http://www.crucial.com/support/firmware.aspx Crucial M4 SSD - Release Date: 08/25/2011 Change Log: •Changes made in version 0002 (m4 can be updated to revision 0009 directly from either revision 0001 or 0002) •Improved throughput performance. •Increase in PCMark Vantage benchmark score, resulting in improved user experience in most operating systems. •Improved write latency for better performance under heavy write workloads. •Faster boot up times. •Improved compatibility with latest chipsets. •Compensation for SATA speed negotiation issues between some SATA-II chipsets and the SATA-III device. •Improvement for intermittent failures in cold boot up related to some specific host systems.
Yep installed. I have that cold boot issue, as i always turn my pc off at mains when not using it. I did notice that if installing firmware on current usable drive, it takes a while before speeds were corrected/ maybe a trim reorganise or something? No increase in boot up time for me though, read speed marginally faster, but im limited by my x58 chipset anyway to 380mbs.
how do you install the firmware? I take it that if I wish to raid m4s they need to be the same version
They don't *have* to be the same f/w, but it helps... What you normally have to do (other than with the SFs where you need an alt boot mechanism) is to download the (bootable) f/w & put it onto a bootable usb stick or CD/DVD... Reboot - go into the bios (or add on card settings) - & set the controller it's on to ide... Reboot (again) & use the bootable thing to install the f/w... & then reboot (for a 3rd time) - go back into the bios (or add on card settings) - & set the controller it's on to ahci or raid (whichever it was on/you prefer)... Otherwise, some add on cards won't let you do this (or rather the f/w updating program cannot see the drives) so you may have to use the mobo's ports temporarily... ...& make sure you back up beforehand just in case. Oh... There's actually instructions next to where you download the f/w... it basically says what i've written - other than some (imho) rubbish about the SSD having to be on the 1st port... Assuming it's like other f/ws for HDDs or (non-SF) SSDs, you 'should' be able to upgrade more than one at once...
OMG. I cant believe the performance increase on seq reads, just from a firmware update. The main thing for me is the cold boot issue has been fixed, i ordered another drive as i thought mine was dying again, but the firmware has fixed it. seen reports and pics that show seq going from 420mbs to a staggering 510mbs, just from a firmware upgrade. Big win for Crucial on price and performance. http://wccftech.com/crucial-m4-0009-firmware-update-increases-speeds-20/
That's a very decent improvement... ...though i'd like to see some no's from Anand as he's the king of SSDs... ...however, accepting that they're kind of correct, just to look at the marginally -ve side of things, whilst the M4 & intel 510 use the same controller (so the previous difference is just down to f/w), it's important to remember that the former uses 25nm Micron NAND, whilst the latter uses 34nm intel NAND... in terms of longevity, this makes no odds when it comes to read speeds but, alongside the 34nm nand generally having a greater longevity, there's also a limitation with nand that as you increase the write speeds, the longevity falls - & so it's also important to note that there is an improvement in write speeds & the effect that this can have. [NB this is the key reason why SFs compress data so that less nand is affected at faster write speeds.] Having said all of that, looking at the actual figures, it also has to be noted that the increases aren't 'that' huge on the writes - in ASS SSD, you're looking at a 7.7% increase in QD1 4Ks, 5% in QD64 4Ks & a 5.6% increase in sequentials... - & that 'if' you're using the SSD as a OS/apps drive without downloading all of the internet to it (or using it heavily as a scratch drive, for converting media, running the world of b/ms, etc) then, whilst there will be a cumulative effect, it shouldn't be one that materially alters anything within a 3 year warranty period... ...however, as the warranty periods for all SSDs are conditional on use, along with accepting that you will invalidate the warranty 'if' you over use them anyway (hence why there's enterprise drives with superior SLC or eMLC nand), 'if' you have a very heavy r-e-w cycle usage (esp if you don't leave a decent amount of free space & do not consider increasing the OP) then this increase is more likely to have a noticeable -ve effect. [NB this is the reason for the Duraclass function in the SFs - with excessive cycles, the controller temporarily slows the write speeds down to protect longevity... Whilst i can't instantly find it, OCZ have previously stated that they 'could' release a f/w which disabled this function completely, however it could reduce the lifespan down to a matter of months depending on the usage - which kind of shows how important write speeds are important to longevity. This is separate to the rebooting to regain speeds (with reasonably large cycles) with the 6Gb/s ones as that's down to when Win7 passes on certain commands to the SSD controller.] Anyway, i'm *really* not saying that this new f/w is disastrous & shouldn't be used (as, not forgetting the cold boot fix, provisionally it's great news for the vast majority of users who want a more budget orientated SSD - & 'should' be another step in helping to drive down the prices of the better SSDs) - simply pointing out that the SSD market is differentiated, so you need to think about your usage vs the tech used in making a decision... Well, not everything can be solely down to trying to save pennies & this improved f/w doesn't prevent the intel 510 & the high end (non-asynchonous nand) SFs being generally superior consumer drives.
Nice info there. Yes the cold boot issue was the main one for me, as (touch wood) i have no more detection issues on cold boot now. I think as you say, to drive prices down is only going to be a good thing. I would of said that people not having money to spend on luxury items would more likely be what drives the prices down, but it appears theres still more spendable income on pc gear than i can believe. Maybe people stop eating and paying the rent, before they will give up on new pc tech?
No problem at all... Yeah, the theory is that all of the SSD controller makers will shift to some form of compression over time d.t. the die shrinks reducing longevity. Perhaps it's that people look to alt ways to improve their current systems than replacing the whole thing? Well, spending ~£85-£210 on a 60-128GB SSD is far cheaper than replacing everything wholesale & will have a significant benefit... Just a half-arsed theory though. Otherwise, something i missed out (though have generally mentioned in the past) is that the rating of any nand in r-e-w (or p/e - means the same thing) cycles is based upon the likelihood of x% of cells (& hence blocks) failing within a given no of cycles at a given speed... ...now certainly this used to be ~3% across the board but, whilst it's not going to be that dissimilar, afaik, there's a bit of variance between the validation that different nand manufacturers use. & obviously (as i explained it in the previous post) the rated cycles will be for a given write speed - so if you increase the write speeds on a specific SSD then you reduce the no of cycles before a given %age failure rate. [NB this doesn't mean that faster SSDs are necessarily more prone to nand failure as, for example, the type & quality of nand can vary & the no of nand channels & devices per channel can be increased without altering the write speed per nand device... ...simply that if you have a specific SSD & the write speeds are increased then this will have a -ve effect upon the longevity.] So, just for clarity, it's not that your SSD will fail after x cycles but that, along with reducing the OP over time (which reduces the maintaining of speeds as only part of the function of the OP is to provide replacement blocks - another reason for increasing the OP naturally), the warranties are fixed to that cycle count & it's stored in the SMART data on the SSD... [NB what happens when all of the 'real' OP has been used up in replacing blocks with failed cells in isn't abundantly clear... Well i know that one of the staff admins on the OCZ forum had modded a f/w for an indilinx drive that had all of the OP used in failed cells to increase the 'real' OP, which suggests that this may the actual limitation point (as the indilinxes could/can be OP'd by under partitioning then what would have been the point otherwise?)... - so it's possibly another good reason for choosing SSDs with natively increased 'real' OP - like the 510 or the SFs which both have 12.7% vs the (other) Marvell controller ones (ie the Crucial & Corsair P3) with 6.9% - ...but it's never been 100% clarified to my knowledge.]
