Crucial BX500 1TB any good? (SATA)

[Nicnac]

I haven't followed SSD progress recently and I don't know much about the BX series. I've liked crucial products before and just want a decent game drive for my steam library. Can anyone recommend these? Or do you guys have any other recommendations?

[Spakes]

Well, for starters, what is your budget? Also, BX500 has nice NAND chips, but it's cacheless, which is not good...

[kokosnh]

I would go with the cheapest from these

 

Crucial MX500

Kingston KC600

WD blue 3D

samsung 860 evo

Plextor M8VC

micron 1300 

 

but well, depends on the difference in prices vs BX500.

 

[porina]

I have a BX500 1TB as a game drive. It was enough cheaper than the slightly higher performing MX500, but I didn't think that was important just to install a bunch of games.

[Nicnac]

1 hour ago, Spakes said:

Well, for starters, what is your budget? Also, BX500 has nice NAND chips, but it's cacheless, which is not good...

does that impact gaming performance though? I don't think chache is too important for that and I don't need the fastest ssd ou there... I would like to get the cheapest thing that does the job. Anything is better than my dying hdds anyways

[Nicnac]

37 minutes ago, kokosnh said:

I would go with the cheapest from these

 

Crucial MX500

Kingston KC600

WD blue 3D

samsung 860 evo

Plextor M8VC

micron 1300 

 

but well, depends on the difference in prices vs BX500.

 

the bx beats all those by far in terms of price tho

[kokosnh]

how much? it all depends on how much more... Where do you buy? Amazon? I see BX500 and MX500 for 99$ 

[NewMaxx]

The 1TB BX500 is QLC-based (the old 960GB SKU is TLC) which reduces its relative value a bit. Worth keeping in mind. It's still fine for games.

[Nicnac]

40 minutes ago, kokosnh said:

how much? it all depends on how much more... Where do you buy? Amazon? I see BX500 and MX500 for 99$ 

95 euros at mindfactory. all other brands are above 100

[Nicnac]

18 minutes ago, NewMaxx said:

The 1TB BX500 is QLC-based (the old 960GB SKU is TLC) which reduces its relative value a bit. Worth keeping in mind. It's still fine for games.

i will probably use that thing until its not worth selling it anymore

[kokosnh]

the best i see in the mindfactory.de is WD blue 3D 1TB for 107.90 so 10 euro more then BX500.

 

For 1TB SSD i would pay 10 more for the better one, but that's just me. 
 

[Nicnac]

2 hours ago, kokosnh said:

the best i see in the mindfactory.de is WD blue 3D 1TB for 107.90 so 10 euro more then BX500.

 

For 1TB SSD i would pay 10 more for the better one, but that's just me. 
 

Yea I confused it the box was an amazon deal at 90. Mindfactory has a shitty sandisk for 95 and the Mx for 99 sol I’ll probably go with the Mx then as you said

[Spakes]

Whatever you do, stay away from WD SSDs... Unlike their HDDs, those are terrible af... Better go for 860 evo.

On 1/28/2020 at 10:51 PM, Nicnac said:

does that impact gaming performance though?

It does impact loading times and how much free RAM you have (since DRAM-less SSDs use PC RAM for cache purposes).

[Zando_]

4 minutes ago, Spakes said:

Whatever you do, stay away from WD SSDs... Unlike their HDDs, those are terrible af... Better go for 860 evo.

Or a Crucial MX500 which performs about the same but is often cheaper. Same as the MX300 vs the 850 Evo (both of which I've had, same as the MX500 and 860 Evo). 

Also what's so bad about WD SSDs? My WD Black NVMe was excellent, my WD Blue SATA M.2 has given me 0 issues. 

[Spakes]

2 minutes ago, Zando Bob said:

Also what's so bad about WD SSDs? My WD Black NVMe was excellent, my WD Blue SATA M.2 has given me 0 issues. 

They die all of a sudden (which is way too unexpecting for an SSD from a reputable brand), and their Blue line up with 3D NAND? Doesn't even reach its stated resource before dying. WD greens are ain't that bad, but they die suddently, too... Can't say anything about blacks, though... Source: https://3dnews.ru/938764/page-3.html#Western%20Digital%20Blue%20SSD

[IcarusLSC]

MX500 1TB is dirt cheap in many stores right now fwiw...

 

[NewMaxx]

1 hour ago, Spakes said:

their Blue line up with 3D NAND? Doesn't even reach its stated resource before dying

Mind you I know your source and quite a lot about the source material, so allow me to illuminate:

1. WD's flash, or SanDisk's flash - same company - is basically the same as BiCS3 (64L Toshiba TLC) on these drives. It's definitely the worst of the big three or four: Samsung > IMFT (Intel or Micron) > Toshiba/WD/SanDisk, that is. Although the flash on their NVMe drives is superior to Toshiba.
2. The methodology used by the site you linked is inherently flawed. I do love reading through their results, but it's not realistic for consumer usage. A large reason for this is SLC caching which often has an additive effect to NAND wear, especially with pure writes on dynamic SLC - this is why the SU800, a drive with legendarily high TBW, fails so quickly on their testbed: it's full-drive SLC caching. However in reality, consumer usage would make the most of that SLC. You do know the tested SU800 has 20% overprovisioning? (288GiB of flash for 238.4GiB user space)  And you believe its failure before TBW is realistic?
3. The WD Blue 3D (or SanDisk Ultra 3D) has a robust controller in the Marvell 88SS1074. It's used in many client and enterprise drives, even.

So I respectably disagree with your analysis. That being said, I would take the MX500 or one of its clones over the Blue 3D.

[Delicious Cake]

The memory controller is more likely to get bricked before TLC NAND even approaches its wear limit in most cases.

[Spakes]

17 hours ago, NewMaxx said:

The methodology used by the site you linked is inherently flawed.

Exactly how flawed are their methods? I'm not sure you captured every detail of their methodology while translating to your language.

[NewMaxx]

3 hours ago, Spakes said:

Exactly how flawed are their methods? I'm not sure you captured every detail of their methodology while translating to your language.

I think their statement on the SU800 (which I mentioned) is pretty clear: "Therefore, it is possible that the SLC caching scheme chosen by ADATA engineers caused the Ultimate SU800 to have such endurance." No kidding. In general use such a large SLC cache with massive overprovisioning - 20% - would last a lifetime for a consumer. The drive has its deficiencies as the flash was Micron's 32L/384Gb at that time - it's since moved to Samsung 64L - but that flash is rated at 1500 P/E and they always underrate endurance. Which they explain with: "So, most likely, the explanation is that ADATA buys second-rate memory chips from Micron, and the SMI SM2258 controller does not know how to properly handle them." Amazing, something I stated on Reddit many times years ago, clearly they do their research...since I'm the one who stated that. I also stated that the updated firmware for the SU800 has largely fixed these issues (which we know because other, newer drives sharing the hardware, and there are several, are quite improved). But in either case, using this kind of data to support the assertion that "they (WD SSDs) die all of a sudden" is faulty to say the least. (also worth reading is their SU900 analysis - since the 32L/256Gb MLC used on that drive is actually the same as the SU800's, the latter is in 3-bit/cell mode; but of course no SLC caching for MLC)

[Spakes]

3 hours ago, NewMaxx said:

I think their statement on the SU800 (which I mentioned) is pretty clear: "Therefore, it is possible that the SLC caching scheme chosen by ADATA engineers caused the Ultimate SU800 to have such endurance." No kidding. In general use such a large SLC cache with massive overprovisioning - 20% - would last a lifetime for a consumer. The drive has its deficiencies as the flash was Micron's 32L/384Gb at that time - it's since moved to Samsung 64L - but that flash is rated at 1500 P/E and they always underrate endurance. Which they explain with: "So, most likely, the explanation is that ADATA buys second-rate memory chips from Micron, and the SMI SM2258 controller does not know how to properly handle them." Amazing, something I stated on Reddit many times years ago, clearly they do their research...since I'm the one who stated that. I also stated that the updated firmware for the SU800 has largely fixed these issues (which we know because other, newer drives sharing the hardware, and there are several, are quite improved). But in either case, using this kind of data to support the assertion that "they (WD SSDs) die all of a sudden" is faulty to say the least. (also worth reading is their SU900 analysis - since the 32L/256Gb MLC used on that drive is actually the same as the SU800's, the latter is in 3-bit/cell mode; but of course no SLC caching for MLC)

Okay, I see your point about SU800s, maybe I'll suggest retesting newer model to them (with 64L). But in any way, sudden deaths that are hard to trace aren't welcome in any way (which was my primary point). That is what I call "all of a sudden". That is, of course, if they live through guaranteed amount of data rewritten onto them.

[NewMaxx]

The new SU800s are using 64L Samsung TLC, which is also being seen on many other drives like the L5 Lite 3D. It's rated for 1000 P/E (I have a Reddit thread exploring it) which is actually less than the Micron 32L/384Gb flash. Some of these drives have used 32L/256Gb TLC from IMFT which I point out because that's native TLC while the 32L/384Gb is the 32L/256Gb MLC in 3-bit mode (native is superior) - this is important because the SU900 tests far better despite having basically the same flash; on the SU800 it's in 3-bit mode, large SLC, while on SU900 it's 2-bit native and no SLC (MLC). In other words, there should not be so drastic a difference given the 20% OP on the SU800 with that flash, so their testing is causing the SLC to be additive to NAND wear. Micron has a whitepaper on this (linked on my subreddit, or will be soon) that basically points to that being 2x additive to WAF which would bring it in line with my expectations.

 

Whew, anyway, many of these generic or "Chinese" drives also used 32L or 48L Samsung which where forerunners to the current 64L. I've also seen some with Hynix's new 72L flash now - this flash is denser (4-plane, 128Gb/plane, ergo 512Gb) but has not been tested to my knowledge. However relevant here to the SU800: the Samsung is second-tier flash, not high quality, in my testing it is roughly equivalent to the original 32L/384Gb but of course doesn't require the heavy OP. Nevertheless it's been paired on drives with higher-than-normal failure rates which does suggest the SM2258 has issues with some flash, although firmware revisions generally match the flash. The easiest drive to check with the Hynix is the Gold S31, albeit with a unique controller.

 

To get back to the WD drives: the Blue 3D uses effectively Toshiba which, in my opinion, is the worst of the main manufacturers (as I said above), but I don't believe it's bad enough to cause unexpected failures. The WD drives tend to have small, static caches which, if anything, should improve reliability, but I'll have to confirm that. The controller is the Marvell 88SS1074 (as I mentioned) which is utilized in client drives like the Micron 1100 and 1300, the 5100 series, and is modified for other enterprise drives, so I think it's reliable there also. So in my opinion it should not have a particularly high failure rate beyond the bathtub curve you get with any storage device.