OCZ/SF FW2.09 Thottles SF2281 Controller

Toggle sidebar Toggle sidebar
E

e-drood

Member
Jun 15, 2011
169
0
0
Sandforce could only minimize instability in Series 2000 SSD Controllers by "Thottling" to overcome marchitectural issues...

This is not merely disappointing, but also very irresponsible by Sandforce to release to SSD Resellers/ODMs non-compliant lsi... Sata 2.6 was ratified over (2) years ago and always public domain.

What is the "new" excuse offered by Sandforce for foisting this half-baked incompetent controller chip on the global market? CAN THERE BE ANY MITIGATION?
 
M

MrWheezy

Junior Member
Jun 6, 2011
13
0
0
E-drood, is Sandforce the sole problem here?

Couldn't it be the execution of the controller by these companies? I mean OCZ isn't known for it's reliability or their five star customer service. While other companies like G.Skill or OWC haven't faced any problems (yet?).
 
E

e-drood

Member
Jun 15, 2011
169
0
0
I am also thinking about this...

G.Skil & OWC haven't a large user base w/sf2281 ssd's, however Corsair has shipped moderate inventory & has reasonable number of users holding sf2281 ssd corsair label. I anticipate that if Corsair, following "recall" adding resistors to pcb & updating firmware, may have quietly thottled sf controller in "returned user ssd's...

yes it can also be pcb induced cross talk or timing errors, but this is observable/recordable as corrupted data stream

it appears to be sf controller marchitecture design philosophy --- wafer fab / lsi dice / "photo" mask & deposition composition/layering is an art, only partially based on hard science - one learns through extremely costly experience of failed fab runs, "how to design lsi chips" --- I would wager that the contract fab (employed by sf) is already covering itself by pointing finger at sf design as flawed.

please understand that "Thottling" to salvage the present production run of SF2281 SSD Nand Controllers ONLY REDUCES No's & Frequency of Stuttering / BSOD EVENTS Experienced by Users To An "Acceptable" Level of User Demands For Refund/RMA.

The SANDFORCE SF2281 Nand Controller CAN NEVER BE COMPLETELY RELIABLE.

I am very troubled by the Sandforce & OCZ focus on the Enterprise SSD Market - We are talking truly "Mission Critical" Air Traffic Control Systems, Hospitals, Police, Ambulance & other AND LongTerm Personal Records Storage - YOUR Insurance, Healthcare, Property Records are electronic today,

Sandforce management/ownership Does Not Have Any Committment To User Base And is Bereft of Any Sense of Social Responsibility.

First SF1200, SF2281 BOTH failing to meet specifications...
 
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
blah, blah, blah.. more "self proclaimed expert" speculation without any quantifiable links. You fit right in around this joint.

It's also funny how we went from "sf controller chip thermal issues" to.. "Thottling to overcome marchitectural issues"? Talk about your 1-2 steps here. lol

and I don't suppose you have a link to the meaning of "marchitectural" do you? Makes it even harder to follow along with your overly derogatory thought process here when you start using words that aren't even in the dictionary, eh?
 
E

e-drood

Member
Jun 15, 2011
169
0
0
Sandforce via Firmware 2.09 has imposed Thottling in their SF2281 SSD/NAND Controller to minimize Stuttering & BSOD Events resulting from flawed logic, pathway and thermal design considerations in a desparate attempt to avoid recalling all initial production SF2281 controller chips.

Purchasers of OCZ SSD's having Sandforce SF2281 Controllers should be aware of the now mandated performance specification reduction of this product range and the continuing possibility of random Stuttering & BSOD Events resulting in unpredictable loss of data.
 
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
they should?
what if they didn't have problems to begin with?
Should they still worry?
should I and others who have been successfully using them for a year and half without issue begin to worry with all this new info you've given us?
are you aware that not all are seeing any form of "throttling" as you so dramatically state?
 
E

e-drood

Member
Jun 15, 2011
169
0
0
distracting & misdirecting

this is 2nd ssd flawed controller from Sandforce, SF1200 & now SF2281, both failing to meet design spec, but this latest fumble is really beyond any explanation

Sandforce is not "gaining" and building on experience, rather forgetting earlier experience while making yet more mistakes... no possible explanation for this one
 
M

MrWheezy

Junior Member
Jun 6, 2011
13
0
0
It should also be noted that OCZ gets a custom firmware from SandForce. While Corsairs issue was a PCB design issue, I've heard an OWC Rep tell me that OCZs issue is due to firmware issues resulting from them over tweaking it and poor quality NAND.
 
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
much better post there and I can hang with that.

this is 2nd ssd flawed controller from Sandforce, SF1200 & now SF2281, both failing to meet design spec, but this latest fumble is really beyond any explanation
Click to expand...
COMPLETELY agree with you there about the second gens being flawed out of the gate again. Not sure where/what you're going on about the "failed design spec" and I doubt you have much more than a speculatory assumption in regards to the actual "specs/design" of that chip. If you did?.. you'd not be wasting your time here and getting a big piece of that action with lips tighter than a frog's ass to avoid losing your latest bonus and stock option. lol
Sandforce is not "gaining" and building on experience, rather forgetting earlier experience while making yet more mistakes... no possible explanation for this one
Click to expand...

couldn't disagree more there. I and many others bitched and bitched and even bitched some more about the overly aggressive Duawrite throttling algorithms and slowish recovery(their GC/internal block consolidation/data rotation is far better than many, if not all others, though) and guess what they did?... Increased the recycling engine's size to help speed low activity recovery and on the fly performance maintenance with much more limited throttling with the added ability to temporarily lift throttles when throughput is needed most. If you did in fact own one of these new controllers and did in fact have an older one to compare it to?.. you would appreciate the strides made in those departments.

Speaking of "distracting & misdirecting".. we're still waiting on those screenshot's by the way(I know by now that you've seen mine before as you trolled around here and over there). Your roll, nub
 
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
MrWheezy said:
It should also be noted that OCZ gets a custom firmware from SandForce. While Corsairs issue was a PCB design issue, I've heard an OWC Rep tell me that OCZs issue is due to firmware issues resulting from them over tweaking it and poor quality NAND.
Click to expand...

what!!?.. an OWC rep said that? that's really odd that they would say such a thing. Their advertsing integrity has always been known to be top notch. LOL

That place is a hole in the wall compared to what OCZ puts out for volume. Of course they want a piece of that action and will do whatever it takes to get it. Big money in these SSD's, you know.

Did you know that they made the silent switch to 25nm nand too?

Do you also realize that the very forum environment that OCZ maintains is in total reverse of what OWC allows? I would hardly call a blog with a couple hundred posts an environment that would give any real clear or concise picture as to their actual problems or resolutions stemming from them. Talk about swallowing limited info/spin advertising.. hook, line and sinker there. :sneaky:
 
E

e-drood

Member
Jun 15, 2011
169
0
0
pls see http://techinsidr.com "Is A Recall Brewing At OCZ" now that performance has been reduced in Vertex3 & Vertex3MaxIOPS SSD's, being an over-the-counter listed company can be a liability. Sandforce is a privately held company & hides away during all the sf-2281 user problems. A big FAIL by both SF & OCZ!

RECALL SSD's With Sandforce Controllers!
 
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
LOL.. NEVER gonna happen.. drood'y.

It's also rather funny that all the revamped Corsair drives going back into the channel are still having similar issues. Wanna know why?.. no new firmware revisions as of yet. See a trend there?

It's even funnt that that link there still lumps the Corsair PCB rework recall into the same issue as the firmware/compatibility problems. Seperate issues entirely. Especially for those keen enough to wonder why the other models using the EXACT same nand and controllers have NOT been recalled.

So shout and muck about on the forums all you want about these controllers. They are not going to be going anywhere.

Eventually they'll be some public info released as to some of the other direct issues related to the underlying issues at play here(some pretty big names aside form Sandforce and OCZ) and threads like this will look even more foolish. But for now firmware will be tweaked to work around the other issues at play and then eventually they'll all play ball together once Intel releases their Sandforce based drives. As usual Intel will create the trickle down effect for others to enjoy.
 
C

curlysir

Member
Feb 21, 2011
43
0
0
I have a Vertex 3 Max with the new Firmware. The throttling in my case is less then 4% on the ATTO benckmark. I can now use the drive as a boot drive in my hot swap bay whereas I could not before. The slight reduction is speed is worth the stability and functionality. In my opinion anyway.
 
Yellowbeard

Yellowbeard

Golden Member
Sep 9, 2003
1,542
2
0
curlysir said:
I have a Vertex 3 Max with the new Firmware. The throttling in my case is less then 4% on the ATTO benckmark. I can now use the drive as a boot drive in my hot swap bay whereas I could not before. The slight reduction is speed is worth the stability and functionality. In my opinion anyway.
Click to expand...

Interesting, I for 1 appreciate your post. SandForce and their new drives are taking a LOT of criticism but, this is the bleeding edge.
 
E

e-drood

Member
Jun 15, 2011
169
0
0
Dirk, thanks for adding this... unrecoverable thottle link

Curlysir, briefly, which motherboard & is sata3 default?

my posts have a focus - your posts have no relevent focus
 
C

curlysir

Member
Feb 21, 2011
43
0
0
Yellowbeard said:
Interesting, I for 1 appreciate your post. SandForce and their new drives are taking a LOT of criticism but, this is the bleeding edge.
Click to expand...

I would hope so, as it is one of your cases (800D) hot swap bays that I couldn't use until now with the Vertex 3 @ 6Gbps :D
 
Last edited:
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
DirkGently1 said:
What about the lifetime throttle. Hit the drive too hard for too long and you get an unrecoverable throttle.

http://www.xtremesystems.org/forums...25nm-Vs-34nm&p=4887914&viewfull=1#post4887914

http://www.xtremesystems.org/forums...25nm-Vs-34nm&p=4867450&viewfull=1#post4867450

http://www.xtremesystems.org/forums...25nm-Vs-34nm&p=4868047&viewfull=1#post4868047
Click to expand...

I plan on going back over there to help them understand a bit better but haven't had time to dig screens up for demonstartion purposes.

There are various reasons that a secure erase may not clear the drive and even some latest versions of the toolbox for those Sandforce drives has caused issues with incomplete SE's being done on some systems.

So to be clear here.. there is NO such thing as a unrecoverable throttled state. SE does in fact clear the Durawrite maps for fresh speed. I know for fact because I actually do it on a regular basis to avoid throttled state's simply because I don't want to lose 110MB/s write speeds when moving highly compressed video to my array.

Tony will tell you the same thing but will be quick to point out that inevitably it will "settle back down to a normal daya to day state". That's the true implementation of a lifetime throttle and nets around 70% of the original fresh speed with incompressible data. You cannot see it when using more easily compressed data and(for example) ATTO will not even show its existance.

and as I asked prior as well.. I too would like to know the meaning of "marchitectural".

ohhh.. still waiting for proof that the OP even owns the drives he states he has so much firsthand knowledge about. Could even become a big scandal like an anti-abortionist politician who actually had an abortion. Something stinks here and I know it's just more than my attitude towards this OP's views/posts.


Please keep the inflammatory rhetoric dialed down just a smidgen...this is the Memory & Storage technical forum, if you want to invoke politics and hot-topics like abortion into your posts then you need to take it to Politics and News.

Re:
Could even become a big scandal like an anti-abortionist politician who actually had an abortion
Click to expand...
Idontcare
Super Mod
 
Last edited by a moderator:
E

e-drood

Member
Jun 15, 2011
169
0
0
after midnight, saturday, here

note & desk keyboards have cn radical & modifier & english letters on key faces - user select keys default cn & will not support greek alphabet/scientific symbols and ms office 2010 is cumbersome to use to insert greek alphabet into anandtech post

and so I wrote marchiture instead of mu-arch which is recognized term identifying (originally) cpu complex structure

this is a compliment to Sandforce to attempt an ssd controller of significant complexity and nearly succeed; as their work-in-process, further refinement may ultimately be completely compatible with sata 2.6 (6.0Gbps) interfaces, until that day only sata2 (3.0Gbps) interfaces appear to provide reliable data transfer.
 
G

groberts101

Golden Member
Mar 17, 2011
1,390
0
0
now that's funny! Especially considering that the problems they are seeing are also occurring on sata 2 interfaces.

They don't work and are unreliable?.. but yet.. they are? LOL
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom