Personal computing discussed

Moderators: renee, morphine, Steel

  • 1
  • 5
  • 6
  • 7
  • 8
  • 9
  • 11
 
Chrispy_
Maximum Gerbil
Posts: 4670
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Samsung 840 & 840 EVO Bug

Wed Apr 29, 2015 7:17 am

just brew it! wrote:
Something is definitely fishy with the HD Tune results people have been posting. The reported Burst Rate is sometimes lower than the Minimum and Average rates, which should be impossible?


Both HDTune and HDTach are legacy spinning disk programs that don't really have updates for the modern SSD era. You should probably be worrying more about the -1% CPU usage ;)

The exact nature of the details is not really that relevant - what matters is that in real-world usage, users are reporting awful performance; Pre-Diskfresh (or Samsung's new firmware doing the same thing), the HDTach/HDTune benchmarks are all over the shot with some staggeringly low numbers to back up the lousy real-world performance that users are whining about. After the Diskfresh or firmware-based refresh, real-world performance is good and the HDTach/HDTune benchmarks are a nice straight line again.

All these old benchmark tools do is pick a random series of blocks at that position on the drive and measure how fast the read of those blocks is. With SSD's it all over the place because the physical location of the blocks no longer corresponds to a single physical location - it's up the SSD controller where in the NAND cells the virtual block pointers are pointed at.

The graph plot tools are just secondary evidence to support awful read speeds on these defective drives. The concern is not with the exact output of the tools, just that the tools are an easy visual representation of an obvious problem that goes away after a Diskfresh. Straight line = good, drive is working within spec. Erratic line = unmistakable performance inconsistency to back up user claims of slow SSDs.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
Chrispy_
Maximum Gerbil
Posts: 4670
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Samsung 840 & 840 EVO Bug

Wed Apr 29, 2015 7:38 am

localhostrulez wrote:
I ran hdtune on my M500 in the 6530b (SATA II). The results look like the drive's write speeds (120GB M500's are slow at that), not the read speeds. Huh? http://i.imgur.com/IAYZLQY.png


I think this comes down to the age of the graphing tools as mentioned in my post above - these tools are not SSD-aware and don't have a concept of partition alignment, page vs block, Trim, etc. You can't really treat them with the same level of confidence as results from a tool that is SSD-aware like CrystalDiskmark.

Whilst they don't give accurate quantitative results (the exact MB/s numbers coming out of HDTune/HDTach don't correspond with the results of a properly SSD-aware, alignment-aware, page-aware benchmark tool), they do give qualitative results showing that performance is massively inconsistent - both erratic and slow before any kind of refresh. Regardless of what the absolute read speeds are, the tools should at least show consistent performance in vaguely the right ballpark - in other words a straight line with no major dips into low write speeds.

Here's an ancient JMicron controller SSD. It's a cheap drive using a cheap controller and low-cost NAND. Even on SATA3, it's not even able to reach SATA2 speeds. It's one of the worst drives I've used, but it's consistent:
Image

Then there's this Intel 330-series which is hardly a high-end product, but again it's perfectly capable of delivering a consistent experience:
Image

These are old, heavily used drives that were low-budget, low-performance options when they were introduced (long before even the original 840-series). Comparing these plots to any of the 840 or 840EVO plots in this thread, you know that something's seriously wrong with the Samsung drives.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
localhostrulez
Minister of Gerbil Affairs
Posts: 2481
Joined: Sun Mar 09, 2014 11:26 pm

Re: Samsung 840 & 840 EVO Bug

Fri May 01, 2015 6:34 pm

Oh. My. God. I'm transferring a bunch of my stuff off the 840 evo in the t440s and onto the second drive in the 8200 via gig-e. (I just bought a WD Blue 1TB. Gets ~170MB/s reads and ~130MB/s writes sequentially on benchmarks.) And some of it's going at ~10-15MB/s (smaller files mainly). Guess what the bottleneck is? Yep, the SSD... the hard drive seems to be handling it just fine. Never thought I'd see that - a hard drive and an SSD, from about the same time period, where a transfer between them is bottlenecked by the SSD. What the hell? I've seen pathetic read speeds on the 840 evo before, but never tested it in an embarrassing manner like that.

I'm also beginning to wonder if the t440s is defective. It just locked up on me when I plugged in the power brick, acted like nothing happened when I unplugged it, and locked back up when plugged in. Inconsistent. I just asked my mom to ship the original t440s HDD and my 840 evo box up to me. I might just call Lenovo on the warranty if this continues, and at any rate, I'm probably going to sell both and find a different config. What a bunch of garbage. (Meanwhile, in a land a desk's length away, a bunch of HP equipment running in the same way is humming along happily, enjoying the nice Oregon spring weather. OK, maybe not the last bit, though I am. :lol:)
 
Welch
Grand Gerbil Poohbah
Posts: 3582
Joined: Thu Nov 04, 2004 5:45 pm
Location: Alaska
Contact:

Re: Samsung 840 & 840 EVO Bug

Fri May 01, 2015 8:34 pm

I'd be a bit leery of those speeds your seeing. Small files never have the oppertunity to get up to the full speed of the drive in question before the drive has to handle a new request for another file. How big are the files your talking about that are reading that slow?
"I think there is a world market for maybe five computers."
Thomas Watson, chairman of IBM, 1943

1600x | Strix B350-F | CM 240 Lite | 16GB 3200 | RX 580 8GB | 970 EVO | Corsair 400R | Seasonic X 850 | Corsair M95 / K90 | Sennheiser PC37x
 
cobalt
Gerbil First Class
Posts: 171
Joined: Mon Oct 30, 2006 11:28 am

Re: Samsung 840 & 840 EVO Bug

Fri May 01, 2015 9:26 pm

Hey, for kicks, here's my current 840 (vanilla, non-EVO) HDTune results. Is this pathetic, or what?

Image

What's that about only the EVO's being affected, Samsung? If they just owned up to it and allowed the tool to work on the normal 840's, even if they didn't update the firmware, I wouldn't be so pissed. It's for repeatedly denying it that they've lost my trust. It's dishonest and disrespectful, and I'll be looking elsewhere in the future if that's what they think of their customers.

Well, off to try diskfresh! (edit: yes, diskfresh seems to be doing the trick. Thanks, everyone. It's 5% done, and my first 5% has gone from 25-50MB/sec to 400+MB/sec.)
 
Welch
Grand Gerbil Poohbah
Posts: 3582
Joined: Thu Nov 04, 2004 5:45 pm
Location: Alaska
Contact:

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 12:34 am

WOW I think that is the most anemic score I've seen yet. Agree, no reason they couldn't enable 840 Vanilla to be able to run the restore utility at least.

Have you ran DiskFresh before? You've known about this issue for how long now? How often if multiple times, does it take for performance to degrade to that level again?
"I think there is a world market for maybe five computers."
Thomas Watson, chairman of IBM, 1943

1600x | Strix B350-F | CM 240 Lite | 16GB 3200 | RX 580 8GB | 970 EVO | Corsair 400R | Seasonic X 850 | Corsair M95 / K90 | Sennheiser PC37x
 
Chrispy_
Maximum Gerbil
Posts: 4670
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 5:53 am

Welch wrote:
WOW I think that is the most anemic score I've seen yet. Agree, no reason they couldn't enable 840 Vanilla to be able to run the restore utility at least.

Have you ran DiskFresh before? You've known about this issue for how long now? How often if multiple times, does it take for performance to degrade to that level again?


That graph looks like about 6 months to my eyes, based on the various 840's I'm keeping an eye on.
My drive is due for an auto-scheduled Diskfresh this Sunday night so it has gone almost a month without one and my performance graph has already dropped from a 520MB/s straight line to a 125-450MB/s wild squiggle with a 290MB/s average. It's bad, but it's not 9MB/s bad.

My advice to Samsung-related Diskfreshers: Change your scheduled task from quarterly to monthly (or even more frequently!). They're already in pretty terrible shape after a month.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
just brew it!
Administrator
Posts: 54499
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 6:15 am

Chrispy_ wrote:
They're already in pretty terrible shape after a month.

And *that*, right there, is a pretty darned serious indictment of Samsung's product validation and testing procedures. How the heck could they possibly have missed this?
Nostalgia isn't what it used to be.
 
thecoldanddarkone
Minister of Gerbil Affairs
Posts: 2449
Joined: Wed Mar 26, 2003 4:35 pm

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 8:56 am

cobalt wrote:
Hey, for kicks, here's my current 840 (vanilla, non-EVO) HDTune results. Is this pathetic, or what?

Image

What's that about only the EVO's being affected, Samsung? If they just owned up to it and allowed the tool to work on the normal 840's, even if they didn't update the firmware, I wouldn't be so pissed. It's for repeatedly denying it that they've lost my trust. It's dishonest and disrespectful, and I'll be looking elsewhere in the future if that's what they think of their customers.

Well, off to try diskfresh! (edit: yes, diskfresh seems to be doing the trick. Thanks, everyone. It's 5% done, and my first 5% has gone from 25-50MB/sec to 400+MB/sec.)


Wow that's so bad it's slower than the ssd in my original latitude xt (well for a pretty big percentage of the drive) and it had a ziff drive. :lol:
I7 4930k, 32 GB Ballistix [email protected] , 1.2 TB Intel 750 AIC, 500 GB mx200, Sapphire R9 Fury, asus x79 ws, HP ZR24w, edifier s730
HP Pro x2 612- i5-4302Y, 8 gigs of memory, 256 ssd
 
Captain Ned
Global Moderator
Posts: 28698
Joined: Wed Jan 16, 2002 7:00 pm
Location: Vermont, USA

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 9:09 am

just brew it! wrote:
Chrispy_ wrote:
They're already in pretty terrible shape after a month.
And *that*, right there, is a pretty darned serious indictment of Samsung's product validation and testing procedures. How the heck could they possibly have missed this?

Are you sure they missed it (says my inner cynic)?
What we have today is way too much pluribus and not enough unum.
 
morphine
TR Staff
Posts: 11600
Joined: Fri Dec 27, 2002 8:51 pm
Location: Portugal (that's next to Spain)

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 11:02 am

Captain Ned wrote:
Are you sure they missed it (says my inner cynic)?

Exactly what I thought. You don't miss things like these. Every tester reports them but the project goes ahead for a release anyway.

See also: EA, Ubisoft,...
There is a fixed amount of intelligence on the planet, and the population keeps growing :(
 
cobalt
Gerbil First Class
Posts: 171
Joined: Mon Oct 30, 2006 11:28 am

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 4:57 pm

Welch wrote:
WOW I think that is the most anemic score I've seen yet. Agree, no reason they couldn't enable 840 Vanilla to be able to run the restore utility at least.

Have you ran DiskFresh before? You've known about this issue for how long now? How often if multiple times, does it take for performance to degrade to that level again?


I've only been loosely tracking the issue, as I took messages about "EVO only" at face value for a while. The machine hasn't seemed as snappy as it used to, but not so bad that I took the time to investigate. Some occasional things that took really long, but could blame it on running in a VM or something else. (And even then, I thought the symptom was 100-200MB/s instead of 400-500MB/sec, which wouldn't be so apparent, so I wasn't thinking it could have been this issue anyway. And some early things I heard that leaving it on, which I do, would prevent it.) It's my home machine, so it's not heavily used anyway.

Never run DiskFresh before, it's been running as my OS drive for about 10 months.

After DiskFresh completed, I'm back at 380MB/s or higher across the whole drive.
 
IAmGhostDog
Gerbil Elite
Posts: 638
Joined: Sun Oct 10, 2004 10:56 am
Location: Milwaukee, Wisconsin

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 8:31 pm

Before & After firmware update/advanced performance optimization
Pretty dramatic. We'll see how long it lasts.

Image
Gigabyte GA-990FXA-UD3 | AMD FX8350 Black Edition | Samsung 840 EVO 500GB | Corsair XMS3 16GB DDR3 | XFX R9 390 | Win7 x64
HT Omega Claro | Technics SL-1210 MKll w/ Stanton 681EEE MKlll | Klipsch Promedia v4.1 | Sennheiser HD 580 Precision
 
cobalt
Gerbil First Class
Posts: 171
Joined: Mon Oct 30, 2006 11:28 am

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 9:10 pm

And here's my "after" result. I gotta say, it's the slow access times that worry me possibly even more than the slow reads. At some point I suspect it will have degraded so far it won't be able to read that data at all. I'm happy that the access times are fixed as well.

Image

As per Chrispy's advice, set DiskFresh to launch once a month.

(And thanks, all, for the moral support and advice!)
 
localhostrulez
Minister of Gerbil Affairs
Posts: 2481
Joined: Sun Mar 09, 2014 11:26 pm

Re: Samsung 840 & 840 EVO Bug

Sat May 02, 2015 9:49 pm

cobalt wrote:
And here's my "after" result. I gotta say, it's the slow access times that worry me possibly even more than the slow reads. At some point I suspect it will have degraded so far it won't be able to read that data at all. I'm happy that the access times are fixed as well.

And viewtopic.php?f=5&t=111578


This, my friends, is why I'm selling my 840 evo and getting a Crucial drive (decided to get an MX200 this time), even though some suggested keeping it for the new laptop.

Though one interesting thing - I recall someone mentioning that the burst rates from hdtune seem a bit funny. Take a look at a result from my HDD: http://i.imgur.com/mMQi19q.png Burst rate is a decent bit slower than the max, though a bit better than the average. Huh? Though this drive is actually pretty decent for speeds. I'd still rather have SSDs for boot drives though, particularly for laptops (physical durability).
 
just brew it!
Administrator
Posts: 54499
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Samsung 840 & 840 EVO Bug

Sun May 03, 2015 8:36 am

Captain Ned wrote:
just brew it! wrote:
Chrispy_ wrote:
They're already in pretty terrible shape after a month.

And *that*, right there, is a pretty darned serious indictment of Samsung's product validation and testing procedures. How the heck could they possibly have missed this?

Are you sure they missed it (says my inner cynic)?

While possible, that would be a pretty stupid move. It is impossible for an issue of this magnitude to go unnoticed for long. And regardless, it would still be an organizational failure for Samsung as a whole; it just shifts more of the blame from the engineers to management.
Nostalgia isn't what it used to be.
 
cobalt
Gerbil First Class
Posts: 171
Joined: Mon Oct 30, 2006 11:28 am

Re: Samsung 840 & 840 EVO Bug

Mon May 04, 2015 9:43 pm

Okay, and I thought my vanilla 840 result was bad? Here's my 840 EVO result before:

Image

... and after:

Image

Holy crap, look at that 'before' result -- most of the drive is less than 5MB/sec?!

WHAT THE HELL, SAMSUNG? HOW THE HOLY FLIPPING PANCAKES DID YOU NOT NOTICE THIS BEFORE RELEASING THIS DRIVE?
 
thecoldanddarkone
Minister of Gerbil Affairs
Posts: 2449
Joined: Wed Mar 26, 2003 4:35 pm

Re: Samsung 840 & 840 EVO Bug

Mon May 04, 2015 10:13 pm

cobalt wrote:
Okay, and I thought my vanilla 840 result was bad? Here's my 840 EVO result before:

Image

... and after:

Image

Holy crap, look at that 'before' result -- most of the drive is less than 5MB/sec?!

WHAT THE HELL, SAMSUNG? HOW THE HOLY FLIPPING PANCAKES DID YOU NOT NOTICE THIS BEFORE RELEASING THIS DRIVE?


Wow. Um just wow.
I7 4930k, 32 GB Ballistix [email protected] , 1.2 TB Intel 750 AIC, 500 GB mx200, Sapphire R9 Fury, asus x79 ws, HP ZR24w, edifier s730
HP Pro x2 612- i5-4302Y, 8 gigs of memory, 256 ssd
 
just brew it!
Administrator
Posts: 54499
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Samsung 840 & 840 EVO Bug

Mon May 04, 2015 11:10 pm

That "after" result is still a little wonky too... WTF is with the weird spikes? Any chance something else was accessing the drive during the "after" test?
Nostalgia isn't what it used to be.
 
localhostrulez
Minister of Gerbil Affairs
Posts: 2481
Joined: Sun Mar 09, 2014 11:26 pm

Re: Samsung 840 & 840 EVO Bug

Mon May 04, 2015 11:57 pm

Yikes... *looks up the speed of a 3.5" floppy disk* And this is WITH the previous firmware they pushed out?

At that point, I have to wonder how close the drive is to loosing the data entirely, like some people were finding/speculating with theirs. Confused error correction left and right I suppose? Makes my spinning rust drive (mass data storage) look insanely fast by comparison. Or any hard drive I have lying around, for that matter. How slowly was that machine booting? I mean, a Windows 7 32-bit VM boots from that spare drive in ~40 seconds (VM's POST is way quicker than a real machine though).

Well, at least these drives seem to be holding plenty of value on ebay. Must be the Samsung name. (seriously)
 
Welch
Grand Gerbil Poohbah
Posts: 3582
Joined: Thu Nov 04, 2004 5:45 pm
Location: Alaska
Contact:

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 3:36 am

But... but... only a small subset of users were reporting the issue. Compared to exactly ZERO people having any issue what-so-ever with their vanilla 840.......

/sigh, Cobalt, if I had any cake to give, you would have taken it with that before shot. God knows you deserve it you poor bastard :lol: .
"I think there is a world market for maybe five computers."
Thomas Watson, chairman of IBM, 1943

1600x | Strix B350-F | CM 240 Lite | 16GB 3200 | RX 580 8GB | 970 EVO | Corsair 400R | Seasonic X 850 | Corsair M95 / K90 | Sennheiser PC37x
 
Topinio
Gerbil Jedi
Posts: 1839
Joined: Mon Jan 12, 2015 9:28 am
Location: London

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 4:33 am

localhostrulez wrote:
Yikes... *looks up the speed of a 3.5" floppy disk* And this is WITH the previous firmware they pushed out?

In which case, the firmware does nothing much, which means without "Magician" running its periodic DiskFresh-alike data refresh you're toast. Brilliant for non-Windows users, that.

localhostrulez wrote:
Well, at least these drives seem to be holding plenty of value on ebay. Must be the Samsung name. (seriously)

Isn't it kinda unethical though, don't think I could inflict it on someone else so mine will just go to being a scratch drive.
Desktop: 750W Snow Silent, X11SAT-F, E3-1270 v5, 32GB ECC, RX 5700 XT, 500GB P1 + 250GB BX100 + 250GB BX100 + 4TB 7E8, XL2730Z + L22e-20
HTPC: X-650, DH67GD, i5-2500K, 4GB, GT 1030, 250GB MX500 + 1.5TB ST1500DL003, KD-43XH9196 + KA220HQ
Laptop: MBP15,2
 
just brew it!
Administrator
Posts: 54499
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 5:45 am

Topinio wrote:
localhostrulez wrote:
Yikes... *looks up the speed of a 3.5" floppy disk* And this is WITH the previous firmware they pushed out?

In which case, the firmware does nothing much, which means without "Magician" running its periodic DiskFresh-alike data refresh you're toast. Brilliant for non-Windows users, that.

Well, the new NEW firmware is supposed to fix that, provided the drive isn't left powered off for extended periods. I guess I will get to find out.
Nostalgia isn't what it used to be.
 
Chrispy_
Maximum Gerbil
Posts: 4670
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 7:48 am

just brew it! wrote:
That "after" result is still a little wonky too... WTF is with the weird spikes? Any chance something else was accessing the drive during the "after" test?

Nah, that looks like the sort of thing you'd expect on plenty of SATA3 drives running over a SATA2 interface. Some of them are capable of capping the interface at an even 235MB/s or so, others try to burst too much data and then dial it back a bit which is what I'm seeing here (and have seen on many drives from numerous vendors). My hunch from witnessing this pattern several times over is that the spikes occur roughly every x gigabytes where x = available free nonpaged memory on your PC at the time of the benchmark. It's probably a windows quirk rather than an SSD quirk.

No matter how bad the 840 vanillas are, Welch's EVO result is a new low, As I've said, the worst I've seen is 9MB/s over the majority of a drive; His is an order of magnitude worse.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
cobalt
Gerbil First Class
Posts: 171
Joined: Mon Oct 30, 2006 11:28 am

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 8:31 am

Chrispy_ wrote:
just brew it! wrote:
That "after" result is still a little wonky too... WTF is with the weird spikes? Any chance something else was accessing the drive during the "after" test?

Nah, that looks like the sort of thing you'd expect on plenty of SATA3 drives running over a SATA2 interface.


Thanks for the SATA3/SATA2 explanation -- that's exactly the setup I've got. This was an old C2Q system, given new life as an HTPC.

(Which means it runs nothing but WMC, reads and writes from a spinning disk, and is never rebooted, so at least I have an excuse for not having noticed anything weird.)

Chrispy_ wrote:
No matter how bad the 840 vanillas are, Welch's EVO result is a new low, As I've said, the worst I've seen is 9MB/s over the majority of a drive; His is an order of magnitude worse.


I think both the 840 and 840 EVO results you're referring to are mine. I'd hate to implicate anyone else in not running DiskFresh before now -- based on what I've seen, it almost feels like neglect at this point! :oops:
 
Chrispy_
Maximum Gerbil
Posts: 4670
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 2:40 pm

cobalt wrote:
I think both the 840 and 840 EVO results you're referring to are mine.

Yep, Dunno why I thought they were Welch's
I've only recently upgraded my Core2 SATA2 HTPC to an Ivy/SATA3 setup but SATA2 was never the bottleneck; An SSD's 0.1ms access time is the thing that actually matters.

What is scary about your HDTune plot is that there are lots of little yellow dots up above 20ms, which makes it slower than the slowest, oldest hard drives that qualify as "SATA" that I can find, even down the back of the "really old junk" drawer at work.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
localhostrulez
Minister of Gerbil Affairs
Posts: 2481
Joined: Sun Mar 09, 2014 11:26 pm

Re: Samsung 840 & 840 EVO Bug

Tue May 05, 2015 2:53 pm

Chrispy_ wrote:
cobalt wrote:
I think both the 840 and 840 EVO results you're referring to are mine.

Yep, Dunno why I thought they were Welch's
I've only recently upgraded my Core2 SATA2 HTPC to an Ivy/SATA3 setup but SATA2 was never the bottleneck; An SSD's 0.1ms access time is the thing that actually matters.

What is scary about your HDTune plot is that there are lots of little yellow dots up above 20ms, which makes it slower than the slowest, oldest hard drives that qualify as "SATA" that I can find, even down the back of the "really old junk" drawer at work.

I've dealt with earlier Core 2 laptops (ex. HP 6510b/6710b) that use earlier SATA drives, and those are hardly any faster than laptops I've seen with IDE. ~40MB/s sequential. Not sure about the access times.
 
Welch
Grand Gerbil Poohbah
Posts: 3582
Joined: Thu Nov 04, 2004 5:45 pm
Location: Alaska
Contact:

Re: Samsung 840 & 840 EVO Bug

Wed May 06, 2015 1:26 am

Probably thinking of me because I keep posting lol, that or it must be my handsome boyish looks....... or not.

Anyhow, Chrispy, great observation about the SATA 2 and SATA 3 bursting "issue". I don't recall if the SATA controller or the drives controller is responsible for link speed negotiations? It would seem this is a simple driver or firmware bug, not necessarily a Windows issue or we should mostly all be seeing the same issue. I've ran lots of these 840 EVOs on SATA 2 interfaces unfortunately and have never seen them break over the 300 MB/s mark. Most of them cap out around 289ish, which leads me to believe that they are being limited to SATA 2 speeds in the first place on drive controller.

Cobalt, what sort of system are you running that on? Could you by chance give us a motherboard/controller model #, and heck, while your at it grab the driver version and date. With the speeds your seeing I'd almost wonder if Chrispy isn't onto something. You said it was an old system, curious how old. Unless the system your running is THAT old... even running on SATA 2 you should be pretty much be close to 270-285ish. It's possible that your SATA controller is just junky and holding back the already funky 840 EVO.
"I think there is a world market for maybe five computers."
Thomas Watson, chairman of IBM, 1943

1600x | Strix B350-F | CM 240 Lite | 16GB 3200 | RX 580 8GB | 970 EVO | Corsair 400R | Seasonic X 850 | Corsair M95 / K90 | Sennheiser PC37x
 
Chrispy_
Maximum Gerbil
Posts: 4670
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Samsung 840 & 840 EVO Bug

Wed May 06, 2015 7:31 am

Welch wrote:
even running on SATA 2 you should be pretty much be close to 270-285ish.


I have found over the years of running SATA2 SSDs that anything over 250MB/s is rare. I couldn't cite the reasons but Indilinx, Sandforce, & Intel drives of the SATA2/SATA3 switchover era would never usually get more than 235-250MB/s depending on the accuracy of the test.

Peak sequential speeds really don't matter that much though, especially on an older SATA2 system. An SSD needs to be able to deliver sequential speeds in the hundreds and access times in the <0.5ms category. Even ancient JMicron devices like the one I posted an HDTune for a while back is perfectly capable of this, desite being (by any modern standard) a terrible SSD. To put it this way, I'd rather have a 50MB/s drive with 0.1ms access times over a 1GB/s drive with 5ms access times for client workloads.

Cobalt's EVO is reporting plenty of access times in the >20ms category. That's clearly the biggest problem here. Not only is it 400x slower than a typical, correctly-functioning SSD, it's also notably slower than even the slowest of spinning disks. Most of us SSD converts lament having to use a laptop with a 5400rpm drive, but even those manage to average 15ms access times. Having an SSD that feels 2-3x slower than that is a real, real-world problem.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
cobalt
Gerbil First Class
Posts: 171
Joined: Mon Oct 30, 2006 11:28 am

Re: Samsung 840 & 840 EVO Bug

Wed May 06, 2015 9:58 am

Let's see -- poking through old ordering records since I'm not in front of the thing, I believe this is the:
Gigabyte GA-EG45M-UD2H. ICH10R, so hopefully not too junky! :D I don't think there's another SATA controller on there; the JMicron 368 onboard only provides IDE.
(Decent board, but just the slightest hint of flakiness in the Realtek LAN that I've stuck with Intel LAN GigE controllers since. And went back to ASUS, too, though I don't have anything against Gigabyte.)
Core 2 Quad Q9550 with 4GB Corsair XMS2 (2x2GB) DDR2-800.
Should be running in AHCI, and NTFS is that matters.

Can't check drivers until tonight -- you're looking for the Intel storage driver version?
  • 1
  • 5
  • 6
  • 7
  • 8
  • 9
  • 11

Who is online

Users browsing this forum: No registered users and 1 guest
GZIP: On