Overclock.net › Forums › Components › Hard Drives & Storage › SSD › Samsung 840 EVO read speed drops on old-written data in the drive
New Posts  All Forums:Forum Nav:

Samsung 840 EVO read speed drops on old-written data in the drive - Page 291

post #2901 of 3279
Quote:
Originally Posted by Tivan View Post

I generally agree, though not sure if it's the/a new read algorithm? I was under the assumption that it simply rewrites data in the background more aggressively.

It was suggested that this is just something that was lost in translation on the way from the engineers to the writers in marketing (and on the way to English). One of the tech journalists writing about this has an account here and explained what he looked at while testing the firmware update. I think he said, he couldn't find anything suggesting that the drive was internally doing something big while sitting around idle. I don't quite remember, but I think he tracked things like power use of the drive. He concluded the drive was getting faster for some other reason, not because of rewriting stuff.
post #2902 of 3279
Quote:
Originally Posted by opcow View Post

I don't put it past Samsung to fudge things. I think that's exactly what they did with the first "fix." They may very well have hoped that the original fix worked, but by rewriting the data they made it impossible to know if it really did and they effectively kicked the can down the road. But I don't for one second believe they are faking reads to fool benchmarks. Even if doable (which it almost certainly isn't) that would be clear malfeasance and a huge mistake. This issue could be expensive, but it's not big enough for them to do something that stupid.

With that out of the way, I have a couple of questions.

1) Do the drives have an autonomous function which periodically rewrites old data or do I need to have Magician running for that to happen?

If it's the latter then I still consider the drive to be broken. The fact that the drive was much faster immediately after the firmware update isn't enough to convince me the drive is fixed (I'm keeping an open mind, but I'm still wary). It seems ok now, but what about a few months from now? I run a RAID0 so Magician isn't an option. I'm curious that that functionality would exist if it isn't needed.

2) Why can't Samsung's software see which drives I'm running in my RAID? Are their programmers just lazy?

My understanding is that other manufacturers don't have this problem with their software, and I can open up rapid storage and see my drives' Model and serial numbers and the firmware revision they're using, but even Samsung's data migration software which shouldn't have needed to do anything low-level refused to work because my drives were RAID'd and it didn't see them.

1. The FW update came with a new read algorithm and a Auto refresh feature that might run when your drive is idle, but from what i've read, it only does so on degraded blocks.

2. Only Sandisk's program can see their drives in RAID. No other company, as far as i know and researched, can see drives in a RAID setup to do FW updates.

*EDit*
Forgot to add to #1, Magician is not needed.
My Stuff
(13 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-6700K ASUS Z170-A  EVGA GeForce GTX 980 K|NGP|N EVGA GeForce GTX 980 K|NGP|N 
RAMHard DriveCoolingOS
G.SKILL Ripjaws 4 Series 16GB (4 x 4GB) Samsung 850 Pro 256GB Corsair Hydro Series H100i GTX Windows 10 Pro 
MonitorKeyboardPowerCase
AOC G2460PG 24" G-SYNC 1ms (GTG) 144hz  Logitech Wireless Illuminated Keyboard K800 Antec HCP-1000 Platinum thermaltake chaser series chaser mk-i 
Mouse
Logitech G700S 
  hide details  
Reply
My Stuff
(13 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-6700K ASUS Z170-A  EVGA GeForce GTX 980 K|NGP|N EVGA GeForce GTX 980 K|NGP|N 
RAMHard DriveCoolingOS
G.SKILL Ripjaws 4 Series 16GB (4 x 4GB) Samsung 850 Pro 256GB Corsair Hydro Series H100i GTX Windows 10 Pro 
MonitorKeyboardPowerCase
AOC G2460PG 24" G-SYNC 1ms (GTG) 144hz  Logitech Wireless Illuminated Keyboard K800 Antec HCP-1000 Platinum thermaltake chaser series chaser mk-i 
Mouse
Logitech G700S 
  hide details  
Reply
post #2903 of 3279
Anyone please confirm 850 EVO has this bug or not?
post #2904 of 3279
Quote:
Originally Posted by nhc511 View Post

Anyone please confirm 850 EVO has this bug or not?

So far nothing has been found with the 850 evo, but its only been out for 5-6 months now,
My Stuff
(13 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-6700K ASUS Z170-A  EVGA GeForce GTX 980 K|NGP|N EVGA GeForce GTX 980 K|NGP|N 
RAMHard DriveCoolingOS
G.SKILL Ripjaws 4 Series 16GB (4 x 4GB) Samsung 850 Pro 256GB Corsair Hydro Series H100i GTX Windows 10 Pro 
MonitorKeyboardPowerCase
AOC G2460PG 24" G-SYNC 1ms (GTG) 144hz  Logitech Wireless Illuminated Keyboard K800 Antec HCP-1000 Platinum thermaltake chaser series chaser mk-i 
Mouse
Logitech G700S 
  hide details  
Reply
My Stuff
(13 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-6700K ASUS Z170-A  EVGA GeForce GTX 980 K|NGP|N EVGA GeForce GTX 980 K|NGP|N 
RAMHard DriveCoolingOS
G.SKILL Ripjaws 4 Series 16GB (4 x 4GB) Samsung 850 Pro 256GB Corsair Hydro Series H100i GTX Windows 10 Pro 
MonitorKeyboardPowerCase
AOC G2460PG 24" G-SYNC 1ms (GTG) 144hz  Logitech Wireless Illuminated Keyboard K800 Antec HCP-1000 Platinum thermaltake chaser series chaser mk-i 
Mouse
Logitech G700S 
  hide details  
Reply
post #2905 of 3279
Quote:
Originally Posted by Techie007 View Post

    To me, this borders on fanaticism and demonstrates a complete lack of knowledge as to what is actually possible.  Samsung is already doing their best trick with RAPID.  This is not a matter of being outclassed, it is a matter of understanding how data transfers are handled between storage devices and software.  If Samsung's programmers were so great, we would not have had this issue in the first place, let alone a second time with the October fix.  I know the many stages the data goes through to get from the SSD to where it can be used by programs, and I will repeat: There is no way that the SSD can pretend to "read faster" than it really does.  You can insert caching mechanisms like RAPID in the middle, but the end of the matter is that the SSD will have to retrieve data eventually because everything doesn't fit in RAM, and at that point, the SSD's real performance will show.
    Once the SSD has fulfilled the data read request, there's no going back.  Either the request was fulfilled successfully (no error code) and in full (all the data that was requested), or it was not.  From a programs' standpoint, the fulfillment of the request comes with the data.  There's no pretending to a benchmark program (especially ones like SSD Read Speed Tester and File Bench that go through the filesystem) that the reads were fulfilled sooner than they were.  In other words, when the read request returns, it's returning with all the data, and the SSD can't say later "Oh, I've got more/revised data for you now."  Besides actually improving their algorithms or overclocking the SSD (limited by its already designed thermal capabilities), the only corner they can cut is error correction.  And if just a little corrupted data is returned—depending on when—your computer:
  • Will BSOD instead of booting
  • Won't be able to successfully boot to the desktop
  • Will crash programs on a regular basis, particularly when loading them
  • Will frequently report filesystem errors or missing files

    So for these reasons, they can't bypass error correction.  And error correction is rather black and white: Either the chunk just read (internally by the SSD) passes checksum, or it does not and needs to be read again.  That leaves one possibility: The new read algorithm has a lower error rate and/or a better way of reevaluating data to pass checksum quickly.

    While I'm fully with you guys that Samsung did a horrible job at customer service, I have the feeling from some of you that no matter what Samsung did to fix the 840 EVO, you would not be pacified.  BUT: economically, I believe that Samsung made the best decisions.  Coming clean publicly or giving more feedback to us about the issue—while appealing to technically inclined customers—would greatly spread word that there is something wrong with their SSDs.  As things stand now, most people are using the old, initial firmware blissfully unaware of the slower performance they are experiencing, and meanwhile, tons of positive reviews on the SSDs are pouring into places like Amazon daily.  Those of you who have been deeply offended are a very tiny part of their customer base.


You are under the mistaken impression that your software (and any other) can't be fooled by a few lines of code.

Check for programs X,Y,Z. Do this.

Easy to do and is the level that Samsung has fallen too many times before. Any benchmark program including your utility doesn't do anything with the data, it merely returns 'data was read'. Therefore no error correction or bypassing of such needed.

My theory also explains the new BSOD's, btw. While your position is based on a Samsung that wants to do their best for it's customers and they have blatantly shown time and again that is not what they're here to do.

Step back from how unlikely it is for Samsung to do such dastardly deeds and then maybe you too will see how all (or at least some) of your support for Samsung is misplaced.

You and others here should not be defending Samsung, that is their job. But they're not and that is the first clue that their goals are not as altruistic as you and a few other here think they are.

And because of this, I'm sure that we'll be discussing this same/related issue in a few months from now again. Because fix #3 will be in the wild for an issue that originated in the middle of 2013.
post #2906 of 3279
So you are saying Samsung coded their firmware to check for those specific user-created benchmarks, and return false results just for them?

That seems...farfetched, I guess is the word I'm looking for - something slightly below bat-crap crazy?
Edited by Forceman - 5/18/15 at 8:22am
post #2907 of 3279
@SATDK:

What you suggest feels actually criminal. It would mean sabotaging one of the most basic Windows functions that all programs use, just to fool one program. This would be risky for the data. Your idea is too crazy.

I'd like to test things under Linux just for you, but I do not have a 840 EVO, so can't do it. tongue.gif
post #2908 of 3279
Quote:
Originally Posted by Forceman View Post

So you are saying Samsung coded their firmware to check for those specific user-created benchmarks, and return false results just for them?

That seems...farfetched, I guess is the word I'm looking for - something slightly below bat-crap crazy?

I think what they did is code the firmware to "look ahead" and mitigate based on read patterns. This is why we see instant (but not 100% full speed) read improvement. My guess is they ARE detecting when the drive is being benchmarked, and mitigating on the fly. If the firmware was truly magical like some seem to think, we'd see 100% read speed improvement immediately, which is not the case.
EconoGamer#
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 4790K 5 GHz (delidded) ASRock Z97 Extreme4 nVidia Titan Xp Team Xtreem 16GB/2400 
Hard DriveHard DriveCoolingOS
Corsair Force GT 120G SSD (OS) Crucial MX200 500G SSD (Games) Corsair H75 Hydro Cooler Windows 10 x64 
MonitorKeyboardPowerCase
Acer Predator Z35 Corsair K95 RGB Platinum Corsair HX750 v2 80+ Gold Thermaltake Core X9 (M0dd3D!) 
MouseMouse PadAudioAudio
Logitech G403 wireless XTrac Ripper XL (cloth) Realtek ALC1150 (on-board) Audio Technica ATH-MSR7 Headphones 
  hide details  
Reply
EconoGamer#
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 4790K 5 GHz (delidded) ASRock Z97 Extreme4 nVidia Titan Xp Team Xtreem 16GB/2400 
Hard DriveHard DriveCoolingOS
Corsair Force GT 120G SSD (OS) Crucial MX200 500G SSD (Games) Corsair H75 Hydro Cooler Windows 10 x64 
MonitorKeyboardPowerCase
Acer Predator Z35 Corsair K95 RGB Platinum Corsair HX750 v2 80+ Gold Thermaltake Core X9 (M0dd3D!) 
MouseMouse PadAudioAudio
Logitech G403 wireless XTrac Ripper XL (cloth) Realtek ALC1150 (on-board) Audio Technica ATH-MSR7 Headphones 
  hide details  
Reply
post #2909 of 3279
Quote:
Originally Posted by Sheyster View Post

I think what they did is code the firmware to "look ahead" and mitigate based on read patterns. This is why we see instant (but not 100% full speed) read improvement. My guess is they ARE detecting when the drive is being benchmarked, and mitigating on the fly. If the firmware was truly magical like some seem to think, we'd see 100% read speed improvement immediately, which is not the case.

That seems not possible for a benchmark that continually (without pause) tries to read as fast as it can for the whole hundreds of GB of contents on the drive (so too much to fit into a cache).

I was thinking they perhaps managed to make it so the firmware can learn what correction is need for each single spot with regards to that voltage drift thingy that makes cells with old data different from cells with fresh data.

What's annoying is that they don't talk thoroughly about what they are doing with this new firmware. My idea from the last paragraph sounds like something everyone would be satisfied with, but them not confirming this might mean that it's not the right idea. Their real solution might be different and might be something that people will think is ugly when they hear about it, and that might be why they aren't talking about it a lot.
post #2910 of 3279
Quote:
Originally Posted by deepor View Post

That seems not possible for a benchmark that continually (without pause) tries to read as fast as it can for the whole hundreds of GB of contents on the drive (so too much to fit into a cache).

I was thinking they perhaps managed to make it so the firmware can learn what correction is need for each single spot with regards to that voltage drift thingy that makes cells with old data different from cells with fresh data.

What's annoying is that they don't talk thoroughly about what they are doing with this new firmware. My idea from the last paragraph sounds like something everyone would be satisfied with, but them not confirming this might mean that it's not the right idea. Their real solution might be different and might be something that people will think is ugly when they hear about it, and that might be why they aren't talking about it a lot.

From what i understand, the reason it doesn't go all the way up to as if the data was brand new, without the diskfresh or advanced optimization, is because on any drive that has stale data, the performance will start to go down, its just that the EVO wasn't catching it correctly and so the performance would continue to drop the longer it got stale. Drives aren't going to have solid straight lines their entire lifetime as data gets older.
My Stuff
(13 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-6700K ASUS Z170-A  EVGA GeForce GTX 980 K|NGP|N EVGA GeForce GTX 980 K|NGP|N 
RAMHard DriveCoolingOS
G.SKILL Ripjaws 4 Series 16GB (4 x 4GB) Samsung 850 Pro 256GB Corsair Hydro Series H100i GTX Windows 10 Pro 
MonitorKeyboardPowerCase
AOC G2460PG 24" G-SYNC 1ms (GTG) 144hz  Logitech Wireless Illuminated Keyboard K800 Antec HCP-1000 Platinum thermaltake chaser series chaser mk-i 
Mouse
Logitech G700S 
  hide details  
Reply
My Stuff
(13 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-6700K ASUS Z170-A  EVGA GeForce GTX 980 K|NGP|N EVGA GeForce GTX 980 K|NGP|N 
RAMHard DriveCoolingOS
G.SKILL Ripjaws 4 Series 16GB (4 x 4GB) Samsung 850 Pro 256GB Corsair Hydro Series H100i GTX Windows 10 Pro 
MonitorKeyboardPowerCase
AOC G2460PG 24" G-SYNC 1ms (GTG) 144hz  Logitech Wireless Illuminated Keyboard K800 Antec HCP-1000 Platinum thermaltake chaser series chaser mk-i 
Mouse
Logitech G700S 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: SSD
Overclock.net › Forums › Components › Hard Drives & Storage › SSD › Samsung 840 EVO read speed drops on old-written data in the drive