Lolight

No, it's not the case. Intel desktop CPUs are designed and fully capable of entering the highest possible power-saving C-states. Mobiles don't seem to have any advantage in that respect, otherwise everyone on this German thread would be running nothing but laptop CPUs. https://www.hardwareluxx.de/community/threads/die-sparsamsten-systeme-30w-idle.1007101/page-96 Please check out the list: https://docs.google.com/spreadsheets/d/1Qlps7qDw9ocdJ82x0BRxEbXeBWnqkWiya6yszSnurv4/edit#gid=0

Within the same Intel CPU generational architecture? AMD desktop and mobile designs have significant differences. An Intel mobile CPU would be a permanently under-clocked, under-volted, featuring a lesser specced iGPU and generally offering a fewer number of cores versus a desktop version of the same generation processors. How could it consume less wattage to complete the same work load as compared to a faster running desktop chip of the same generation, with the same number of cores? It's pretty much the same chip, with the exact same capacity of processing instructions per cycle, only slower. Though It can only be said when comparing the same gens of processors since each sequential generation normally gets improved in every possible way including its manufacturing methods by continuously shrinking process nodes. I'd argue that it's quite possible for the mobile chip system to actually consume a little more power overall, since the CPU will not be allowed to fall back into the low power idling state until the task is completed. While the faster desktop CPU will get work done sooner, then enter a low power mode while allowing the rest of the system components to also reduce their power consumption. I'm not aware of a specific mobile vs desktop test but found this video to be very informative:

The 50C warning is for HDDs only. Those temps are not considered to be too hot for active NVMe's. You can adjust the warning value higher in the main tab by clicking on the device.

In general All HBAs consume more power than ASMedia or JMB582/585 cards. HBAs have been shown to always interfere with the CPU's idle power saving features (C-States). HDDs: sleep - 0.5W idle - 5W max data transfer - 10W initial spin up - up to 30W There is a huge thread discussing ways to minimize power consumption: You might also want to check this German-language forum: https://www.hardwareluxx.de/community/threads/die-sparsamsten-systeme-30w-idle.1007101/page-96

Some motherboards disable specific on-board SATA port(s) in certain scenarios - have to consult the mobo's manual. I'd say 14-16TB is the most cost-effective range at the moment. The storage space provided by bigger HDDs might even eliminate your need for extra SATA ports for a long while. And by then you might decide to move to better, newer hardware anyways. It all depends on your particular situation. It's generally recommended to start with the biggest HDDs you can afford at the moment. Each additional HDD consumes electricity and becomes another potential point of failure.

What amount of data do you plan to have in the system? 4 x 20TB HDDs with a parity will provide 60TB of usable space.

Just be aware that HBA's have downsides, especially in the view of power efficiency. A good SATA expansion card could be another, maybe better option.

Welcome! That article expresses an opinion. There are differing opinions on that point. HBA's have downsides and should be avoided if possible., in my opinion. How many HDDs in your system?

Welcome! Assuming that you're purchasing new components... Are you sure that you will need to have an HBA in your system? There are readily available motherboards featuring 8 SATA ports on-board.

As stated on Plex forums... Current Plex implementation is: Intel: HW transcoding - HDR → SDR tone mapping with Intel Quick Sync Video AMD: HW transcoding - SDR only (HDR is not mapped) Nvidia: HW transcoding - HDR → SDR tone mapping with CUDA https://forums.plex.tv/t/do-i-need-quick-sync-synology-alternative/853278/6

This LimeTech should really consider offering such an option, preferably at a lower price. Many users are not interested in "advanced" features, especially at the cost of additional complexity and associated with it lower stability. Would also greatly expand Unraid's adoption rates in the lower income countries.

Sounds like another case of a shady product, likely originating from the same pool of cheaply made circuitry as the rest of fakes. Only this time sold and branded by a popular computer hardware outlet! Apparently Micro Center's purchasing managers couldn't be bothered (or woudn't care) with conducting any research as related to the product's origin and quality and have made their sourcing decision based purely on price. And why wouldn't they? Since USB drives are commonly treated as throw-away devices.

The first post of the topic links a bunch of the "best" flash drives to purchase which is based on the 3+ year old test which doesn't even mention any widespread cases related to proliferation of fakes. It performs a bunch of durability tests that don't have any relevance as applied to the specific Unraid environment. It also doesn't spell out differences between USB 2.0 and USB 3.0 as related to thermal properties either. It's very likely that Spaceinvader wasn't even aware of those factors at the time of making the video. With all do respect, a good back up strategy is a must but I wouldn't call it a good advise to "not worry too much". The quite frequently posted reports of failed boot devises on this board and also everywhere else speak volumes about the importance of choosing a long lasting drive. With users reporting a failed USB while not having, displacing or forgetting to maintain a proper configuration backup. Which is real common since USB drive backups are often treated as an afterthought. It would only be wise, based on all available info at the moment, to make at least a minimum effort in selecting the potentially longer lasting device.

It's been concluded on this thread that a mere presence of an HBA in the system would generally prevent the CPU from reaching anything higher than C2. As opposed to SATA cards that don't present the same limitation.

I've compared few of my smaller 8-32 GB USB drives and found out that their capacity is not being the determining factor. I don't have bigger capacity drives to test. They were all tested via a USB 2.0 port. In another similar thread I've already mentioned about significant differences in the USB drives current requirement designs. IMO, the USB drive controller is what determines how much current will be supplied by the port. A badly designed drive (or a fake) will be drawing the maximum current allowed by the USB 2.0 standard which is 500 mA. One of my tested USB 2.0 drives did exactly that. And it would likely be even worse if connected to the USB 3.0 port. This article discusses design variances: "if your USB Device exceeds the default 100 mA, but you do not design it to control the various current levels as they are enabled, then there is really nothing that the USB Host can do to limit your current. The USB specifications are designed with the assumption that USB Devices will properly advertise their current needs, and properly control their current draw based upon listening to what the USB Host permits. The only thing that the USB Host does is say "Yes" or "No" to a given USB Configuration. After that, it's up to the USB Device to control the current." http://dangerousprototypes.com/docs/Designing_USB_Devices_for_proper_current_and_MaxPower The rest of my USB 2.0 and USB 3.0 drives were drawing in the range of 100-225 mA. I don't think any of the USB 3.0 ones appeared to draw less than 200 mA. Based on that observation I'd say that in order for the boot device to run the coolest it would be desirable to get a genuine USB 2.0 drive that's been verified to draw only 100 mA. And then to make sure that it's attached to a USB 2.0 port or header. To check for the drive's current metric in Windows go to Device Manager --> USB Root Hub --> Power Or even better, use the USB Device Tree Viewer: https://www.uwe-sieber.de/usbtreeview_e.html

With HDR tone mapping on?

Plex transcoding is not fully supported on AMD. It's still work in progress. You'd need to go with Intel. https://forums.plex.tv/t/got-hw-transcoding-to-work-with-libva-vaapi-on-ryzen-apu-ryzen-7-4700u/676546/569

The PSU. You'd need to replace it anyways. Soon.

"The second batch of Storaxa will be available for ordering on October 1st, 2023 - 10am(EST). There will be different shipping options available from 3 business days to 14 days. Customers will receive a confirmation email with the tracking number." https://www.arrotrack.com/store So... They've shown no real finished product to date. Have offered to send an AMD sample unit to NASCompares for a review but then decided to delay. The funds are still locked by Stripe with no concrete date on when they'll be released. Yet now offer their "second batch" for sale.

Not needed.

PSUs do "wear" down over time as opposed to CPUs and RAM. Get at least Tier B, mid-range model from the list, in the smallest capacity you can find. https://cultists.network/140/psu-tier-list/

Because it's still a prevalent but IMO probably outdated way of thinking, especially in the view of power efficiency. As applied to the vast majority of consumer NAS users, running SATA drives... I think the biggest reason for their popularity is based on the HBA's uniform build quality since it's "server-grade" hardware. But it's largely an outdated tech, that is guaranteed to mess up modern consumer hardware by disabling CPU's power management features (higher C-States). HBAs on their own also suffer from relatively high power consumption which often leads to overheating when placed in consumer cases - usually solved by adding an aftermarket dedicated fan to keep its temps in check. On the other hand, the SATA card market is hugely fragmented, mostly cheaply built and marketed by off brand vendors, with largely varying build quality. But the right card will preserve your system's ability to sip power at idle, perform just as well as the motherboard's SATA controller, not consume much power on its own without any requirement for additional cooling. When shopping you'd need to pay attention to the PCIe card controller chip and overall configuration - look for the card that is built based on the ASM1166 chip. Most of them normally come with an outdated firmware, which can be "usually" re-flashed with an update that is maintained by Silverstone since their cards are based on the same chip. https://www.silverstonetek.com/en/product/info/expansion-cards/ECS06/ And yeah, you'd need to be really selective and look for the card's user feedback prior to deciding on which one to choose. e.g. @Gragorg seem to be satisfied with his model. You could also get guaranteed quality by going with the card built by Silverstone, but it comes at a much higher price point: https://www.amazon.com/SilverStone-Technology-ECS06-Non-RAID-Express/dp/B08Y64R5WV/ref=sr_1_1?crid=1HQAJ4FKTZE37&keywords=silverstone+ECS06&qid=1695977101&sprefix=silverstone+ecs06%2Caps%2C375&sr=8-1

Yeah, but why would the OP be concerned with TDP (thermal design power) as applicable to his use case? To make his system more power efficient? There's a common misconception that a lower TDP CPU would always be more efficient as compared to a higher TDP equivalent.

It should be sufficient for light loads. btw, I'd advise to replace the PSU with a newer, known quality model.