HungryHamster

SN850(X) or KC3000.
 
iCUE enabled PSU offer the ability to monitor and control various aspects of PSU performance from the desktop.
 
At least keep the ML fans on the H150i. They are able to generate more static pressure so the cooler performs to spec.
 
It is possible to have a mixture of ML and QL fans in a system. They simply can't have lighting control off the same controller. See https://forum.corsair.com/forums/topic/172000-running-ql-and-ml-rgb-fans-together/

You know it wouldn't hurt to have the extra headroom. If you plan to upgrade to the 5000 or 6000 series of GPUs, your CPU might still be able to push it just fine at higher graphical settings. At that point, who knows what power requirements the GPUs might have, but it's most likely to increase over the 4090. The only reason I went with a 750w PSU is bc I couldn't find a better SFX model from a top tier brand.
 
I still think 1000w will serve your current system just fine 🙂 

bro. you're worrying about this way too much. The 13900k is a power sucker at full load. I don't remember what you're going to do with your system, but a quality 1000w PSU will be able to handle it. I thought you were going with the 13700k!!!
 
Luckily the 4090 doesn't have the terrible transient issues of the 3090, so at least that's one positive thing.
 
Do you plan to do something that will utilize the CPU and GPU at 100% for extended periods? Do you plan to OC on top of the stock numbers? If you plan to OC both components, then yes you should upgrade to a 1200 or 1600w unit.

Your RTX 4090 will consume 450-500w ... your 13900K will peak to around 300 watts, if you let it. The rest of the components in the computer are unlikely to consume more than 100w
You can set the PL limits in BIOS and that will cap the power consumption to around 250w (that's the stock/default limits, 300w is with limits unlocked) and you'll lose very little performance, and depending on motherboard you can probably set the limits lower so that the cpu would be capped at around 150-200w (with more significant performance losses)
 
If you go over 1000w by a big amount (usually power supplies will have a 50-100w buffer above 1000w)  the power supply should simply shut down.

Most motherboards worth buying already come with a M.2 heatsink attached, no need to spend extra to get a second one. 
 
Besides, if you happen to get a board without them and  want one in the future, a heatsink can be had for $5 on eBay and AliExpress. 

Agreed, the 4090 is mainly for rendering frames though for animations, plus eventually I'll get a higher refresh rate monitor when QD OLED's become more available

S’all good.  Works now.  I personally have found 20gbUSB to be very close to an orphan system.  Kicking it down to 3.1(10gig) won’t do much. The issue is your motherboard and your case were designed in different generations.  I suggest getting one 3.2->3.1 and a 3.2->3.0 though instead and just leaving the 2.0 alone.  That way you’ll have 3.0 on your front panel instead of 2.0.  You’ll need to get the adaptors for your case connections though.  They are widely available. Mo’ money of course.

I haven't seen a 4090 in person yet, so it's tough to tell. If they increased the fan size, then the GPU holder you've found might not work at all. Otherwise your case looks like it can support 7 slots, it just depends if your motherboard starts the first PCI-E 16x slot at expansion slot 1 or 2 (counting from the top). If it is 1, then yes technically this holder "should" fit. If your motherboard starts the GPU slot at 2, then probably not as your GPU is 3+ slots and this holder requires an additional 3 slots. I mean it does have free returns at Amazon for prime members. Can't hurt!

The length of the gpu is basically 14 inches and the length of the case inside is  22in so it should have sufficient space

If you're just gonna enable XMP and call it a day, good luck noticing a performance difference, and good luck getting the 6600 MT/s CL32 kit to enable XMP. 
 
If you want to do RAM overclocking, 6400 CL38 and 6600 CL32 are the same memory IC (Hynix M die) and both can do the other's settings. M die is pretty consistent and even OEM sticks can do 6800MT/s CL34 (given an unlocked PMIC), with the only real differences coming down to subtimings, max frequency, and tRCD/tRP, which subtimings aren't part of the XMP profile, max frequency only ranges between 6800 and 7000MT/s so both of those kits should clock the same, and 2 ticks on tRCD isn't gonna be noticeable in anything but SuperPi 32M. 
 
No matter what, you should just get the one that's cheaper, and that will almost certainly be the 6400MT/s kit. Heck, it's probably not worth going for a kit faster than 5600MT/s CL40 since that's also a Hynix bin if you want to overclock (though it's not guaranteed like 6400 CL40 is), and at XMP there isn't gonna be much of a noticeable difference (maybe 5% faster in most realistic tasks, not specific RAM benchmarks).

Really depends on if the CPU and motherboard can do those speeds. The 6600 C32 is going to be theoretically faster, but the other thing to look for is first word latency. If the c32 is 28ns while the c38 is 12ns, the c38 may actually come out ahead in day to day tasks.

I'd argue that you don't actually want a Gen 5 SSD, let alone 2 of them, especially on a LGA 1700 based system. The reason why is because the only way to get a Gen 5 SSD on a LGA 1700 CPU is to steal lanes from the GPU because the CPU only has 16 lanes of Gen 5, 4 lanes of Gen 4. In order to get a Gen 5 slot, it would have to bifurcate the PCIe slot your GPU is using, reducing it down to 8 lanes and using those 8 lanes to go to an M.2 slot, and since no LGA 1700 based CPUs natively support 8+4+4 lanes of Gen 5, you can only get 1 M.2 Gen 5 slot without resorting to some fancy expansion cards instead. 
 
If you are dead set on getting a Gen 5 SSD (remember, the difference between Gen 4 and Gen 3 was barely noticeable for the majority of consumers, Gen 5 is well into the point of diminishing returns outside of some very specific scenarios), you're realistically better off going AM5 instead since that's a platform actually designed around Gen 5 SSDs, and since I'm guessing you're also planning on going 4090, you don't want to be stealing 8 PCIe lanes away from your GPU like you'd have to on Z690/Z790

From the chipset diagram, looks like no: 

 
No 5.0 splitout for M.2 drives from the CPU or chipset. Apparently AMD's new AM5 can support em though. 

Thanks for the info. The case is the Corsair 5000T so it does have holes above the PSU shroud. I like this PSU because of the design. It has cool RGB and an OLED screen that shows the wattage.

Depends on the motherboard and memory configuration. The photo below is for the MSI Meg Z790 series and with 1 dimm per channel, 1 rank it supports up to 6800 mhz, with 1 dim per channel, dual rank or 2 dims per channel, single rank then you're looking at 6400mhz, if you have 2 dims per channel and they're dual rank the maximum is 5600mhz. In your specific case the worst case scenario is 6400mhz if those memory sticks are dual rank. Nonetheless at the end of the day I don't think the difference in memory speed matters that much 5600mhz is still plenty fast for any cpu.

Usually, a well built PSU should have some immunity on dust. The solder joint usually covered in a layer of plastic to avoid dust from short circuiting pins and contacts. Still, there's possibility of dust build up that can trap heat and reduce the component's life.
 
If you practise good maintenance of your system (proper cleaning your PC at least once every half year), you should not have to worry.

That only lasts a few years. It comes with a 3 year warranty for these features for a reason.

Just blow compressed air into the psu and call it a day. Do make sure that the case you choose has holes at the top for the psu to take air from. Else just flip it around and mount it fan down like usual.
 
Any reason for this psu? A rmx1000w is at least 100$ cheaper and just as good. This asus you pay for the design.

A good quality 1000w is plenty!
 
13700k=12900k roughly. Paper math the 12900k isn't going over 300w stock. A rtx 4090 isn't going over 450w stock. That's 750w. Plenty of room on a 1000w. 
 
13900k? Let's go worst case scenario and say it uses 350w under stock load (it's gonna be less) with a 450w GPU. That's 800w. You're still good. 
 
Overall, it's all good. 

Whenever I hear this, it has often been in the context of measuring power from the wall to answer questions like how much power does a PC actually use.
 
It shouldn't have anything to do with whether you should or shouldn't use a surge protector (answer: you should for your PC).
 
A PSU requires more power from the wall than what it can provide to the PC parts its connected to since PSUs arent 100% efficient at converting power for parts' Voltages, etc.
 
I think measuring power at the wall tends to be more accurate than software (HWinfo, etc), but software measures the amount of power used by parts.
 
EDIT: As for when power from the wall matters:
* Picking a UPS (battery backup), because those have limits on how many Watts they can provide, and the more watts used then the shorter time your PC stays on.
* Really high power usage, like 1200W+, because that has the potential to limit other power available to a circuit (e.g. 15W), and if exceeded then the circuit breaker will trip off
* When connecting to a generator (at your own risk, make sure its grounded), making sure not to use more power than what it can generate
 
Kill-a-watt is a name brand physical power usage monitor that I use for my own reference since I care about all 3 for PC and other things.

surge protector is not gonna hinder your power draw, assuming it's working properly. A good surge protector can even be quite practical to help protect your PC.

IIRC even the 13600k will suit you just fine. Depending on what you're running.
I mean if you reeeeeally don't want to lose a single frame, then yeah, you're gonna have to go with the fastest CPU.

I think the CPU bottleneck is more of a concern with lower frame rates, like if you are getting 30 to 40 fps in a game where a better CPU could achieve 100+. Or if you plan to pair your 4090 with an i3. I don't think there is anything unsettling about it. You are getting the 13700k anyway, which is supposed to have more cores and 10%+ IPC improvements over the 12700k.
 
You are correct. In the case of Tomb Raider, they were only using high settings and ray tracing was disabled. Maybe this was a testing methodology thing for GN to use high settings, but you can easily push the graphics settings higher even at 1440p. And if you are still CPU bottlenecked, so what... you should be around 200 fps without RT. By definition it might be a CPU bottleneck, but so? lol CPU bottlenecks on the higher end aren't as big of a deal. As newer games come out that tax the GPU more, that CPU bottleneck will fade away.

In 4K and higher it won't happen yet. 1440P yeah, but we are talking about 200 FPS anyway, so does it matter that it "bottlenecks" ? When more graphically demanding games come out it will "unbottleneck".

Thanks for everyone's response. Perhaps I should have stated that I'm talking about a bottleneck while playing in 4k resolution. It looks like we'll just have to wait and see which is a little unsettling because I've already pre-ordered the 13700k.