LSI SandForce 5 Series SSD Firmware - TRIM Lost and Found, Performance Investigated

[Guest]

Please feel free to comment about our story entitled "LSI SandForce 5 Series SSD Firmware - TRIM Lost and Found, Performance Investigated"

[parsec]

Re: LSI SandForce 5 Series SSD Firmware - TRIM Lost and Found, Performance Investigated

WOW, I'm the first to post a comment?! This is an important issue... actually two, and I'm not sure which one is really the most significant.

OMGawd, the no TRIM in LSI/SF firmware 5.02 is a major story, that has been kept out of the PC enthusiast news stories and SSD reviews on so many PC Hardware web sites! Extreme salute to Tweak Town for finding and publishing this, your journalistic credibility is very high! All the other PC review web sites, including the well known, big names, that are not reporting this, are now forever suspect. The phrase tip of the iceberg comes to mind.

I wonder if LSI has the clout to intimidate or ask those sites to "do LSI/SF a favor..." and not report this. We all know why they would, SF SSDs with 5.02 based FW would be avoided by enthusiasts. Also, just as SF has gotten some respect back, after the FW issues they had in 2011, they don't need this at all.

Do the other web sites think they are doing the lesser of two evils, for the good of the industry (what consumer?) Since the consumer isn't really damaged when their SSD does not perform as well as it could or should, but the SSD industry would be at least temporarily damaged if this fact was generally released to the public. Think of all the non-enthusiasts that are not up to updating FW, the RMAs will be piling up.

Then there is the FUD factor, where any issue, real of imagined, is amplified by the masses ("... all SF SSDs don't have TRIM!"), which I can sympathize with... to a degree.

Regarding a quick fix for this issue, yes if it was a one or two lines of code thing, then the testing and validation could occur faster, but otherwise, two weeks is a blink of the eye for software validation. Sure, some companies pump out FW all the time, while Intel spent up to a year validating the SF 2281 SSD controller and FW. Intel has custom FW for their SF drives, which is not that unusual, since using purely stock FW is apparently rare. Then again, the Intel FW is mentioned in most of their SF SSD reviews, because it matters.

I can't explain why more people aren't outraged by the bigger issue, the general non-reporting of this issue in the PC enthusiast press, or is that censorship? I haven't decided how I completely feel about this, since we may not have all the information (spin?) about it. But if a comment posted under the article about a manufacture censoring a post about this topic is true, that speaks volumes.

Thanks to Tweak Town for printing this info, you've earned my respect.

[chrisram]

Re: LSI SandForce 5 Series SSD Firmware - TRIM Lost and Found, Performance Investigated

I'm right there with you on everything you said except for the part about LSI or SandForce strong arming reviewers.

I talk with LSI SF reps at least once a week and they never once asked TweakTown to sweep it under the rug or to look away. They were actually able to assist us figure out what was going on even though they knew the findings would give them a black eye. They took it like a man and will let the wound heal.

I can say that it wasn't just a line of code, the 5 Series is a big jump from the older 3 Series and it was a complicated fix but it's fixed other than an additional fix that some OEMs requested, that is why most companies are waiting for 5.0.4.

[parsec]

Re: LSI SandForce 5 Series SSD Firmware - TRIM Lost and Found, Performance Investigated

Thanks for your reply. It's great to know that LSI/SF was completely cooperative with you, not a surprise really. You also mentioned Corsair provided you with an updated firmware for testing.

While considering why only you and a very few others have written about this topic, the conspiracy theorist side of me came up with the please don't mention it idea. My impression is you certainly are more objective in your reviews and commentary, than some of the other PC hardware review web sites, and are less likely to be swayed by strong arm tactics.

I've read both SanDisk Ex reviews (and own one), and am aware the 5.0.3 firmware is based more so on the version 3 firmware than 5.0.2. As a programmer, it bothers me when people demand software fixes to be released overnight, and can't understand why it takes "so long". Intel's verification of the LSI/SandForce controller and firmware is an example, albeit extreme, of how much time it can take to do the job right. Yet the AES 256bit encryption bug was not found during that time, another perfect example of how difficult verification is.

So it seems it all boils down to the other sites not wanting to reference Tweak Town in their pages. That's a statement in itself. I can't wait to see what they write when the firmware is released. Keep up the good work!

[chrisram]

Re: LSI SandForce 5 Series SSD Firmware - TRIM Lost and Found, Performance Investigated

Hello,

I'm sitting down to write the SanDisk Extreme 240GB on 5.0.3 to simulate the upcoming 5.0.4 public release FW.

As for the other sites linking to us. That has more to do with our current NVIDIA product launch reviews...or I guess we can call them prelaunch since we are writing them up a month before other sites even know the products are coming. I wouldn't link to our SSD reviews either if I were them, we are so far ahead of what everyone else is doing right now, it just makes them look silly. We are getting ready to change the format a bit too and when that happens it'll be game over because we are taking real world testing to a whole new level.