Personal computing discussed

Moderators: morphine, Steel

 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Preliminary results for Crucial BX200 degraded read speeds

Tue Mar 22, 2016 8:13 am

Some of you might remember the thread I created earlier about the ADATA SP550 suffering from read speed degradation.
In which I said I would test several SSDs to find out if they suffered from this issue as well and let you know if that was the case.

I'm still running my tests, but for one of these SSDs it's read speeds dropped so quickly I thought I should let you know.
As you may have already guessed considering the title of this thread that SSD is a Crucial BX200.


At first I thought that the results were due to HD Tune Pro sometimes needing some time to get warmed up.
However after three weeks of testing I analyzed the results a bit more closely and noticed that it read speeds went back to normal about the same time as it reached the unused areas of the SSD.

And as I got more results it became more and more clear that it's read speeds were dropping and dropping faster than those of any SSD I'm currently testing.


Here you can see the results of the tests I've performed every week to measure read speeds.

Image

This is after one week and 20GB.
You can see that read speeds are back to normal around 20GB.


Image

This is after two weeks and 40GB.
Read speeds recover at about 40GB.


Image

This is after three weeks and 60GB.
Like with the others read speeds are fine after it reaches the unused portion of the SSD.
Here it is starting to become clear that read speeds are dropping thanks to that curve becoming a bit more uneven making it less likely that the earlier results were solely due to HD Tune Pro.


Image

This is after four weeks and 80GB.
You can see that where there was a drop in read speeds earlier it has recovered quite a bit but at the same time read speeds for the more newly written sections are dropping.


Image

This is after five weeks and 100GB.
Here you can see the biggest drop in read speeds so far which suggests that LDPC can raise read speeds but that the effect is temporary.


Image

This is after six weeks and 120GB.
Don't know I have to add anything else other than that the read speeds keep dropping.


Image

This is after seven weeks and 140GB.
Like before you can see that where before there was a particularly large drop in read speed it has recovered quite a bit at the same time read speeds for the more newly written sections are dropping.




I believe that the results from the first four weeks may be affected by HD Tune Pro right at the start of the read but even if I feel a bit skeptical of the really low read speeds right at the beginning you can still see that the read speeds have started to drop.
The results also show that they are dropping fast, with several examples of it dropping 50-100MB/s from one week to the next.

And even if LDPC ECC is better at adjusting for things like age and temperature than the more common BCH ECC and BX200 can use it to increase read speeds after I have read the entire SSD it doesn't prevent read speeds from dropping.


Perhaps you would like to know how I have tested the BX200 so far?
2016-01-29 I added one folder at 20GB filled with files at 20MB and after having left it unpowered for a week I tested its read speeds with HD Tune Pro 5.60 and added another folder at 20GB filled with files at 20MB.
Which has been repeated every week.


Other SSDs besides the BX200 may use the same NAND.
The Patriot Blast SSD does.
I suspect that for example the Corsair Force LE does.
And I have heard that ADATA intends to replace the SK Hynix NAND in the SP550 with the 16nm TLC NAND found in the BX200 if they haven't already.

While the ADATA SP550 is not particularly fast with the original SK Hynix NAND and can suffer from degraded read speeds in certain situations the NAND found in the BX200 is even slower and (according to these tests) significantly more prone to read speed degradation.
Let's put it this way: The BX200 has been the SSD for which read speeds have dropped the fastest of the SSDs I'm currently testing with the ADATA SP550 taking second place.
Read speeds for the ADATA SP550 have dropped by about 50MB/s over eight weeks of testing, while the BX200 has seen drops in read speeds over 100MB/s from one week to another.
And that's excluding the really low results because of the difficulty in determining if HD Tune Pro affected them.
Also note that both SSDs were tested in the same way and in the same conditions.


As I said earlier in the thread about the ADATA SP550 these results do not neccesarily mean that the BX200 has issues with low read speeds regardless of situation, most of which these tests do not cover.
That it is left without power for a week hardly affects its read speeds positively.
But the fact that I scan the entire SSD every week gives it the opportunity to identify and correct those areas with low read speeds which does affect read speeds positively.
Maybe it is just very sensitive to being left unpowered or that it needs time to go through the SSD when it is idling to fix this issue before it has gone as far as it already has.
But that hasn't been part of the tests I have performed (so far at least).
Last edited by Glaring_Mistake on Wed Mar 23, 2016 4:04 pm, edited 1 time in total.
 
Duct Tape Dude
Gerbil Elite
Posts: 718
Joined: Thu May 02, 2013 12:37 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 8:37 am

Interesting. Out of curiosity do your testing methods include using TRIM and waiting a few minutes for steady-state?
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 10:06 am

Duct Tape Dude wrote:
Interesting. Out of curiosity do your testing methods include using TRIM and waiting a few minutes for steady-state?


TRIM is always active, but is not forced before testing.
Nor do I test it in steady-state.

I start the read test after the OS has booted up properly in order to prevent the (pretty unlikely) event that it should resolve any issues before the test is run.
I also shut down unnecessary applications to minimize the risk of one of them affecting the results.
 
DPete27
Silver subscriber
Grand Gerbil Poohbah
Posts: 3652
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 3:45 pm

This just looks like the progression to steady-state performance to me. The performance drop happens in the same range as the amount of data that's currently on the drive. TRIM should recover most of that lost performance though (not all, you'd have to secure erase the drive to get back to "out-of-box" performance). Have you tried manually running TRIM from the Crucial Storage Executive sofware and retesting?
Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 5:11 pm

DPete27 wrote:
This just looks like the progression to steady-state performance to me. The performance drop happens in the same range as the amount of data that's currently on the drive. TRIM should recover most of that lost performance though (not all, you'd have to secure erase the drive to get back to "out-of-box" performance). Have you tried manually running TRIM from the Crucial Storage Executive sofware and retesting?


I'll try to remember to manually run TRIM when it's time to test it again to see if that has any effect.
You should get an update on that during the weekend.
 
Zoomastigophora
Gerbil Elite
Posts: 667
Joined: Tue Nov 11, 2008 7:10 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 6:20 pm

DPete27 wrote:
This just looks like the progression to steady-state performance to me. The performance drop happens in the same range as the amount of data that's currently on the drive. TRIM should recover most of that lost performance though (not all, you'd have to secure erase the drive to get back to "out-of-box" performance). Have you tried manually running TRIM from the Crucial Storage Executive sofware and retesting?

I thought TRIM only affects write speeds?
 
LostCat
Minister of Gerbil Affairs
Posts: 2018
Joined: Thu Aug 26, 2004 6:18 am
Location: Alphanumeric symbols.

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 8:32 pm

It looks like its pretty much the same story with the BX200, SP550, and Trion 100 from what I saw. I'm picking up an Arc 100 instead for my laptop at about the same cost.
Meow.
 
just brew it!
Gold subscriber
Administrator
Posts: 52313
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Preliminary results for Crucial BX200 degraded read spee

Tue Mar 22, 2016 8:50 pm

Zoomastigophora wrote:
I thought TRIM only affects write speeds?

That was my impression as well.
Nostalgia isn't what it used to be.
 
DPete27
Silver subscriber
Grand Gerbil Poohbah
Posts: 3652
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 8:59 am

You're right, TRIM traditionally affects/improves write speeds. Read speeds generally don't degrade. (example) However, it seems oddly convenient that only the blocks that have been written to are affected for reads in the above tests. That's the reason I suggested running TRIM manually (to ensure it's been done) to see if that improves performance.

I didn't mean for my initial comment to come off as "this is typical behavior for all SSDs", but in retrospect, it certainly did.
Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
just brew it!
Gold subscriber
Administrator
Posts: 52313
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 11:29 am

DPete27 wrote:
You're right, TRIM traditionally affects/improves write speeds. Read speeds generally don't degrade. (example) However, it seems oddly convenient that only the blocks that have been written to are affected for reads in the above tests. That's the reason I suggested running TRIM manually (to ensure it's been done) to see if that improves performance.

There is an exception, however. While it should not affect the read speed of valid data, TRIM *can* affect read speeds of blocks that have been TRIMmed (i.e. the data in them is no longer considered valid). Reading from unwritten blocks on a fresh drive (or blocks on a used drive which have subsequently been completely deallocated by the file system and TRIMmed) does not require the drive to read the contents of the flash AT ALL, other than the small amount of internal meta-data the drive firmware needs to determine that the blocks in question are invalid. Once the drive knows that a given block is invalid, it can just shovel zeros out the interface as fast as it can without even touching the flash, since it knows the file system doesn't care about the contents.
Nostalgia isn't what it used to be.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 4:05 pm

DPete27 wrote:
You're right, TRIM traditionally affects/improves write speeds. Read speeds generally don't degrade. (example) However, it seems oddly convenient that only the blocks that have been written to are affected for reads in the above tests. That's the reason I suggested running TRIM manually (to ensure it's been done) to see if that improves performance.

I didn't mean for my initial comment to come off as "this is typical behavior for all SSDs", but in retrospect, it certainly did.


Well, I think that the results are more consistent with it having the same issue as the 840/840 EVO do.
Meaning that the cells lose charge over time reducing read speeds because of ECC having to process them longer than if they were newly written.
Or if they are left empty.

Additionally if you read my earlier thread about the ADATA SP550 you can see just how similarly the two behave.
Read speeds go up after doing a full read and may be improved in time for the next read, especially if done soon after.

That is likely because of ECC being able to see where there are issues and makes it easier for it to do something about it.


While leaving the ADATA SP550 powered but idle which gives it time to do garbage collection did not seem to improve read speeds.


Additionally during one of the tests of the ADATA SP550 I noticed that Current Pending Sector went up from 0 to over 100,000 before going back to 0 again without any increase in reallocated sectors which I think is because ECC was working on those files.
In the article I said I thought it was StaticDataRefresh doing those things, but I now think LDPC would fit the bill better since it can try to take into account things like age and temperature and would also explain why no files seemed to be rewritten.


Still, I will run TRIM during my next test to exclude it as a possibility.
I might do the test with SSDReadSpeedTester however instead of HD Tune but that should work as well even if the graph is a bit different.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 4:07 pm

Savyg wrote:
It looks like its pretty much the same story with the BX200, SP550, and Trion 100 from what I saw. I'm picking up an Arc 100 instead for my laptop at about the same cost.


Have you seen any read speed issues with any of those besides in my posts?
If so I would be interested in seeing them, especially those for the Trion 100.
 
DPete27
Silver subscriber
Grand Gerbil Poohbah
Posts: 3652
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 4:11 pm

Are you suggesting this is TLC bit validation overhead?

Found this from the Samsung 840 Evo review:
TechReport wrote:
Because there is some cell-to-cell variance in the silicon, writing to the NAND involves a verification step that reads the contents of the cell. Reading entails applying a series of control voltages to hone in on the charge in the floating gate. In TLC NAND, that charge can represent one of eight possible values between 000 and 111. MLC NAND only has to contend with four values between 00 and 01, while SLC is limited to two: 0 and 1. The more bits per cell, the more iterative steps are required to verify the data, slowing the write process.
Last edited by DPete27 on Wed Mar 23, 2016 4:22 pm, edited 1 time in total.
Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
DPete27
Silver subscriber
Grand Gerbil Poohbah
Posts: 3652
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 4:22 pm

Glaring_Mistake wrote:
Well, I think that the results are more consistent with it having the same issue as the 840/840 EVO do.
Meaning that the cells lose charge over time reducing read speeds because of ECC having to process them longer than if they were newly written.
Or if they are left empty.
Additionally if you read my earlier thread about the ADATA SP550 you can see just how similarly the two behave.
Read speeds go up after doing a full read and may be improved in time for the next read, especially if done soon after.

I see what you're getting at now. That thought had crossed my mind, but you didn't mention length of time between writing the data and testing. The Samsung TLC bug needed some time to let the cells discharge. Simply writing the drive full and testing right away didn't show degraded performance.

FWIW, when the 840 EVO bug struck, I filled and tested a Sandisk Ultra II (TLC) and it did not show any signs of read speed reduction.
Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
diasflacog
Gerbil Team Leader
Posts: 228
Joined: Tue Dec 18, 2007 3:58 am

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 4:37 pm

I love crucial drives so that's why I feel like I need t post here ^^

I have been rocking 2 crucial ssd's now. The MX100's. I have two of them. A 256 for my OS and a 512 for my steam games.

I love these drives. Had em for over a year now and not a single problem. They still read just as fast as when I bought them.

As far as I know the BX line up is a lower tier one so maybe that's why the read speeds have degraded?

I know they also have the MX200 which is the more higher tier ones like the older MX100.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 5:11 pm

DPete27 wrote:
I see what you're getting at now. That thought had crossed my mind, but you didn't mention length of time between writing the data and testing. The Samsung TLC bug needed some time to let the cells discharge. Simply writing the drive full and testing right away didn't show degraded performance.

FWIW, when the 840 EVO bug struck, I filled and tested a Sandisk Ultra II (TLC) and it did not show any signs of read speed reduction.


I first put one folder of 20GB on it and then tested read speeds after it was left unpowered for a week and then added another folder of equal size and did the same thing the week after that and the week after that and... Well you get what I mean.
May not have made this as clear as I could have in my first post however.

Appreciate the test of the SanDisk Ultra II.
It's good to see more results of read speed tests for SSDs even if that doesn't prevent from running some tests myself.
I have a Trion 100 among the SSDs I'm running tests on which uses the same NAND but a different controller.
And the one in the Trion 100 was made with TLC NAND in mind unlike the one in SanDisk Ultra II so we'll see which does best.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 5:47 pm

diasflacog wrote:
I love crucial drives so that's why I feel like I need t post here ^^

I have been rocking 2 crucial ssd's now. The MX100's. I have two of them. A 256 for my OS and a 512 for my steam games.

I love these drives. Had em for over a year now and not a single problem. They still read just as fast as when I bought them.

As far as I know the BX line up is a lower tier one so maybe that's why the read speeds have degraded?

I know they also have the MX200 which is the more higher tier ones like the older MX100.


Well, you could say that it is because of it being a low tier SSD.

But in reality it is a bit more complicated than that.

It didn't used to be like that (with an exception for the 840/840 EVO).

But when so many companies have gone over to using 2D TLC NAND at small litographies it gets difficult to compete (or perhaps more likely, get the same price margins) using MLC NAND.
Regardless of brand that means it's going to be more difficult to prevent the memory cells from leaking electrons which will if left unchecked result in low read speeds like the ones I have reported.
And they're likely going to sink a lot more than that before I'm through with my tests.

So today for the lower tier SSDs that means you're (most likely) going to get one with lower tolerances than the previous lower tier SSDs using MLC NAND.

Additionally the 840 EVO which had this issue was not excempt from criticism despite being a lower tier SSD than the 840 Pro.
And though the BX200 admittedly can raise read speeds (with varying success) if you do a full read of the SSD, read speeds in this test seem to sink at a rate that (I think) even an 840 EVO would have trouble matching.
 
MOSFET
Gold subscriber
Gerbil XP
Posts: 316
Joined: Fri Aug 08, 2014 12:42 am

Re: Preliminary results for Crucial BX200 degraded read spee

Wed Mar 23, 2016 10:13 pm

1) what jbi said

2) It's because the drives are not powered up for any length of time - try leaving the drive in the system powered but idle for the next week (unless that negates your research plans).

3) I also love Crucial drives, especially since they released ssd software last year (which is excellent).

4) After one Corsair Nova five years ago, I will never ever buy a "budget" ssd again. Life is too short. In that case, give me back a 320GB WD Blue and at least I'll know what to expect.
Be careful on inserting this (or any G34 chip) into the socket. Once you pull that restraining lever, it is either a good install or a piece of silicon jewelry.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Thu Mar 24, 2016 9:45 am

MOSFET wrote:
2) It's because the drives are not powered up for any length of time - try leaving the drive in the system powered but idle for the next week (unless that negates your research plans).

3) I also love Crucial drives, especially since they released ssd software last year (which is excellent).


The fact that they are left unpowered has an effect on read speeds but so does reading the entire SSD so there is a +/- effect going on if you leave them unpowered and then test read speeds.
And like I said these tests do not actually demonstrate whether read speeds would be perfectly normal when plugged in.

After read speeds have dropped low enough however I do intend to keep them plugged in for a time to see how well they recover which may not be the same as leaving them plugged in from the start but should still give some indication of how well they can fix read speeds by checking out files when they're idling.
If I do decide to test how much read speeds would drop when they're constantly plugged in, I'm likely going to do just one or a select few since keeping all of them plugged in would require getting an additional computer, maybe two.


And well, I was also pleasantly surprised by how well Crucial Storage Executive seems to work.
 
diasflacog
Gerbil Team Leader
Posts: 228
Joined: Tue Dec 18, 2007 3:58 am

Re: Preliminary results for Crucial BX200 degraded read spee

Thu Mar 24, 2016 1:16 pm

I never been a big fan of low tier SSD's. As far as I know the BX series is their lowest tier.

I rather pay a lil more for the more decent quality one's like the MX one's that I own or in this case the MX200's that replace the 100's.

Im guessing that when you buy the lowest tier ssd's things like this happen.
 
MOSFET
Gold subscriber
Gerbil XP
Posts: 316
Joined: Fri Aug 08, 2014 12:42 am

Re: Preliminary results for Crucial BX200 degraded read spee

Thu Mar 24, 2016 1:33 pm

After read speeds have dropped low enough however I do intend to keep them plugged in for a time to see how well they recover


When people I know archived CD's 15 years ago, they bought the most premium Gold CD-R's they could find. I'm not sure where your "cheap SSD archive" research is headed, but it's not where I would spend my time. Have at it, though. This is not how SSD's are designed to work.

diasflacog wrote:
Im guessing that when you buy the lowest tier ssd's things like this happen.


Yes, and he's causing it to happen.
Be careful on inserting this (or any G34 chip) into the socket. Once you pull that restraining lever, it is either a good install or a piece of silicon jewelry.
 
just brew it!
Gold subscriber
Administrator
Posts: 52313
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Preliminary results for Crucial BX200 degraded read spee

Thu Mar 24, 2016 2:04 pm

It's useful info regardless. If nothing else, we can point to this thread in the future to dissuade anyone who thinks they can use cheap SSDs for archival storage!
Nostalgia isn't what it used to be.
 
meerkt
Graphmaster Gerbil
Posts: 1039
Joined: Sun Aug 25, 2013 2:55 am

Re: Preliminary results for Crucial BX200 degraded read spee

Fri Mar 25, 2016 10:29 am

Makes you wonder about the retention durability of modern UFDs. I would assume most of them use the cheapest reject TLC the vendor can find.
 
Ryu Connor
Gold subscriber
Global Moderator
Posts: 4340
Joined: Thu Dec 27, 2001 7:00 pm
Location: Marietta, GA
Contact:

Re: Preliminary results for Crucial BX200 degraded read spee

Fri Mar 25, 2016 11:45 am

just brew it! wrote:
It's useful info regardless. If nothing else, we can point to this thread in the future to dissuade anyone who thinks they can use cheap SSDs for archival storage!


Honestly it's a good thread for anyone thinking of using NAND flash for archival storage period.
All of my written content here on TR does not represent or reflect the views of my employer or any reasonable human being. All content and actions are my own.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2667
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Preliminary results for Crucial BX200 degraded read spee

Fri Mar 25, 2016 12:09 pm

In general, SSDs shouldn't be left un-powered for long periods of time. The closer you get to that magical point when the ECC can no longer keep up the slower the drive will get.

TLDR, normal operation for anything flash.
Desktop: Z170A Gaming Pro Carbon | 6700K @ 4.4 | 16 GB | Radeon VII | XSPC RX360 | Heatkiller R3 | Samsung 4K 40" | 2048 + 240 + LSI 9207-8i (128x8) SSD
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Fri Mar 25, 2016 2:03 pm

MOSFET wrote:
When people I know archived CD's 15 years ago, they bought the most premium Gold CD-R's they could find. I'm not sure where your "cheap SSD archive" research is headed, but it's not where I would spend my time. Have at it, though. This is not how SSD's are designed to work.

Yes, and he's causing it to happen.


I'm testing a variety of SSDs of varying quality and the reason I report on those of lower quality first is, well it is fairly obvious.

And SSDs are designed for having a data retention of (depending on the environmental conditions) a year after it's run out of it's rated write cycles.
My tests may not exactly represent the most common use, but they're not even close to being a worst case scenario.



You're a bit focused on the effect of it being constantly powered and ignoring that it is being read in its entirety regularly.

Check out my thread on the ADATA SP550 and look at the last pictures.
It is about 14 hours between them with it being powered on the entire time between the two tests yet read speeds have gone down in the last picture.

That would suggest that doing a full read of the SSD can be more beneficial for read speeds than just leaving it to idle.
Which indicates that ECC tries to fix read speeds when it needs to read those files meaning it looks more like it is reactive rather than proactive when it comes to fixing read speeds.


Even if I have not tested SSDs that have been left powered on extensively that does not mean we have to assume that them being powered on would have to make them immune to these kinds of effects.


And like I said before: My tests are nowhere near complete.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Fri Mar 25, 2016 2:03 pm

Did not run TRIM this time since I did not find an option for it in Crucial Storage Executive, though I think we have established that it is most likely not in stead state.

Nonetheless here you can see the latest results with SSDReadSpeedTester.

Image

Let me know if you want to see the results for every file.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Thu May 05, 2016 1:38 pm

Update may be a bit late this time but here you can see the results from the the latest test.

Image


I’ll have to say that it is an impressive result seeing as the average read speeds are higher than last time despite it being unpowered for four times as long as usual.
It is also made clear that BX200 can adjust for voltage drift and that it may be triggered when reading files with lower read speeds even if it doesn’t have as much of an impact every time.
 
Glaring_Mistake
Gerbil
Topic Author
Posts: 31
Joined: Wed Feb 03, 2016 2:35 pm

Re: Preliminary results for Crucial BX200 degraded read spee

Thu Jun 02, 2016 4:08 pm

Posting a bit late once again but here you can see the results from the latest test.

Image

So, read speeds are once again decreasing.
 
meerkt
Graphmaster Gerbil
Posts: 1039
Joined: Sun Aug 25, 2013 2:55 am

Re: Preliminary results for Crucial BX200 degraded read spee

Thu Jun 02, 2016 6:09 pm

After being unpowered for a month? How much on time did it have a month ago?

Who is online

Users browsing this forum: No registered users and 3 guests