OCZ/SF FW2.09 Thottles SF2281 Controller

[e-drood]

thank you for finally admitting the sata2 problems, in addition to the continuing sata3 incompatibility issues.

it's refreshing to read that OCZ's volunteer field support finally admits sf-2281 sata interface problems

many thanks

 

[e-drood]

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

thanks for your informed confirmation of sandforce sf-2281 sata2/sata3 interface compatibility problems

I appreciate how difficult it has been for OCZ staff to field the continuing complaints about sandforce sf-2281 based ssd's compatibility problems

 

[groberts101]

finally admiting?.. one only needs to drop by the OCZ support forums about a week ago to see the posts that some are making about the few issues there are with sata2 setups. Not any big cover up or anything. If posts get deleted or moved to the complaints section?.. well.. they were probably just being contraversial pricks like some others who will go un-named(eh hmm).

Notice however that I didn't make any "ALL sata 2 interfaces" blanket statements. It's not an "all or nothing" situation and as I've already said many times there is far more to it than what some are quick to believe. far more involved here than just Sandforce, that's for sure. Things will get fixed by the various players involved and life will go on without a PCB recall like some like to lump OCZ's problems with.

 

[e-drood]

OCZ/Sandforce Firmware Update FW-2.09 HAS NOT resolved continuing SF-2281 Controller Sata Interface issues. Some users are now "downgrading" to previous firmware FW-2.06 and living with random bsod's... instead of drive slowdown...

The SATA 3.0Gbps / 6.0Gbps Interface resides in silicon, NOT firmware - the FIRMWARE Operational Instruction Set merely sets operational parameters (upper & lower limits within limited range)

Stating that ALL Other Players, ie: Chipset & Motherboard Firms are responsible for continuing compatability issues of SF-2281 is simply untrue and serves to deflect the responsibility from Sandforce. The Serial Interface Standards HAVE ALWAYS BEEN Public Domain and will not be changed to accomodate Sandforce.

The Sandforce 2000 Series Controllers are either compatible with "Reference Compliant Sata 3.0Gbps & 6.0Gbps Interfaces" widely found on many motherboards OR Sandforce must recall these non-compliant controller chips.

Six Weeks & firmware still needs further revision - NO ONE ACCEPTS RANDOM BSOD's As Price To Pay For Bleeding Edge Promises of Performance...

 

[Makaveli]

These issues are a great example of why you don't let benchmark numbers control your purchasing decisions!

 

[AdamK47]

No problems here with my Vertex 3 120GB using the initial firmware connected to the Marvell 9123 controller on my Asus P6X58D Premium. I guess I'm lucky. The controller isn't as fast as the Intel SATA-3 interface, but it is faster than the Intel SATA-2 interface.

 

[grimpr]

Running a Vertex 3 120GB, 24/7, for 2 months now on a ASUS M4A89GTD PRO motherboard with an AMD SB850 southbridge, updated to latest bios and Win7 SP1 with the new and updated MSAHCI driver. Runs like champ, TRIM & idle time Garbage Collection works and the performance stays top notch.

 

[e-drood]

I give OCZ "Genuine Heroic Effort" (sincerely, yes)

http://www.ocztechnologyforum.com/forum/showthread.php?90316-Intel-management

The Sandforce SF-2281 SSD Controller appears to require a very narrow range of Sata3 (6.0Gbps) interface + bios + (total) driver feature sets to function --- when other hdd's/ssd's are far more tolerent of "motherboard/interface settings/values"

and suggesting (by ocz) that ("normal") overclocking is discouraged (and I ALSO DISCOURAGE GROSS EXTREME OVERVOLT O.C.'g) is unacceptable to most users of P67/Z68 chipsets

Again I give genuine HighMarks to OCZ --- but the "solution" has become excessively complex --- perhaps a quiet word w/sf is urgently needed at this point --- enterprise product credibility is on the line, here...

if a consumer product line can't be properly supported, the 7/24 mission critical enterprise market is beyond reach...

 

[HendrixFan]

Enterprise markets won't be overclocking. What on earth are you talking about?

 

[e-drood]

I give OCZ "Genuine Heroic Effort" (sincerely, yes)

http://www.ocztechnologyforum.com/forum/showthread.php?90316-Intel-management

The Sandforce SF-2281 SSD Controller appears to require a very narrow range of Sata3 (6.0Gbps) interface + bios + (total) driver feature sets to function --- when other hdd's/ssd's are far more tolerent of "motherboard/interface settings/values"

and suggesting (by ocz) that ("normal") overclocking is discouraged (and I ALSO DISCOURAGE GROSS EXTREME OVERVOLT O.C.'g) is unacceptable to most users of P67/Z68 chipsets

Again I give genuine HighMarks to OCZ --- but the "solution" has become excessively complex --- perhaps a quiet word w/sf is urgently needed at this point --- enterprise product credibility is on the line, here...

if a consumer product line can't be properly supported, the 7/24 mission critical enterprise market is beyond reach...

Last Para --- the real money is in the enterprise market & the present v3/v3m consumer intro isn't a thing of beauty at this point...

 

[HendrixFan]

You just said the problem was overclocking. How does that affect enterprise users?

 

[e-drood]

hendrixfan - I Did Not Write that the consumer user problem is o.c.'g - I Did Not Write that enterprise users o.c. enterprise platforms - You have expressed those opinions...

the rollout of the sf-2281 is "fraught" without a real solution presently on the boards...

the rollout of the sf 2500 series enterprise ssd controllers is presently occuring AND BY ASSOCIATION many potential enterprise users are wondering what kind of support to expect from ocz & sf...

support by enterprise definition goes far beyond consumer rma's...

various dedicated enterprise admin / user communities discuss system components, w/ emphasis on hardware reliability and availability of 7/24 support... and there are recognized established hardware mfr's having well earned credibility

 

[Makaveli]

hendrixfan - I Did Not Write that the consumer user problem is o.c.'g - I Did Not Write that enterprise users o.c. enterprise platforms - You have expressed those opinions...

the rollout of the sf-2281 is "fraught" without a real solution presently on the boards...

the rollout of the sf 2500 series enterprise ssd controllers is presently occuring AND BY ASSOCIATION many potential enterprise users are wondering what kind of support to expect from ocz & sf...

support by enterprise definition goes far beyond consumer rma's...

various dedicated enterprise admin / user communities discuss system components, w/ emphasis on hardware reliability and availability of 7/24 support... and there are recognized established hardware mfr's having well earned credibility

You didn't directly say it but to anyone reading your post it would seem you are implying so.

You shouldn't have said anything about overclocking if your post was ment about enterprise drives.

 

[HendrixFan]

Every post you have just rambles around, literally without even sentences to hold thoughts together. Paragraphs that actually delve into the subject would be an even bigger plus.

As it is I am stuck trying to decode your "snippets" of thoughts.

Going back to the OP, the firmware does throttle but the dropoff in performance seems to be no greater than 5%, and that isn't an across the board dropoff either. This seems to have provided stability for chipsets that had trouble supporting the specs.

Do you disagree with any of that? I don't see how any of that is a big problem, certainly for enterprise users.

 

[e-drood]

hendrixfan and makaveli i completely agree o.c.'g irrelevent & my bad & distracting in my post

hendrixfan enterprise platforms are inherently fault tolerent while consumer level generally is not (different os + different hardware) - ssd random fault is different issue - fault tolerent platforms can actually amplify no's of (ssd) random faults if (ssd) feature set is not supported

candidly ocz is attempting to enlarge it's specialist consumer ssd market niche (of merely 3yrs) into enterprise storage provider - using startup sandforce (of merely 4yrs) "anticipated" improved performance ssd controllers - no provider maturity anywhere in this niche provider relationship

this is not a criticism - this is a statement of fact

candidly sf is bundling presently immature user features into new release controllers without exhaustive field/in the wild/realworld user experience

predictive analysis studies, by sf, of worst case faults, is not the same as multi user feedback...

 

[e-drood]

makaveli, i support reasonable/moderate o.c.'g of consumer m.b./cpu's as widespread mainstream activity that results in stable functional productive pc's & workstations

hendrixfan, the f.w. imposed thottle aside (yes reduced instability) how is raise actually performing/user experience

 

[e-drood]

hendrixfan, re: thottle + other issues sf-2281 controller, briefly:

current "design" mb having p67/b3 or z68/initial stepping, of above average build quality represents sata3 native support

if sf 2281 ssd is installed to port1/sata3/intel we see random user events/issues

this motherboard is fully compliant & has adequate power supply resources

this motherboard represents a dynamic data environment in which various data access & bandwidth resources are frequently reallocated

the sf 2281 ssd must constantly poll & incrementally adjust/reallocate it's controller resources --- this is where the issues originate

the motherboard resources are adequate, but dynamic - the ssd controller must be capable of recognition & response to dynamic nature of available mb resources...

this is not intented as a criticism - applying thottle recognizes this reality, but will not fully resolve