OmniDyne

You’re overstating the SLC caching “issue” with SSDs that utilize 3D TLC NAND, and you’re also incorrect about read speeds “tanking”. SLC caching (also known as pSLC, or pseudo SLC), when consumer SSDs are concerned, mostly helps with sustained writes, random or sequential. When this buffer (cache) is overridden, writes either fall on the base NAND or writes are flushed to the NAND and written to the cache simultaneously. Most recent SSDs (like the EX900 or 970 EVO) will write directly to the base NAND after the cache is filled. Overriding the cache typically happens from very long sustained writes or short extremely fast writes, e.g., a faster SSD writes to a slower SSD. Once the cache is exhausted, you can expect drops in write performance to be multiple hundreds of megabytes per second; still far above a hard disk, if using a semi-decent SSD. Your performance does not “turn to shit” when you override the pSLC cache. Performance is what it is. You’d be hard pressed to find a better and faster solution, within reason. What does this mean? Can you elaborate further on this? Plenty. Long duration read speeds? Elaborate on this, as well. Your post really isn’t clear with as to what you’re attempting to accomplish.

A drive with a lower TBW rating isn’t any less reliable than one with a higher TBW. The WD Green SSD was/ is a pretty awful SSD.

The TBW rating really doesn’t matter for anyone gaming or just using an SSD as a boot drive. I think you’d be hard pressed to find many people (content creators or enthusiasts aside) that write more than 80 TB to an SSD over a 5 year period. 250 TBW is far beyond anything most people would write to an SSD during the lifetime of the computer it is in. Look at TBW ratings for older SSDs that used planar NAND and you’ll find TBW ratings of 80 TB and below, and I’ve seen many people far exceed the TBW rating and the drive still functions as if new. Remember, TBW is only a warranty metric, not a guarantee of anything. Also, WD is a brand. As with all brands, they produce great drives, and they also produce terrible drives. Brand alone means nothing. Look at the reviews of the individual products themselves. The WD Blue 3D happens to be one of their better drives.

Macrium Reflect does a great job of cloning hard disks to SSDs.

Okay, and? The P1 wouldn’t provide a noticeable difference in boot times compared to really any other SSD; certainly not an additional 20+ seconds. To the OP: How full is the drive? When you say you rebuilt the computer, what changes did you make besides the SSD? Did you clone from the old SSD to the new SSD? As previously stated, installing utilizing UEFI/ GPT is important. Did you previously have fast boot enabled? Is it possible you forgot to enable it after the rebuild? Boot times can be drastically different on different motherboards; I use an SN750 at work and boot times are significantly slower on the newer Gigabyte motherboard I recently received when compared to the NZXT board I was using, for example.

What background services?

If it's read-heavy, the amount and length of access isn't going to matter much; DRAM isn't going to help much, if at all, in this situation. SSDs that omit DRAM caching don't just automatically operate at hard disk speeds. In consumer SSDs, DRAM is typically utilized for logical block addressing, and that's it. That being said, SSDs utilizing QLC NAND may operate at or below hard disks speeds during write-heavy sequential workloads; random performance will ALWAYS far outpace a hard disk during heavy workloads.

This is false. DRAM-less SSDs will outperform a hard disk in any task that isn't an "elevated" workload, as it's put. What does this mean? If you're a workstation user. Is gaming a workstation use scenario? No. Are OS tasks workstation scenarios? No. What tasks are you subjecting the server to? If it's streaming, no.

I would check to be sure that you installed Windows in AHCI mode. I believe that some SSDs will bypass the SLC cache if the SSD isn't overburdened. I wouldn't consider 300MB/s poor or unusual. Additionally, as you see in the picture you posted from Anandtech, 520MB/s is an "up to" figure and if you read the review from Anandtech, in particular the 'Light' storage test, the performance of their drive was quite similar to yours. They do have higher performance metrics during the 'Sequential Performance' test, but they're also hitting higher queue depths, something you likely aren't subjecting the drive to, so performance will be lower at QD 1. In my view, I don't see anything unusual with what you're experiencing.

How much space is remaining on the drive? What are you doing with the drive that you’re experiencing 300MB/s? If you’re transferring from drive to drive, please tell us what make and model the other drive is.

There likely would not be a noticeable difference.

Write amplification is something that is native to SSDs no matter what you do. You can’t prevent it and it’s even caused by normal SSD on-board tasks, such as wear-leveling.

This generally isn’t accurate. As far as write amplification is concerned, I’d say this is extremely inaccurate. Manufacturers set aside a certain percentage of flash that is inaccessible to the user. This flash cannot be written to by the user; it is used solely by the controller for maintenance and to prevent exactly what you stated: slow downs. This is know as over-provisioning and every modern consumer SSD I’m aware of utilizes this technique. Is filling any storage device to maximum capacity ideal? No. Does it matter? Not really. You have nothing to worry about. Use your SSD as you normally would.

An M.2 PCIe SSD will not function as a SATA SSD. However, motherboards often do support both SATA and PCIe over M.2. MSI, Gigabyte, Asus, ASRock; they do not specifically state M.2 PCIe SSDs utilize NVMe in the manual or on the product page. They all state "M.2 PCIe x2/x4 support". The only mention of NVMe I've found regarding storage has to do with configuration or PCIe x16 NVMe support or booting via NVMe. That's not what you stated. You stated "there is no real performance difference". This is false. Consumers can absolutely realize performance differences when transferring from drive to drive. What motherboards specifically state support for M.2 PCIe NVMe? I haven't seen one yet. As I stated above, there are only specific cases where NVMe is mentioned and it's never regarding M.2 PCIe support. https://download.gigabyte.com/FileList/Manual/mb_manual_z390-aorus-ultra_1001_190219_e.pdf https://dlcdnets.asus.com/pub/ASUS/mb/LGA1151/TUF_Z370-PLUS-GAMING-II/E14720_TUF_Z370_PLUS_GAMING_II_UM_WEB.pdf?_ga=2.239004355.2100305124.1584368493-1709474572.1582643030 https://download.gigabyte.com/FileList/Manual/mb_mamual_z370p-d3_1003_e.pdf https://download.asrock.com/Manual/Z390 Phantom Gaming 7.pdf There are consumer motherboards that only support PCIe AHCI?

M.2 is the form factor. PCIe is the interface. NVMe is the protocol. My explanation was not "in depth". It's labeled very clearly in the manual multiple times that it supports M.2 PCIe. It's not going to say NVMe because NVMe is the protocol PCIe uses. It can't support M.2 PCIe without NVMe. There's a massive performance difference. PCIe NVMe SSDs can hit well above 4000 MB/s. SATA is limited to ~500 MB/s. That being said, for OS or gaming applications you will not notice a difference unless transferring massive files from drive to drive. 2.5" SATA SSDs are the same as M.2 SATA SSDs. Same interface, same protocol.