NVMe vs M.2

Im no particular fan, I just have one as it was cheap, but it’s not as badly designed as you say, you just didn’t buy the right drive for your niche use case. For the vast majority QLC is fine.

I don’t think any TLC SSDs use an MLC cache do they? Never seen it mentioned.

Using encryption like bitlocker etc. is mainsteam these days, not niche. Not even taking into account the poorer durability of these drives.

The vast majority of SSD's are TLC or MLC so the only cache they can have is SLC, I would not expect a QLC drive to use SLC cache.
 
Using encryption like bitlocker etc. is mainsteam these days, not niche.

The vast majority of SSD's are TLC or MLC so the only cache they can have is SLC, I would not expect a QLC drive to use SLC cache.

Hardware encryption is built in, and doesn’t require a whole drive write. So you’re down to the the few users who must use software based encryption only for whatever reason.

Regarding the MLC cache I can’t find a TLC drive that uses MLC as a cache, which should be possible, and QLC appears to be the same. I don’t know why only SLC caches are used, whether it’s a technical limitation or performance related.

A little research would have let you know what you were trying to do was a bad idea. It’s easy to blame bad design but I’m not convinced that’s true.
 
Hardware encryption is built in, and doesn’t require a whole drive write. So you’re down to the the few users who must use software based encryption only for whatever reason.

Regarding the MLC cache I can’t find a TLC drive that uses MLC as a cache, which should be possible, and QLC appears to be the same. I don’t know why only SLC caches are used, whether it’s a technical limitation or performance related.

A little research would have let you know what you were trying to do was a bad idea. It’s easy to blame bad design but I’m not convinced that’s true.

Software encryption is widely deployed because it offers better security features such as multiple cipher cryptography and nested partitions, it also has better compatibility.

However you are missing the point for the most part, only the initial encryption of the disk requires a full write, after that the drive can use a cache as normal. The point was that such a low usage of the disk cripples it's performance to the point it is useless.

If I wanted a few GB of really fast storage I can use a ramdisk lol.

There is far less reason for a TLC drive to use MLC cache than a QLC drive so it's hardly a surprise you don't see that.
 
Last edited:
The encryption is really a moot point as it would perform the same if I did a non encrypted full format. Yes I chose not to use the build in UEFI encryption because it does not meet my security or compatibility requirements.

I was aware of QLC limitations in the general sense, that does not mean I trawled over benchmarks trying to replicate this exact scenario with this exact drive... one does not expect a 2019 SSD to perform worse than a usb pen drive made 10 years ago under any circumstances! Theoretically an empty 1TB QLC drive should have a 500GB TLC cache, however after reaching only 15% encryption the rate plummeted to <50% of the typical TLC drive throughput. By comparison it takes less than 1/4 of the time to encrypt my 1TB EVO 850 array.

You come across as quite defensive of the product, all I said was that I would not consider one based on my experience which you acknowledge as being a valid one.

You seem to have confused defensive with me pointing out that you contradict yourself and make incorrect assumptions/statements in pretty much every post. Still this is the storage forum, so feel free to ‘transfer’ if it makes you feel better :D

You say you understood QLC’s limitations, but you trigger a full drive write and then complain when you obviously knew this would happen once you exceeded the SLC cache.

You say using encryption and doing a full write is a moot point as you’d get the same result from a full format, but ignore that you wouldn’t normally do so on a flash device.

You say you shouldn’t have to trawl reviews to replicate your usage scenario, but out of the 3 reviews I checked at well known review sites, all make it clear a write that exceeds the SLC (not TLC) cache will result in QLC native speeds in very simple terms.

You claim QLC uses TLC cache, but every post above this has made it clear it’s SLC. You also seem to think that a 1TB drive would have a 500GB TLC cache when it’s clearly stated as 12-140GB of SLC on 1TB.

The issue here isn’t QLC, it’s that you didn’t understand the product and/or your own usage scenario, because no reasonable person who actually spent 5 minutes reading a single review and understood what they wanted to do with the drive would be trying to do what you did, let alone still be trying to deflect blame onto the product on a forum for people who are actually interested in the underlying technology.

Again, most users won’t dump 140GB of sequential writes at full speed to a 1TB drive in a single sitting, so for most people a QLC drive will be indistinguishable in real world usage from TLC. If you look at the health reports on most drives that are sold used, the health is usually very high 90’s, people generally don’t write much data at all, so for a games drive, QLC is ideal, which is what the op wanted.
 
It’s a solid buy. The end of mechanical drives in home PCs must be pretty imminent.

It is so weird not hearing any noise from HDDs. Also the fact that the windows login screen appears in seconds, as do all the startup programs shortly after.

I'll keep external drives for any big files, but with all my games on demand I'll just download what I am currently playing.
 
On my z87 MSI gd65 I have an msata slot. If I used an adaptor for a M2 ssd would this take out 2 of my sata 3 ports?
 
You seem to have confused defensive with me pointing out that you contradict yourself and make incorrect assumptions/statements in pretty much every post. Still this is the storage forum, so feel free to ‘transfer’ if it makes you feel better :D

You say you understood QLC’s limitations, but you trigger a full drive write and then complain when you obviously knew this would happen once you exceeded the SLC cache.

You say using encryption and doing a full write is a moot point as you’d get the same result from a full format, but ignore that you wouldn’t normally do so on a flash device.

You say you shouldn’t have to trawl reviews to replicate your usage scenario, but out of the 3 reviews I checked at well known review sites, all make it clear a write that exceeds the SLC (not TLC) cache will result in QLC native speeds in very simple terms.

You claim QLC uses TLC cache, but every post above this has made it clear it’s SLC. You also seem to think that a 1TB drive would have a 500GB TLC cache when it’s clearly stated as 12-140GB of SLC on 1TB.

The issue here isn’t QLC, it’s that you didn’t understand the product and/or your own usage scenario, because no reasonable person who actually spent 5 minutes reading a single review and understood what they wanted to do with the drive would be trying to do what you did, let alone still be trying to deflect blame onto the product on a forum for people who are actually interested in the underlying technology.

Again, most users won’t dump 140GB of sequential writes at full speed to a 1TB drive in a single sitting, so for most people a QLC drive will be indistinguishable in real world usage from TLC. If you look at the health reports on most drives that are sold used, the health is usually very high 90’s, people generally don’t write much data at all, so for a games drive, QLC is ideal, which is what the op wanted.

Where did I state QLC uses TLC cache? I quite clearly said the opposite in the case of that drive.

The encryption/format was an illustrative example of how bad Intel's QLC drive performance is, doing everyday tasks like copying raw photos from my camera to the drive would be drastically slowed down if at only 15% of a format we are below mechanical drive speeds.

QLC itself is a perfectly fine underlying technology that I understood the limitations of when I first bought the drive, Intel's unnecessary use of SLC only cache however is something I did not expect.

You seem so bothered by the fact that I personally wouldn't personally consider a drive like the Intel for my use case I just don't get it.
 
Last edited:
You seem so bothered by the fact that I personally wouldn't personally consider a drive like the Intel for my use case I just don't get it.

I think he's suggesting it's more the fact that you did consider the drive for your use case, which was a terrible idea, and then blame it on Intel for designing their drive with an SLC cache not an MLC cache despite every drive in the market using an SLC cache. You may have found something Intel have missed in their massive R&D operation, but it seems unlikely that theyd have used SLC without a good reason.

Then you go on to say you would not consider one for any purpose. Which I dont think is justified as they will work perfectly well for many people, probably the vast majority.

Really, you should have just read a review and avoided it completely from the off but you dont seem to recognise that.
 
Last edited:
Last edited:
The above is the reason I asked, as I thought it may be the same as Asus, and your limited to the same bandwidth as that SATA port so it will be slower.
Not sure I've understood what you're trying to say, but I thought you're trying to attach an SSD to the mSATA slot, so you're never going to get faster throughput than what the mSATA slot is capable of.
 
I am not to trying to do anything I am not the OP. :)

I asked if it was similar to this:


6ac3e7de06d1667edc2ebc94e56376cd.jpg


N7k00SP


If so you lose SATA Port 6 is you occupy the mSATA side of the adapter (its fine if you use the WIFI/BT card that is pre-installed), and the fastest you will see from any device is the same as the SATA ports bandwidth so 6Gb/s minus overheads so 560-580MB/s.

You can use a PCI-E adaptor but you will lose PCI Lanes for the GPU (it will drop to 8 if that bothers you) in all but the single PCIe 2.0 x4 Slot that I use for my Soundcard but that slot may not be fast enough.
 
Last edited:
I think he's suggesting it's more the fact that you did consider the drive for your use case, which was a terrible idea, and then blame it on Intel for designing their drive with an SLC cache not an MLC cache despite every drive in the market using an SLC cache. You may have found something Intel have missed in their massive R&D operation, but it seems unlikely that theyd have used SLC without a good reason.

Then you go on to say you would not consider one for any purpose. Which I dont think is justified as they will work perfectly well for many people, probably the vast majority.

Really, you should have just read a review and avoided it completely from the off but you dont seem to recognise that.

I didn't think that wanting to use an SSD to store photos was a terrible idea as previous MLC and TLC SSD's I used for this purpose handled this fine.

Yes in retrospect I should have done more research on the drives, however would you really expect an SSD to be slower than a hard disk? Some things you just take for granted, if you bought a car you wouldn't expect it only do 30 on the motorway would you?

I still maintain that using SLC cache on QLC drive is just stupid though as using MLC would double the size of the cache and prevent many of the performance issues I encountered, at 50% capacity you only have 128GB of cache which is less than the 256GB memory cards I am using in my camera. Using SLC cache is no doubt cheaper for Intel because it's already a well established technique, rather than having to design new chips to handle MLC/TLC cache, however it's an inferior design.

The reason I say I wouldn't consider it for any purpose is because the price difference between QLC and TLC drives is negligible and the durability is significantly worse which is bad for any use case.

On the topic of hardware encryption...

Many SSDs Don’t Implement Encryption Properly
Even if you enable BitLocker encryption on a system, Windows 10 may not actually be encrypting your data. Instead, Windows 10 may be relying on your SSD to do it, and your SSD’s encryption may be easily broken.

That’s the conclusion from a new paper by researchers at Radbound University. They reverse engineered the firmwares of many solid-state drives and found a variety of issues with the “hardware encryption” found in many SSDs.

The researchers tested drives from Crucial and Samsung, but we definitely wouldn’t be surprised if other manufacturers had major issues. Even if you don’t have any of these specific drives, you should be concerned.

For example, the Crucial MX300 includes an empty master password by default. Yes, that’s right—it has a master password set to nothing, and that empty password gives access to the encryption key that encrypts your files. That’s crazy.

Many consumer SSDs claim to support encryption and BitLocker believed them. But, as we learned last year, those drives often weren’t securely encrypting files. Microsoft just changed Windows 10 to stop trusting those sketchy SSDs and default to software encryption.

In summary, solid-state drives and other hard drives can claim to be “self-encrypting.” If they do, BitLocker wouldn’t perform any encryption, even if you enabled BitLocker manually. In theory, that was good: The drive could perform the encryption itself at the firmware level, speeding up the process, reducing CPU usage, and maybe saving some power. In reality, it was bad: Many drives had empty master passwords and other horrendous security failures. We learned consumer SSDs can’t be trusted to implement encryption.



Now, Microsoft has changed things. By default, BitLocker will ignore drives that claim to be self-encrypting and do the encryption work in software. Even if you have a drive that claims to support encryption, BitLocker won’t believe it.

https://www.howtogeek.com/442114/windows-10s-bitlocker-encryption-no-longer-trusts-your-ssd/

So yeah lol, this is sort of reason why I use software encryption.
 
Last edited:
Is it really so difficult to admit you did little/no research and just assumed SSD=fast? It’s been pretty obvious since you said you were aware of QLC’s limitations and then told us you did a full write. Also that ‘negligible’ is roughly 50%, obviously it’s all relative, but that is a pretty big price difference.

The encryption point is interesting, from a technical perspective firmware is lower level than software, this gives it a technical advantage and like for like makes it more secure, the point is moot if the implementation is flawed though, just as it would be with software. Software is often easier/more likely to be updated by an end user vs hardware.
 
I'm a bit curious about this pricing issue, because I hadn't realised there were a couple of cheap non-QLC brands (the two I've seen in 30 seconds of searching): Sabrent and Silicon Power. It looks like they are priced at ~£250 for 2TB NVMe, whereas every well-known brand is £400+ (for TLC). I'm more than a little bit suspicious of how they achieve these low prices? Has anyone come across deep dive reviews into Sabrent and Silicon Power NVMe 2TB drives?
 
I'd like to add that I have no problem with QLC as long as the warranty is nice and long. The amount of cache they put in front of them should negate any real world performance issues. Just obviously if it's only ~20% more to get all TLC then it seems less attractive, but curious about the brands that have the cheap TLC.
 
I'm a bit curious about this pricing issue, because I hadn't realised there were a couple of cheap non-QLC brands (the two I've seen in 30 seconds of searching): Sabrent and Silicon Power. It looks like they are priced at ~£250 for 2TB NVMe, whereas every well-known brand is £400+ (for TLC). I'm more than a little bit suspicious of how they achieve these low prices? Has anyone come across deep dive reviews into Sabrent and Silicon Power NVMe 2TB drives?

Sabrent is Toshiba NAND with a Phison controller like a lot of the branded drives. I think it's a safe bet.
 
Back
Top Bottom