M.2 Failed ...

[Gazza_P]

So I wanted to ask some advice....

I bought a Rogueware M.2 NX200M December 2022. After alot of fault finding because the PC was freaking the fuck out i determined that it was the drive... I RMA'D it through takealot and the same week I got a new one delivered, 27 November 2023.

Everything was perfect, started having a boot issue where I had to boot to bios and then exit to get into windows but didn't bother me too much I just worked around it. Today I'm playing games and the whole system just freezes and I reboot only to get stuck on the Auros boot logo ... do some digging, reddit had a similar post saying you just spam space and it goes into bios, I did that and I see the M.2 doesn't show on the list of drives, so I remove drive by drive and test. The moment I remove the new drive it posts super fast no issue. Put the drive I'm my wife's rig and same issue, super slow won't go past bios etc.

So basically aren't these drives like decent quality, its only been a few days and to me it looks dead. And also what should I get instead of the NX200M given that the drive is only about R600.

 

[Deviation]

@Ojo can tell you, but Rogueware should be avoided at all cost.
Failure rate is apparently superhigh on all of their products, not only on their storage but monitors as well.

 

[Gazza_P]

@Ojo can tell you, but Rogueware should be avoided at all cost.
Failure rate is apparently superhigh on all of their products, not only on their storage but monitors as well.

What would you recommend as a replacement and is a m.2 even nessesary? Won't a normal SSD be fine?

 

[goldfritter]

What would you recommend as a replacement and is a m.2 even nessesary? Won't a normal SSD be fine?

Yes, unless you are writing very large files frequently enough that the higher speed will make a big difference.

 

[Gazza_P]

Yes, unless you are writing very large files frequently enough that the higher speed will make a big difference.

I see this one on special

Check this out on takealot: Samsung 980 500GB NVMe M.2 SSD

Takealot

www.takealot.com

 

[Oj0]

@Ojo can tell you, but Rogueware should be avoided at all cost.
Failure rate is apparently superhigh on all of their products, not only on their storage but monitors as well.

What would you recommend as a replacement and is a m.2 even nessesary? Won't a normal SSD be fine?

I would recommend writing 1s and 0s on a wax tablet before I'd recommend Rogueware.

Yes, unless you are writing very large files frequently enough that the higher speed will make a big difference.

That's not the be-all and end-all. The massive IOPS can be noticable. I've had more than one person including the boss-man upgrade from an existing NVMe to a Kingston KC3000 and the difference in responsiveness was both perceivable and measurable - that difference will get larger when going from SATA to NVMe.

 

[yourang]

The best M2 drives so far on budget is those mini ones people pull from dell laptops etc, no issue and I get 3000mbs(newer ones) read on them. Pity they super small but you can buy adpater.

 

[Deviation]

What would you recommend as a replacement and is a m.2 even nessesary? Won't a normal SSD be fine?

At PCIe Gen 3, nvme is about 3500/3000MB/s read/write while SSD is about 500/450MB/s.
Prices are roughly the same at Gen 3, so go for nvme. Wootware has some good deals now and then. Search results for: 'nvme' - Wootware

 

[goldfritter]

I would recommend writing 1s and 0s on a wax tablet before I'd recommend Rogueware.


That's not the be-all and end-all. The massive IOPS can be noticable. I've had more than one person including the boss-man upgrade from an existing NVMe to a Kingston KC3000 and the difference in responsiveness was both perceivable and measurable - that difference will get larger when going from SATA to NVMe.

I've seen a handful of blind tests online. Seems confirmation bias is quite a thing. I have personally used various SATA/M.2 SATA/DRAMless NVMe/fast NVMe drives and if you didn't tell me which was which, I'd never have known the difference.

The one time I did notice a difference was with heavy writes, where a faster drive could reduce waiting times by as much as 20 minutes on some workloads.

 

[SoF_ZA]

The best M2 drives so far on budget is those mini ones people pull from dell laptops etc, no issue and I get 3000mbs(newer ones) read on them. Pity they super small but you can buy adpater.

Or, get a free adapter?

 

[yourang]

Or, get a free adapter?

Yup those work also, My friend busy 3d printing these adapters.

 

[Oj0]

I've seen a handful of blind tests online. Seems confirmation bias is quite a thing. I have personally used various SATA/M.2 SATA/DRAMless NVMe/fast NVMe drives and if you didn't tell me which was which, I'd never have known the difference.

The one time I did notice a difference was with heavy writes, where a faster drive could reduce waiting times by as much as 20 minutes on some workloads.

It depends on the use-case. If storage isn't the bottleneck, it's possible to get the same performance out of a few mechanical drives in RAID 0 as you would out of a Gen5 NVMe. I'm going to cherry pick these as not everything will show the same difference, but check this - all are slowest 2.5" SATA vs slowest NVMe vs Kingston KC3000. This is to show not only the difference between 2.5" SATA and a fast NVMe drive, but also between a slow NVMe and fast NVMe drive:

Windows 10 Pro Startup Time:
Crucial BX500 2.5" SATA 480GB: 21.6 seconds
ADATA Falcon NVMe 1TB: 9.5 seconds
Kingston KC3000 NVMe 2TB: 7.4 seconds

VMWare (refresh virtual machine):
Samsung 870 QVO 2.5" SATA 1TB: 136.3 seconds
Kingston A2000 NVMe 1TB: 29.4 seconds
Kingston KC3000 NVMe 2TB: 14.6 seconds

Steam Preload (decrypt and install 20GB game):
Crucial BX500 2.5" SATA 480GB: 620.7 seconds
Lexar NM620 NVMe 1TB: 270.6 seconds
Kingston KC3000 NVMe 2TB: 193.1 seconds

MP3 Indexing (1,000 songs in WinAmp):
Samsung 870 QVO 1TB: 1.7 seconds
ADATA SX6000 Pro 1TB: 1.6 seconds
Kingston KC3000 NVMe 2TB: 0.9 seconds

Extrapolate that to a more reasonable music collection, let's say 20,000 songs, and it's 34 seconds vs 32 seconds vs 18 seconds.

Search Documents (find text within files):
Crucial BX500 2.5" SATA 480GB: 29.2 seconds
Corsair MP400 NVMe 2TB: 25 seconds
Kingston KC3000 NVMe 2TB: 15.6 seconds

Virus Scan (Avast scanning all files within System32):
Crucial BX500 2.5" SATA 480GB: 57.3 seconds
ADATA SX6000 Pro NVMe 1TB: 53.4 seconds
Kingston KC3000 NVMe 2TB: 41.1 seconds

Apple iTunes Installation:
Crucial BX500 2.5" SATA 480GB: 48.7 seconds
ADATA Falcon NVMe 1TB: 34.2 seconds
Kingston KC3000 NVMe 2TB: 32.9 seconds

Google Chrome Installation:
Crucial BX500 2.5" SATA 480GB: 34.9 seconds
HP EX900 Pro NVMe 1TB: 15.4 seconds
Kingston KC3000 NVMe 2TB: 13.9 seconds

Adobe Acrobat Installation:
Crucial BX500 2.5" SATA 480GB: 38.6 seconds
Lexar NM620 NVMe 1TB: 26.6 seconds
Kingston KC3000 NVMe 2TB: 23.7 seconds

Photoshop Editing (crop, move, auto levels, resize, save on 10x 50MP files consecutively):
Crucial BX500 2.5" SATA 480GB: 222.1 seconds
HP EX900 Pro NVMe 1TB: 60.5 seconds
Kingston KC3000 NVMe 2TB: 43.7 seconds

Doom Eternal Load Time:
Crucial BX500 2.5" SATA 480GB: 7.2 seconds
ADATA SX6000 Pro 1TB: 5.1 seconds
Kingston KC3000 NVMe 2TB: 4.9 seconds

Doom is the only game that shows a good percentage difference, but pretty much everyone does more than just game on their PC (otherwise they'd have a console for a cheaper gaming experience, right?). It's a lot of data up there, so I've normalized the 2.5" SATA SSD and this is the relative time for all tests:

Many of those differences are not only measurable, but very easily perceivable - enough that I'd reckon 99% of people could pass a double blind test. All results taken from:

Kingston KC3000 2 TB Review - Faster Than Samsung 980 Pro

The Kingston KC3000 is built using the Phison E18 controller and Micron's best 176-layer TLC NAND flash. In our performance testing, the drive can beat the Samsung 980 Pro and is the fastest SSD we ever tested. It shares that performance throne with the WD Black SN850.

www.techpowerup.com

 

[Oj0]

One more, relative performance between fast and slow NVMe drives:

 

[goldfritter]

It depends on the use-case. If storage isn't the bottleneck, it's possible to get the same performance out of a few mechanical drives in RAID 0 as you would out of a Gen5 NVMe. I'm going to cherry pick these as not everything will show the same difference, but check this - all are slowest 2.5" SATA vs slowest NVMe vs Kingston KC3000. This is to show not only the difference between 2.5" SATA and a fast NVMe drive, but also between a slow NVMe and fast NVMe drive:

Windows 10 Pro Startup Time:
Crucial BX500 2.5" SATA 480GB: 21.6 seconds
ADATA Falcon NVMe 1TB: 9.5 seconds
Kingston KC3000 NVMe 2TB: 7.4 seconds

VMWare (refresh virtual machine):
Samsung 870 QVO 2.5" SATA 1TB: 136.3 seconds
Kingston A2000 NVMe 1TB: 29.4 seconds
Kingston KC3000 NVMe 2TB: 14.6 seconds

Steam Preload (decrypt and install 20GB game):
Crucial BX500 2.5" SATA 480GB: 620.7 seconds
Lexar NM620 NVMe 1TB: 270.6 seconds
Kingston KC3000 NVMe 2TB: 193.1 seconds

MP3 Indexing (1,000 songs in WinAmp):
Samsung 870 QVO 1TB: 1.7 seconds
ADATA SX6000 Pro 1TB: 1.6 seconds
Kingston KC3000 NVMe 2TB: 0.9 seconds

Extrapolate that to a more reasonable music collection, let's say 20,000 songs, and it's 34 seconds vs 32 seconds vs 18 seconds.

Search Documents (find text within files):
Crucial BX500 2.5" SATA 480GB: 29.2 seconds
Corsair MP400 NVMe 2TB: 25 seconds
Kingston KC3000 NVMe 2TB: 15.6 seconds

Virus Scan (Avast scanning all files within System32):
Crucial BX500 2.5" SATA 480GB: 57.3 seconds
ADATA SX6000 Pro NVMe 1TB: 53.4 seconds
Kingston KC3000 NVMe 2TB: 41.1 seconds

Apple iTunes Installation:
Crucial BX500 2.5" SATA 480GB: 48.7 seconds
ADATA Falcon NVMe 1TB: 34.2 seconds
Kingston KC3000 NVMe 2TB: 32.9 seconds

Google Chrome Installation:
Crucial BX500 2.5" SATA 480GB: 34.9 seconds
HP EX900 Pro NVMe 1TB: 15.4 seconds
Kingston KC3000 NVMe 2TB: 13.9 seconds

Adobe Acrobat Installation:
Crucial BX500 2.5" SATA 480GB: 38.6 seconds
Lexar NM620 NVMe 1TB: 26.6 seconds
Kingston KC3000 NVMe 2TB: 23.7 seconds

Photoshop Editing (crop, move, auto levels, resize, save on 10x 50MP files consecutively):
Crucial BX500 2.5" SATA 480GB: 222.1 seconds
HP EX900 Pro NVMe 1TB: 60.5 seconds
Kingston KC3000 NVMe 2TB: 43.7 seconds

Doom Eternal Load Time:
Crucial BX500 2.5" SATA 480GB: 7.2 seconds
ADATA SX6000 Pro 1TB: 5.1 seconds
Kingston KC3000 NVMe 2TB: 4.9 seconds

Doom is the only game that shows a good percentage difference, but pretty much everyone does more than just game on their PC (otherwise they'd have a console for a cheaper gaming experience, right?). It's a lot of data up there, so I've normalized the 2.5" SATA SSD and this is the relative time for all tests:

Many of those differences are not only measurable, but very easily perceivable - enough that I'd reckon 99% of people could pass a double blind test. All results taken from:

Kingston KC3000 2 TB Review - Faster Than Samsung 980 Pro

The Kingston KC3000 is built using the Phison E18 controller and Micron's best 176-layer TLC NAND flash. In our performance testing, the drive can beat the Samsung 980 Pro and is the fastest SSD we ever tested. It shares that performance throne with the WD Black SN850.

www.techpowerup.com

Many are perceivable, not denying that. If you're running VMs and indexing files on a daily basis then sure, NVMe is totally justified. There are certainly power users for whom this would be helpful.

But while most people aren't just gaming on their PC, they also aren't reinstalling chrome or cropping and moving 50 photos or decrypting games every single day.

One task that isn't mentioned is the general "snappiness" of browsing Windows, websites, etc. which I'd argue plays one of the largest roles in a person's experience of the feeling of speed of a system. This is of course difficult to benchmark - which is where a blind test comes in - where I'd guess most people wouldn't be able to tell the difference on a blind test.

Of the list of tasks benchmarked, even people who don't just game, the only ones that are generally being done daily, are booting into Windows and loading a game. So going from the slowest drive to the fastest only saves you 16.5 seconds per day. And then maybe on the odd occasion, when you do need to reinstall acrobat, you save another 15 seconds. Which, sure, might be noticeable - but I guess it's up to OP to determine whether the extra cost justifies that time saving.

 

[NI∀ƆOƆ‾⊥ᴚ∩ʞ]

snip

Where stock?

Kingston KC3000 PCIe Gen4x4 M.2 2280 NVMe SSD — 2TB — Best Deals at Progenix — South Africa

Kingston KC3000 PCIe 4.0 NVMe M.2 SSD delivers next-level performance using the latest Gen 4x4 NVMe controller and 3D TLC NAND. Upgrade the storage and reliability of your system to keep up with demanding workloads and experience better performance with s

progenix.co.za

 

[Oj0]

But while most people aren't just gaming on their PC, they also aren't reinstalling chrome or cropping and moving 50 photos or decrypting games every single day.

And when said apps want to update? Skype wants to update almost daily lately, and Steam every other day. Add to that all the other stuff that runs in the backgrounds, as well as those that can certainly be measured but not consistently benchmarked (such as Windows Update) and the overall experience on a high speed NVMe drive is FAR superior to that of a 2.5" SATA drive (or even a slow NVMe drive).

One task that isn't mentioned is the general "snappiness" of browsing Windows, websites, etc. which I'd argue plays one of the largest roles in a person's experience of the feeling of speed of a system. This is of course difficult to benchmark - which is where a blind test comes in - where I'd guess most people wouldn't be able to tell the difference on a blind test.

I reckon many people will pass the blind test. It's not often that I do builds with a 2.5" SATA SSD, but the system always feels sluggish compared to a system with an NVMe drive. In the beginning I couldn't figure it out as I was in the same camp as you, believing that unless you're copying hundreds of gigs between drives regularly you wouldn't be able to tell the difference. It took me a while (years ) to realize that when it comes to an SSD, IOPS is almost everything.

Of the list of tasks benchmarked, even people who don't just game, the only ones that are generally being done daily, are booting into Windows and loading a game. So going from the slowest drive to the fastest only saves you 16.5 seconds per day. And then maybe on the odd occasion, when you do need to reinstall acrobat, you save another 15 seconds.

See first paragraph of this post.

Which, sure, might be noticeable - but I guess it's up to OP to determine whether the extra cost justifies that time saving.

That's the thing. The "premium" for NVMe has fallen away. In fact, NVMe is often cheaper than 2.5" SATA. Using my pricing,

	Cheapest 2.5"	Cheapest NVMe	Premium over 2.5"
240-256GB	R 388	R 423*	R 35
480-512GB	R 642	R 666	R 24
960GB-1TB	R 973	R 1,134	R 161
2TB	R 2,299	R 2,088	- R 211
4TB	R 5,580	R 5,614	R 34

* there's an OEM 256GB NVMe drive for R 387, making it cheaper than 2.5" SATA, but because it's OEM with only a 1-year warranty I left it off.

Unless you're out of M.2 slots, there's basically NO REASON to go 2.5" anymore.

 

[Oj0]

Where stock?

Kingston KC3000 PCIe Gen4x4 M.2 2280 NVMe SSD — 2TB — Best Deals at Progenix — South Africa

Kingston KC3000 PCIe 4.0 NVMe M.2 SSD delivers next-level performance using the latest Gen 4x4 NVMe controller and 3D TLC NAND. Upgrade the storage and reliability of your system to keep up with demanding workloads and experience better performance with s

progenix.co.za

Same place as your wife's pleasure in life - gone.

It'll be restocked in the new year

 

[goldfritter]

And when said apps want to update? Skype wants to update almost daily lately, and Steam every other day. Add to that all the other stuff that runs in the backgrounds, as well as those that can certainly be measured but not consistently benchmarked (such as Windows Update) and the overall experience on a high speed NVMe drive is FAR superior to that of a 2.5" SATA drive (or even a slow NVMe drive).

Cool, so maybe another 20 seconds a day? We're up to a whopping 40 seconds a day saved.

I reckon many people will pass the blind test. It's not often that I do builds with a 2.5" SATA SSD, but the system always feels sluggish compared to a system with an NVMe drive. In the beginning I couldn't figure it out as I was in the same camp as you, believing that unless you're copying hundreds of gigs between drives regularly you wouldn't be able to tell the difference. It took me a while (years ) to realize that when it comes to an SSD, IOPS is almost everything.

You're a power user... You build and run systems on the daily. A high-speed NVMe is completely justified for you.

That's the thing. The "premium" for NVMe has fallen away. In fact, NVMe is often cheaper than 2.5" SATA. Using my pricing,

	Cheapest 2.5"	Cheapest NVMe	Premium over 2.5"
240-256GB	R 388	R 423*	R 35
480-512GB	R 642	R 666	R 24
960GB-1TB	R 973	R 1,134	R 161
2TB	R 2,299	R 2,088	- R 211
4TB	R 5,580	R 5,614	R 34

* there's an OEM 256GB NVMe drive for R 387, making it cheaper than 2.5" SATA, but because it's OEM with only a 1-year warranty I left it off.

Unless you're out of M.2 slots, there's basically NO REASON to go 2.5" anymore.

This I totally agree with. No reason to go 2.5" unless you're out of NVMe slots.

 

[Oj0]

Cool, so maybe another 20 seconds a day? We're up to a whopping 40 seconds a day saved.

Per app* Most people have a good number of things running on the daily which update. Skype, Teams, Steam, Origin, Epic Games Launcher, Ubisoft's thing whatever it's called now, Plex, possibly a "Linux ISO downloader," GPU drivers, Discord, MSI Afterburner, iCUE, Team Viewer, etc etc etc. Pick a bad day where half the above want to update as well as Windows and you're talking about a large difference - not to mention all of those add to startup time (the Windows startup benchmark was done on a fresh Windows installation - more on that at the bottom).

People will pay hundreds to thousands of Rands to go from 120 to 121 FPS (eg picking a Gaming OC instead of a Gaming or TUF OC instead of TUF of the same GPU, or even Aorus Master instead of Eagle or Strix OC instead of TUF) which is a difference not only imperceptible, but so far within margin of error it's likely that given enough runs you may very well find some where the cheaper card comes out on top.

You're a power user... You build and run systems on the daily. A high-speed NVMe is completely justified for you.

Ironically I don't use an NVMe SSD xD But I can certainly tell the difference on a fresh install of Windows, never mind a dirty install with a dozen background apps. I've been able to tell the difference for YEARS, but always shrugged it off and tried to blame something else as I had the mentality that it's anything BUT the storage. It took a fair deal of convincing for me to accept that not only is NVMe vastly superior, but the difference between a slow and fast NVMe shows itself in far more than just copying large files around.

---

About reviews in general. Something that always irks me is that benchmarks are never done in a real-world scenario, but rather absolute best case. Who games (or whatever else) on a fresh install of Windows with nothing but the drivers and game (plus launcher if applicable) installed? Who formats and re-loads every time there's a new driver (ANY driver, not just GPU driver)?

Hell, some people won't even run the latest drivers because their preferred game is a bit faster on a driver one or two versions older.

A typical way to game would be on a dirty installation - even if it's only a few days old, it's got all your apps and utilities loaded, and people seldom close EVERY. OTHER. APP. to game. Very often, at the very least Chrome is running, maybe a mail client as well - I'm not even talking about background tasks such as Steam, Afterburner, etc which are ALSO running.

Even chronic Windows-reinstallers typically only reinstall every six months, and they too will start by installing all their apps and utilities so a best case review is really quite irrelevant for anything other than relative performance.

All of this does have an impact on performance. About a year ago I found a single video which I can't find now where a comparison with the same hardware was done on both a fresh installation and a typical installation, and the difference was quite noticeable - something like half the difference of going "one up" in the GPU SKU stack.

I would love reviews to be done with at least the common apps loaded - Teams, Discord, Afterburner, Steam, Epic Games Launcher, motherboard and RGB software, an anti-virus (pick something common - maybe Bitdefender or Malwarebytes or even just leave MS Defender enabled)... you get the idea.

/rant

 

[goldfritter]

Per app* Most people have a good number of things running on the daily which update. Skype, Teams, Steam, Origin, Epic Games Launcher, Ubisoft's thing whatever it's called now, Plex, possibly a "Linux ISO downloader," GPU drivers, Discord, MSI Afterburner, iCUE, Team Viewer, etc etc etc. Pick a bad day where half the above want to update as well as Windows and you're talking about a large difference - not to mention all of those add to startup time (the Windows startup benchmark was done on a fresh Windows installation - more on that at the bottom).

People will pay hundreds to thousands of Rands to go from 120 to 121 FPS (eg picking a Gaming OC instead of a Gaming or TUF OC instead of TUF of the same GPU, or even Aorus Master instead of Eagle or Strix OC instead of TUF) which is a difference not only imperceptible, but so far within margin of error it's likely that given enough runs you may very well find some where the cheaper card comes out on top.


Ironically I don't use an NVMe SSD xD But I can certainly tell the difference on a fresh install of Windows, never mind a dirty install with a dozen background apps. I've been able to tell the difference for YEARS, but always shrugged it off and tried to blame something else as I had the mentality that it's anything BUT the storage. It took a fair deal of convincing for me to accept that not only is NVMe vastly superior, but the difference between a slow and fast NVMe shows itself in far more than just copying large files around.

---

About reviews in general. Something that always irks me is that benchmarks are never done in a real-world scenario, but rather absolute best case. Who games (or whatever else) on a fresh install of Windows with nothing but the drivers and game (plus launcher if applicable) installed? Who formats and re-loads every time there's a new driver (ANY driver, not just GPU driver)?

Hell, some people won't even run the latest drivers because their preferred game is a bit faster on a driver one or two versions older.

A typical way to game would be on a dirty installation - even if it's only a few days old, it's got all your apps and utilities loaded, and people seldom close EVERY. OTHER. APP. to game. Very often, at the very least Chrome is running, maybe a mail client as well - I'm not even talking about background tasks such as Steam, Afterburner, etc which are ALSO running.

Even chronic Windows-reinstallers typically only reinstall every six months, and they too will start by installing all their apps and utilities so a best case review is really quite irrelevant for anything other than relative performance.

All of this does have an impact on performance. About a year ago I found a single video which I can't find now where a comparison with the same hardware was done on both a fresh installation and a typical installation, and the difference was quite noticeable - something like half the difference of going "one up" in the GPU SKU stack.

I would love reviews to be done with at least the common apps loaded - Teams, Discord, Afterburner, Steam, Epic Games Launcher, motherboard and RGB software, an anti-virus (pick something common - maybe Bitdefender or Malwarebytes or even just leave MS Defender enabled)... you get the idea.

/rant

Ok, I guess we'll have to agree to disagree here.

My position remains, for a general user, the performance difference from a faster drive to a slower drive won't make a daily difference big enough to justify going from a cheaper drive to a more expensive faster one (e.g. nearly R 2000 more from a an entry-level 1TB NVMe to a KC3000).

But each user is free to examine the data and make a judgment call based on their own use-case - "value" is a subjective term and making blanket statements about it has its pitfalls.

As a point about the going from a cheap GPU to an expensive one (and while we're on pet peeves) - there are reasons people buy the more expensive one despite little to no performance gain, other than just bragging rights. Often the more expensive GPU has a better (=quieter) cooler, and aesthetic considerations play a role as well.

 

[Oj0]

My position remains, for a general user, the performance difference from a faster drive to a slower drive won't make a daily difference big enough to justify going from a cheaper drive to a more expensive faster one (e.g. nearly R 2000 more from a an entry-level 1TB NVMe to a KC3000).

The cheapest 1TB NVMe is R 1,134, the KC3000 is R 1,617 - that's not R 2k more? It's R 483 more, which gets you far more than just a faster drive.

Using my stats from the last two years, and based on all capacities, the Legend 710 has a failure rate of 6% (since availability, which started in Feb this year) vs the KC3000 at 1.2% (since availability, which started June last year)
The Kingston has 1.6x the warranty duration
The Kingston has 3x the endurance rating

3-5 years down the line if/when the Legend fails, you'll be buying another. Assuming the price drops by 30% over that time (unlikely - thanks RoE), you're still paying R 1,928 for the original drive plus its replacement vs R 1,617 for a single KC3000.

Statistically, I think the chances of the cheaper drive dying in the 3-5 year period are high enough for it to be something to take into consideration, as the KC3000 has been available for 50% longer yet the Legend 710 already has 6x the failure rate.

You're also discounting the difference between 2.5" SATA performance and entry-level NVMe performance, which was the original point.

But each user is free to examine the data and make a judgment call based on their own use-case - "value" is a subjective term and making blanket statements about it has its pitfalls.

Value CAN be purely objective:

	Cheapest 2.5"	Cheapest NVMe	Difference
240-256GB	240GB @ R 1.62/GB	256GB @ R 1.65/GB	+ R 0.03/GB
480-512GB	512GB @ R 1.25/GB	512GB @ R 1.30/GB	+ R 0.05/GB
960GB-1TB	960GB @ R 1.01/GB	1TB @ R 1.11/GB	+ R 0.10/GB
2TB	R 1.12/GB	R 1.02/GB	- R 0.10/GB
4TB	R 1.36/GB	R 1.37/GB	+ R 0.01/GB

Averaged out, that's R 1.27/GB for 2.5" SATA and R 1.29/GB for NVMe, or + 1.5 %. Such a tiny difference makes a blanket statement easy to make. More on that in the closing paragraph.

As a point about the going from a cheap GPU to an expensive one (and while we're on pet peeves) - there are reasons people buy the more expensive one despite little to no performance gain, other than just bragging rights. Often the more expensive GPU has a better (=quieter) cooler, and aesthetic considerations play a role as well.

How about TUF to TUF OC? Gaming to Gaming OC? Eagle to Eagle OC? Gaming Trio to Gaming X Trio? Strix to Strix OC? Ventus 3X to Ventus 3X OC? Each pair is aesthetically identical with identical coolers and therefore and noise and cooling performance.

People generally go for the OC models because they want that extra performance even if the difference is so small it's impossible to consistently MEASURE. It's so bad that distis are often only bringing in the OC model as the non-OC only sells when the OC is out of stock.

The OC model is generally around 3% more expensive (sometimes more than 10% more expensive), but people will pay that for a difference which doesn't exist. That plays into the "paying +1.5% more is a reasonable blanket statement" when the difference isn't just consistently measurable, but perceivable.

This I totally agree with. No reason to go 2.5" unless you're out of NVMe slots.

Is that not the exact same blanket statement? I've just given the numbers in my posts so that others can see as well, rather than taking my/our word for it.

 

[TheJudge]

It depends on the use-case. If storage isn't the bottleneck, it's possible to get the same performance out of a few mechanical drives in RAID 0 as you would out of a Gen5 NVMe. I'm going to cherry pick these as not everything will show the same difference, but check this - all are slowest 2.5" SATA vs slowest NVMe vs Kingston KC3000. This is to show not only the difference between 2.5" SATA and a fast NVMe drive, but also between a slow NVMe and fast NVMe drive:

Windows 10 Pro Startup Time:
Crucial BX500 2.5" SATA 480GB: 21.6 seconds
ADATA Falcon NVMe 1TB: 9.5 seconds
Kingston KC3000 NVMe 2TB: 7.4 seconds

VMWare (refresh virtual machine):
Samsung 870 QVO 2.5" SATA 1TB: 136.3 seconds
Kingston A2000 NVMe 1TB: 29.4 seconds
Kingston KC3000 NVMe 2TB: 14.6 seconds

Steam Preload (decrypt and install 20GB game):
Crucial BX500 2.5" SATA 480GB: 620.7 seconds
Lexar NM620 NVMe 1TB: 270.6 seconds
Kingston KC3000 NVMe 2TB: 193.1 seconds

MP3 Indexing (1,000 songs in WinAmp):
Samsung 870 QVO 1TB: 1.7 seconds
ADATA SX6000 Pro 1TB: 1.6 seconds
Kingston KC3000 NVMe 2TB: 0.9 seconds

Extrapolate that to a more reasonable music collection, let's say 20,000 songs, and it's 34 seconds vs 32 seconds vs 18 seconds.

Search Documents (find text within files):
Crucial BX500 2.5" SATA 480GB: 29.2 seconds
Corsair MP400 NVMe 2TB: 25 seconds
Kingston KC3000 NVMe 2TB: 15.6 seconds

Virus Scan (Avast scanning all files within System32):
Crucial BX500 2.5" SATA 480GB: 57.3 seconds
ADATA SX6000 Pro NVMe 1TB: 53.4 seconds
Kingston KC3000 NVMe 2TB: 41.1 seconds

Apple iTunes Installation:
Crucial BX500 2.5" SATA 480GB: 48.7 seconds
ADATA Falcon NVMe 1TB: 34.2 seconds
Kingston KC3000 NVMe 2TB: 32.9 seconds

Google Chrome Installation:
Crucial BX500 2.5" SATA 480GB: 34.9 seconds
HP EX900 Pro NVMe 1TB: 15.4 seconds
Kingston KC3000 NVMe 2TB: 13.9 seconds

Adobe Acrobat Installation:
Crucial BX500 2.5" SATA 480GB: 38.6 seconds
Lexar NM620 NVMe 1TB: 26.6 seconds
Kingston KC3000 NVMe 2TB: 23.7 seconds

Photoshop Editing (crop, move, auto levels, resize, save on 10x 50MP files consecutively):
Crucial BX500 2.5" SATA 480GB: 222.1 seconds
HP EX900 Pro NVMe 1TB: 60.5 seconds
Kingston KC3000 NVMe 2TB: 43.7 seconds

Doom Eternal Load Time:
Crucial BX500 2.5" SATA 480GB: 7.2 seconds
ADATA SX6000 Pro 1TB: 5.1 seconds
Kingston KC3000 NVMe 2TB: 4.9 seconds

Doom is the only game that shows a good percentage difference, but pretty much everyone does more than just game on their PC (otherwise they'd have a console for a cheaper gaming experience, right?). It's a lot of data up there, so I've normalized the 2.5" SATA SSD and this is the relative time for all tests:

Many of those differences are not only measurable, but very easily perceivable - enough that I'd reckon 99% of people could pass a double blind test. All results taken from:

Kingston KC3000 2 TB Review - Faster Than Samsung 980 Pro

The Kingston KC3000 is built using the Phison E18 controller and Micron's best 176-layer TLC NAND flash. In our performance testing, the drive can beat the Samsung 980 Pro and is the fastest SSD we ever tested. It shares that performance throne with the WD Black SN850.

www.techpowerup.com

Thats the kakkest collection of Haikus I've ever read in my life!










Interesting time differences, I still have a SATA NVME drive.

 

[goldfritter]

The cheapest 1TB NVMe is R 1,134, the KC3000 is R 1,617 - that's not R 2k more? It's R 483 more, which gets you far more than just a faster drive.

Pologies, I was looking at the the 2TB.

Using my stats from the last two years, and based on all capacities, the Legend 710 has a failure rate of 6% (since availability, which started in Feb this year) vs the KC3000 at 1.2% (since availability, which started June last year)
The Kingston has 1.6x the warranty duration
The Kingston has 3x the endurance rating

3-5 years down the line if/when the Legend fails, you'll be buying another. Assuming the price drops by 30% over that time (unlikely - thanks RoE), you're still paying R 1,928 for the original drive plus its replacement vs R 1,617 for a single KC3000.

Statistically, I think the chances of the cheaper drive dying in the 3-5 year period are high enough for it to be something to take into consideration, as the KC3000 has been available for 50% longer yet the Legend 710 already has 6x the failure rate.

This is brand new info unrelated to the speed question. We were just talking about speed; I'm happy to concede that reliability factors in as well and could swing the value proposition back to the fancier drive.

You're also discounting the difference between 2.5" SATA performance and entry-level NVMe performance, which was the original point.

Yes, I will also agree that SATA doesn't offer a lot over 2.5" NVMe. That was the original point, but in your post #19 you reiterate the point that a fast NVMe is much better than a cheap NVMe, so I was responding to that point as well.

Value CAN be purely objective:

	Cheapest 2.5"	Cheapest NVMe	Difference
240-256GB	240GB @ R 1.62/GB	256GB @ R 1.65/GB	+ R 0.03/GB
480-512GB	512GB @ R 1.25/GB	512GB @ R 1.30/GB	+ R 0.05/GB
960GB-1TB	960GB @ R 1.01/GB	1TB @ R 1.11/GB	+ R 0.10/GB
2TB	R 1.12/GB	R 1.02/GB	- R 0.10/GB
4TB	R 1.36/GB	R 1.37/GB	+ R 0.01/GB

Averaged out, that's R 1.27/GB for 2.5" SATA and R 1.29/GB for NVMe, or + 1.5 %. Such a tiny difference makes a blanket statement easy to make. More on that in the closing paragraph.

Rand per gb isn't the only way to consider "value." The question of "how many extra rand is a time saving of a minute or two per day worth to you?" is much more subjective.

How about TUF to TUF OC? Gaming to Gaming OC? Eagle to Eagle OC? Gaming Trio to Gaming X Trio? Strix to Strix OC? Ventus 3X to Ventus 3X OC? Each pair is aesthetically identical with identical coolers and therefore and noise and cooling performance.

People generally go for the OC models because they want that extra performance even if the difference is so small it's impossible to consistently MEASURE. It's so bad that distis are often only bringing in the OC model as the non-OC only sells when the OC is out of stock.

The OC model is generally around 3% more expensive (sometimes more than 10% more expensive), but people will pay that for a difference which doesn't exist. That plays into the "paying +1.5% more is a reasonable blanket statement" when the difference isn't just consistently measurable, but perceivable.

Your original point suggested going from an Eagle to a Master and similar model jumps that would make quite a difference in aesthetics etc. I agree going from non-OC to OC won't make much of a difference.

Is that not the exact same blanket statement? I've just given the numbers in my posts so that others can see as well, rather than taking my/our word for it.

Yes, and as I said, blanket statements come with pitfalls. That applies to my own blanket statements as much as others' and I am happy to admit that.

 

[goldfritter]

Thats the kakkest collection of Haikus I've ever read in my life!










Interesting time differences, I still have a SATA NVME drive.

A drive can't be both
SATA and NVME
It's not possible.

 

[TheJudge]

A drive can't be both
SATA and NVME
It's not possible.

Much better.

OK, itty bitty SATA?

Fine, M.2 SATA. 😖

 

[Oj0]

Thats the kakkest collection of Haikus I've ever read in my life!

There are four lines per verse, therefore not a haiku. Fail!

Interesting time differences, I still have a SATA NVME drive.

With an floppy interface, right?

 

[Oj0]

Pologies, I was looking at the the 2TB.

Haha, no prob.

This is brand new info unrelated to the speed question. We were just talking about speed; I'm happy to concede that reliability factors in as well and could swing the value proposition back to the fancier drive.

My point was originally NVMe (any) vs SATA, so the OTHER factors between fast and slow NVMe weren't brought up. Even the second graph I made (post 13) was an afterthought as I was focusing on 2.5" SATA vs NVMe. The KC3000 was chucked in to show that there's also a difference between NVMe drives, and it was chosen due to its speed, reliability, and popularity (between all capacities, the KC3000 accounts for nearly 30% of SSD sales).

Yes, I will also agree that SATA doesn't offer a lot over 2.5" NVMe.

This sounds like @TheJudge with his SATA NVMe drive

That was the original point, but in your post #19 you reiterate the point that a fast NVMe is much better than a cheap NVMe, so I was responding to that point as well.

No, I didn't say "much better," I said:

but the difference between a slow and fast NVMe shows itself in far more than just copying large files around

Which is objectively correct - if there's a difference anywhere other than copying large files, my statement is correct. What I said was that NVMe (any) is much better than 2.5" SATA, which I've also shown.

Rand per gb isn't the only way to consider "value." The question of "how many extra rand is a time saving of a minute or two per day worth to you?" is much more subjective.

Well let's take a common size, 1TB. At an average price difference of R 0.02 per GB, the difference is R 200. Let's work on the very (unreasonably) low side and say the difference is 30 seconds per day, and let's stick to a short time-frame of just one year. That's a saving of 182.5 minutes per year, round that to 180. Do you earn more than R 67/hour? If so, would you accept R 67/hour? If not, you value your time at a minimum of R 67/hour. 180 minutes would therefore be worth at least R 200 to you, or the difference in saving.

If we scale that up to something more reasonable and realistic, such as a one minute saving per day (averaged over a year, this includes things such as the extra time for Windows updates spread across the full year) and say you'll use the drive for three years. The saving is now 1,095 minutes, or more than 18 hours. Your time now has to be valued at less than R 12/hour for it not be worth it to you.

Nobody earning less than minimum wage is looking at SSDs, so we can objectively say that everyone considering an SSD will save significantly more over a three year period than their time is worth.

Your original point suggested going from an Eagle to a Master and similar model jumps that would make quite a difference in aesthetics etc. I agree going from non-OC to OC won't make much of a difference.

I did mention the more similar cards:

(eg picking a Gaming OC instead of a Gaming or TUF OC instead of TUF of the same GPU, or even Aorus Master instead of Eagle or Strix OC instead of TUF)

Strix vs TUF, Gaming Trio vs Ventus 3X, Aorus Master vs Eagle, etc are all differences consisting of more than just performance, which is why I grouped them separately.

On a far larger scale (distributor), the OC version is taken often enough that the distis don't bother with non-OC versions, so it's established fact that people are willing to pay extra for a non-existent difference. A few years ago they'd bring in everything, and I'd watch the OC levels drop closer and closer to zero which the non-OC stays at or close to the original quantity.

This thereby enforces the blanket statement of 1.5% more for a real-world difference is a moot difference to pay (even more so when you consider the absolute Rand value, which varies between an extra R 161 and a SAVING of R 211 depending on the capacity).

 

[goldfritter]

This sounds like @TheJudge with his SATA NVMe drive

Fokof.



As to the rest, we're bickering about semantics more than anything else at this point. I think we're both on very close to the same page, subjective factors not considered.

FWIW when it was up to me I bought a KC2500, the faster Gen 3 drive, even though I didn't believe it would make a big difference to my experience. The reliability and the "price is near enough to be worth it" factors swayed it for me.

From you in fact, IIRC.

 

[Oj0]

As to the rest, we're bickering about semantics more than anything else at this point. I think we're both on very close to the same page, subjective factors not considered.

Yes we are, but it's passing time. I'm not used to not working and not quite sure what to do with myself

 

[goldfritter]

Yes we are, but it's passing time. I'm not used to not working and not quite sure what to do with myself

Go argue on MyBB, it's MUCH more spicy.

I checked their Ts and Cs and they apparently do NO content moderation at all with the exception of actually illegal stuff. Makes for some truly fruity debate.