which makes the 13900KS a rather pointless effort.
13900KS is a bit like 12900KS or 9900KS: there is market for a binned chip with better thermal properties.
I think it’s fair to say customers who overclock their i9 chips aggressively will expect better longevity, rather that vastly superior performance, with these higher priced CPUs.