New Indilinx firmwares coming with up to 500% performance?

[bradley]

Let's celebrate the six month anniversary of OCZ's press release regarding new firmware that "Supports 2xnm NAND Flash and Increases Performance of Indilinx Controllers by up to 500%" Do you remember where you were on the day of the announcement?

 

[RU482]

Let's celebrate the six month anniversary of OCZ's press release regarding new firmware that "Supports 2xnm NAND Flash and Increases Performance of Indilinx Controllers by up to 500%" Do you remember where you were on the day of the announcement?

I remember reading it, thinking "********"


Profanity is not allowed in the technical forums.

Do not use our Forums to post any material, or links to any material, which is knowingly false and/or defamatory, inaccurate, abusive, vulgar, hateful, harassing, obscene, profane, sexually oriented, threatening, invasive of a person's privacy, or otherwise violative of any law. Special exception to the restrictions on vulgarity and profanity are granted ONLY in the social forums.

AnandTech Forum Guidelines

Administrator Idontcare

 

[Xpage]

I Loled too much on "Do you remember where you were on the day of the announcement?"

Would have been a nice boost for my gen 1 ssd

 

[groberts101]

Not really ******** or even that hard to believe at all when you realize how low the small randoms of the original firmware was in the first place. Going from 2MB/s to 10MB/s might not seem like much but at least it's better than the original and still beats the hell out of HDD.

It's just an entry level drive anyways. so who really cares. I'll just be using all mine for raided data storage/scratch disks anyways... so big whoop. Much rather have 500% increases in sequential file performance. lol


Profanity is not allowed in the technical forums.

Do not use our Forums to post any material, or links to any material, which is knowingly false and/or defamatory, inaccurate, abusive, vulgar, hateful, harassing, obscene, profane, sexually oriented, threatening, invasive of a person's privacy, or otherwise violative of any law. Special exception to the restrictions on vulgarity and profanity are granted ONLY in the social forums.

AnandTech Forum Guidelines

Administrator Idontcare

 

[Coup27]

I'm not trying to ignite another full on SF/OCZ flame thread here but I will chip in and say that all the SSD players have had the same set of rules regarding the hardware users own. If you are correct and there are errors in the implementation of various features on Intel chipsets or Windows or whatever else, Crucial/Micron, Samsung & Toshiba (and obv Intel) have all successfully navigated around these problems to produce generally reliable and compatible drive. The fact that SF has caught such a cold on two generations of hardware could possibly be blamed on the big boys, but more they must accept the blame themselves for not being savvy enough to navigate around these problems like the previously mentioned companies did.

 

[groberts101]

very valid point and I would agree almost completely.. well.. for the most part anyways.

Want to know why the Intel(and others) stuff works so well?.. they chose not to utilize the very power mgmt related specs that they pushed so hard for in the first place.

Have some light reading if you're interested?.. but will have to link it later on since i'm on another system right now without access to them at the moment.

 

[Engineer]

LOL...old post (and OP) is old. Sorry guys, just posted what I saw at OCZ and never dreamed it wouldn't materialize. My view of OCZ has fallen greatly over the last year.

 

[Coup27]

Yeah mate send it over. I would like to read it. The LPM bugs you mention, do they effect the H67 range as well as P67 and Z68?

 

[VirtualLarry]

Indilinx - 100% slower, after a week of usage.

http://forums.anandtech.com/showthread.php?t=2201051

 

[Emulex]

you essentially sacrifice lifespan with max iops - read the difference between intel 710 and 320 - same controller, same nand (cherry picked) and heavy overprovisioning and slowing down the write speed by 10-20x.

so if you took a 400gb 320 and hacked the firmware from the 710 over - you'd probably get nearly the same life (minus the poorer quality nand).

intel 710 = A grade cherry picked nand
intel 320 = A- grade nand
(other companies buying intel nand = B grade junk)

 

[taltamir]

At first it sounded like bull... 500% increase? pull the other one! Maybe if all competitors were getting 500% faster performance on same spec hardware due to a serious bug in their firmware...

The revolutionary new Arowana FTL enhances performance of Indilinx controllers with HyperQueuing, which significantly increases sequential write speeds and random IOPS over the previous generation FTL

Reading this, it sounds like they are using some form of agressive command queing. Like TCQ or NCQ. Which effectively allows the drive to convert random writes into sequential writes. Which makes the 500% improvement believable. And if anything a conservative estimate. If anything it sounds like a very conservative figure for such a tech

 

[Coup27]

and hacked the firmware from the 710 over

Good luck with that.

(other companies buying intel nand = B grade junk)

Do you really think that IMFT would have customers like OCZ, Corsair and the likes if they were shipping them vastly inferior NAND to what they use in their 320? The fact that MLC-HET only lasts 3 months after use and not 12 says straight away that it is not just 'cherry picked' regular NAND.

 

[taltamir]

The fact that MLC-HET only lasts 3 months after use and not 12 says straight away that it is not just 'cherry picked' regular NAND.

isn't it 20% larger die due to being physically different and redesigned MLC? Achieving MLC with longer lifespan? That is certainly not "binned grade A vs grade B junk" case.

 

[groberts101]

Yeah mate send it over. I would like to read it. The LPM bugs you mention, do they effect the H67 range as well as P67 and Z68?

Will send it via PM. Be patient though as I have a pretty full weekend of mechanic'in and a B-Day party going on. Should get the links bundled up and sent out in the next few days.

And just to end the question of whether or not this new firmware delivers that increase of small file performance?.. I do actually know firsthand that it does. Unfortunately.. I'm not yet allowed to say how I know that to be fact. lol

Either way.. I'm sure the speculation and opinion will continue.

 

[bradley]

Ocz gets a ton of grief on this forum, and elsewhere -- deservedly so. I just wanted to give credit for their fulfilling a pretty large promise.

Yes, ten months ago, they promised Arowana firmware for first generation drives when it was nowhere close to finalization. Though better late than never.... and never almost came and went.

Indilinx actually did a decent tweaking the firmware, especially access times and writes. More space is being provisioned, but hardly a big deal. The Solid 2 seems to bench markedly faster than the average Vertex with either 1.7 or 3.55 firmware.

 

[VirtualLarry]

Is this new firmware available for the OCZ Agility 30GB Indilinx Barefoot drives?

I have a few of them, and while they are SSDs, they don't have the greatest performance. WEI gives them a 5.9, same as a HD.

 

[exdeath]

And I doubt your proposal of hacking the firmware in such a way is possible. Disabling some function calls, changing return values and so on - sure that's easy. But pretty much changing the inner workings of a complex code with only the assembly (and then that's no detailed documented asm like for x86,..) available seems pretty impossible.

Doubtful it's proprietary. Doesn't Sandforce use ARM or something?

Undocumented, memory map, I/O registers, packet formats, etc on the other hand would take some time to reverse engineer. The code itself is never the hard part. It's about understanding what that code DOES.

 

[exdeath]

you essentially sacrifice lifespan with max iops

I think I'm good with the estimated 9 years left on my Chronos Deluxe and Wildfires. They will be replaced in less than 2 as soon as something better than SATA 6G arrives.