Jonotallica

Yes it's possible that way.
But first here's the manual link -
https://dlcdnets.asus.com/pub/ASUS/mb/SocketAM4/ROG_CROSSHAIR_VIII_DARK_HERO/E17453_ROG_CROSSHAIR_VIII_DARK_HERO_UM_V2_WEB.pdf
 
secondly yes with the config u said there,
it's possible and works like a charm.
 
But now the problem/query is shifted to will either of the second M.2 or 10 gig networking card get bottlenecked bcuz of shared PCIe lanes between them? If yes then how much?
(And then there are only 4 PCIe lanes in between chipset and CPU, that doesn't help at all.)
 
Does this mobo give an option to change manually between 4.0 to 3.0 and 3.0 to 4.0 (considering that PCIe is 4.0)?
Yes it does. Not only this one but others too give this option.
 
Check out the below link
https://amp.reddit.com/r/Amd/comments/cut59x/change_from_autopcie_40_to_30/
 
 
 
_________________________________
 
I will just leave it here. (A line from meme)
 
You CAN use a PCIE 4.0 device (like a graphics card) in a PCIE 3.0 slot and it will work just fine, it’s just that it will run at PCIE 3.0 speeds. Since the very fastest GPU’s (2080 TI) run just a little faster on PCIE 3.0 x16 than on PCIE 3.0 x8, we’ve JUST gotten to the point where the speed difference between PCIE 2.0 and PCIE 3.0 matters for the very fastest cards in a few certain games and situations.
This means it should be another 2–3 years before the difference between a PCIE 3.0 x16 and a PCIE 4.0 x16 matters for the fastest graphics cards and another few years beyond that before it matters for mid-level cards.
So, it’s not that important for several more years, but if you want a PCIE 4.0 motherboard anyway, you shouldn’t need to worry about compatibility, PCIE 3.0 devices should work just fine with your system (and FWIW even PCIE 2.0 too).
 
 
And always thanks for replying.

Inb4 "racist!!!"
 
I've noticed an unusual amount of people on the forum buy into the PewDiePie racism narratives, which is pretty astonishing given the tech and social media awareness of the general user of this forum

But Micron E-die is almost half the cost, at least where I reside.
Ah, already mentioned .. never mind then. Though still I'd get Micron...

Samsung B die or Micron E die.. B die is better but E die is much cheaper (and almost as good).
 
Use DRAM Calculator for Ryzen
and you'll get decent results with either.  It's recommended to use this regardless of which RAM you choose, as it will be better than any stock XMP profile etc.  Choose which model based on availability, pricing, warranty options etc.  It really doesn't matter if it's Corsair, Trident or whatever else.. it's the die type that matters.
 
And look to how much your "infinity fabric clock" on the CPU can do.  3600mhz  (1800mhz infinity FCLK ) is a safe number for most CPU's.  But that's your bottleneck, not the RAM itself (assuming you have B die or E die).  The higher end CPU's like 3900X and 3950X tend to reach the higher clocks.. as high as 1900mhz (or higher).  But you want a 1:1 ratio with the RAM you choose.  If your CPU can't handle this (it's a silicon lottery) you could just get 3200mhz RAM and save some money.  By the way, if you get 3200mhz RAM, it'll handle the higher speeds just fine with the 3000 series AMD CPU's.  It doesn't need to say 3600mhz on the box, but if that makes you feel better you can choose that one.  But in a lot of situations, when you buy 3000mhz or 3200mhz or 3600mhz, you aren't getting higher binning, you're getting the exact same RAM with a higher XMP profile.

None of the chips in the 3000 series have ample headroom  for overclocking.  They are all pretty much maxed out from the box with factory settings.  BIOS updates, your motherboard VRM, a good cooler.. will help slightly to potentially reach higher speeds but it's more of a tweaking thing than anything else.  Or the ability for the PC to run quietly on max load.  Overlocking is pretty much non existent, and in benchmarks with 3600X, 3700X and 3900X.. the stock factory settings often run faster in games.  It's usually only in highly threaded workloads that the Manual OC performs faster.  Sometimes disabling SMT does a higher percentage improvement.. but that's not something you generally want to turn off.  It's possible to play around with affinity in Windows, to disable certain threads for custom applications, especially with a program like Process Lasso.
 
You are right in that no one got samples of 3800X.  There will be a few days or a week delay before the proper reviews start surfacing.  It's likely that it will be very similar to 3700X, maybe with a slight improvement.  The main difference between the two is that thee 3800X is already maxed out, while the 3700X boosts to a higher clock (relative to the 3800X).  But the max clocks should be pretty similar between the two since that's the way AMD's boost is working. There maybe 1-3% improvement with 3800X.  Hard to say anything for sure until we see the proper tests.  For all intents and purposes check out the 3700X benchmarks that are out already, 3800X will pretty much be the exact same thing.  If you are feeling optimistic, add 2%.

The maximum amps a Ryzen 2700X is allowed to draw period is 115A. Otherwise it typically hovers around 60A continuously and spikes up to 90A.
 
Estimating power draw is not as easy as it may seem. There's two factors in play here. The first is the I/O portion of the processor and the Infinity Fabric eat up a non-trivial amount of power. The second is the more cores the processor uses, the more it limits how fast the entire processor can operate at. For example, this is the power consumption trend for Threadripper 2990WX (from https://www.anandtech.com/show/13516/the-amd-threadripper-2-cpu-review-pt2-2970wx-2920x/2)

 
While the power consumption for cores starts off linearly (double the core load roughly doubles the core power consumption), it starts to taper off at 8 cores before essentially plateauing at 17.
 
A similar thing happens even with the 16 and 12 core variants

 

 
EDIT: Basically speaking, the TDP plus about 10% is what you can expect the processor to use power wise (at least for AMD's processors). Since the processor typically runs at 1.35v, you can use some math to figure out how many amps the thing will suck.

First of all..
It's not AMD's 7nm ... the factories belong to Global Foundries (which was owned by AMD but they sold their part and it's somewhat independent) and TSMC. The design of the whole thing was made by Samsung (the machines and stuff)
 
Also ... you can't put an equal sign between  Intel 14nm and tsmc 14nm and global foundries 14 nm
The numbers 7nm or 14nm are more of a " n nm class", which means most of the things in the chip are 7nm or around 7nm, or 14nm or around 14nm.
 
When AMD, TSMC, Global foundries says 14nm, they mean some structures inside the chip can be 14nm... but for example if a component in the chip is 14nm wide, it may have to be at least 20nm long, a limitation of the manufacturing process.
In the case of other components in the chip, let's say less used, it may not be possible to make them lower than 22-25nm in any direction.
 
Intel may have different rules for what it considers 14 nm ... for example they may be more honest and they may say 14nm for when they make structures that are 14 nm wide AND 12-14nm long, so basically BETTER than the alternative solutions.
 
Anyway... in general, the lower the nm number, the smaller the internal components in a chip can be, which mean all the components can be packed closer together in the chip so
* you can squeeze more components inside a chip in the same surface, or make the chip smaller using the same number of components
* smaller transistors and components inside the chip means these components need less power to do their job, so overall the chip consumes less power
* the distance between components inside the chip is smaller, which means there's less power required to push electrons between components... and this means less heat is produced to achieve same things
* also, less distance means there is potential to reach higher frequencies - with older bigger nm designs some things are simply not possible, because the chip waits a lot of time for electrons to move at the speed of light from one side of the chip to another .. speed of electricity is around 280 meters per nanosecond or around 0.01" per picosecond ...
 
... and there's other reasons...