|
Ralphus posted:I re-installed Windows 7 today and just ran the Intel SSD toolbox. It's been updated to 3.1.2 (apparently as of November 27th) so I updated it and flashed the newest firmware to my 335 (240GB). It now shows 0% wear instead of the somewhat worrying ~5% wear it showed when I installed it 2 weeks ago! Huh, I'm curious about this now. Are the SMART details stored somewhere in firmware, rather than the OS partition?
|
#
?
Nov 29, 2012 08:26
|
|
|
|
| # ? Jun 8, 2024 00:45 |
|
|
Yeah, SMART details are generally stored in drive electronics. The Intel 330 and 335 are special cases re: drive wear indicators, as both have bugged ones. The 335 reports wear happening much faster than it actually is (fixed with the current firmware), and the 330 is completely bugged and the wear indicator never changes (to be fixed in December). Wouldn't make much sense to store electromechanical wear indicators in a place that could be damaged, overwritten or erased during the normal use of the drive.
|
|
#
?
Nov 29, 2012 08:45
|
|
|
Odette posted:Huh, I'm curious about this now. Are the SMART details stored somewhere in firmware, rather than the OS partition? Apparently the media wearout indicator data was incorrect. There's a link to an article that explains it a bit more here: http://forums.somethingawful.com/showthread.php?threadid=3454120&userid=0&perpage=40&pagenumber=99#post409669376 I wasn't horribly concerned about it after I read that, it just kinda threw me off when I installed the drive (before that article was published) and it showed that much wear on a brand new drive.
|
|
#
?
Nov 29, 2012 08:47
|
|
|
The OP says regarding Macs: "Samsung 830s work fine, but for the smoothest experience go Sandforce. Overall, the Samsung 830 drives work best in Macs, and they're what Apple ships in their latest systems." Isn't that a contradiction? I'm torn between a Samsung 830 and a Corsair Force 3 for my MacBook Pro. Looking at the 256GB / 240GB capacities. Edit: Duh, I get it, the update was the bit about Sandforce and the paragraph below was kept there as a reference point. Force 3 it is.
|
|
#
?
Nov 29, 2012 12:40
|
|
|
Caged posted:The OP says regarding Macs: A Samsung drive provided by Apple will have working TRIM automatically done by the OS, but not one you buy yourself. OS-based TRIM is only enabled on drives provided by Apple (they may have Apple's firmware on them, an Apple logo on their sticker, etc). If you buy a Samsung drive from some random retailer, TRIM will not be enabled by the OS, and the drive itself does not do its own built-in garbage collection. There is a way to hex-edit a system file or just use a 3rd-party program called "TRIM enabler" to modify the system file to force-enable the OS-based TRIM. That method/program work fine for most people (as far as I know). Basically, you hack your OS to enable TRIM support. On the other hand, SandForce drives are favored by a lot of "non Windows" users, as it doesn't matter if you're using a Mac, Windows XP, DOS, BeOS, PlayStation 3, or whatever, the drive itself has the built-in garbage collection. You just pop it in and go. OWC (not to be confused with OCZ) makes its own SandForce-based drives for Mac, and recommends going with SandForce drives with built-in TRIM instead of relying on OS-based TRIM. http://blog.macsales.com/11051-to-trim-or-not-to-trim-owc-has-the-answer (Of course, since they make drives with the built-in TRIM, they are going to tell people that they should only buy drives with built-in TRIM.) Some like the idea of the "over provisioning" of the SandForce drives: 256GB flash -> 240GB usable, with 16GB for RAISE error-recovery (auto-replaces bad cells if any are found during the life of the drive). I don't know how important this is; I didn't see much talk about it in the OP. The Samsung is regarded as really reliable, and its probably too early to know of issues of cell failure. Most SSD failures I've heard about are just total device failure, not cell failure. It's still a pretty new technology. Maybe in 5-10 years, when the "reliable" drives have lived their full life, individual cell death may be more noticeable. For our work systems (MacBook Pros), we picked up a few 480GB OWC drives (SandForce 2281). For desktops/other workstations at work, we picked up some of them nice Toggle NAND Mushkin drives (SandForce 2281). For my home computer, I picked up a nice Toggle NAND SanDisk drive (SandForce 2281).
|
|
#
?
Nov 29, 2012 13:12
|
|
|
Xenomorph posted:the drive itself does not do its own built-in garbage collection. Nope, Samsung 830 does have idle-time garbage collection
|
|
#
?
Nov 29, 2012 13:15
|
|
|
DeathChill posted:Weighing decisions here: get the Samsung 840 500 GB that's only $400 (CAD) or go with the more expensive 830 512 GB, which is $550. I feel like I want to live dangerously, but mostly cheaply. For what it's worth I found a 512GB Samsung 830 for $479 from this retailer. http://www.bhphotovideo.com/c/product/845683-REG/Samsung_MZ_7PC512B_WW_512GB_SSD_830_SERIES_INTERNAL.html Shipping is 8 business days but seems legit.
|
|
#
?
Nov 29, 2012 15:35
|
|
|
So about that Crucial M4 firmware warning: I made sure that my drive had 000F on it. Last night after an on and off power outage, the drive was not recognized in the BIOS. I unplugged the SATA and power connectors, reset the CMOS and reconnected the SSD and it still would not initiate. I tried the fix and got it up and running by letting it sit for about 20 minutes. While I was doing my research, I found this thread (last post): http://forums.anandtech.com/showthread.php?t=2275790 It may be a 000F bug as well. I have rolled back to 0309 and I will report if I have anymore issues. AsRock Z77 Pro3 (UEFI BIOS) M4 256GB I don't think other specs are relevant to this issue.
|
|
#
?
Nov 29, 2012 16:47
|
|
|
comper posted:I made sure that my drive had 000F on it. Last night after an on and off power outage, the drive was not recognized in the BIOS. I unplugged the SATA and power connectors, reset the CMOS and reconnected the SSD and it still would not initiate. I tried the fix and got it up and running by letting it sit for about 20 minutes.
|
|
#
?
Nov 29, 2012 16:53
|
|
|
Xenomorph posted:On the other hand, SandForce drives are favored by a lot of "non Windows" users, as it doesn't matter if you're using a Mac, Windows XP, DOS, BeOS, PlayStation 3, or whatever, the drive itself has the built-in garbage collection. You just pop it in and go. Out of curiosity - how does this work? I'd think that the drive would need to be able to read the filesystem to see what blocks are marked as safe to wipe. Adding that for common filesystems like NTFS, HFS+, FAT, and so forth probably wouldn't be too hard, but without either understanding the filesystem or the OS sending TRIM hints, I don't see how it would be possible to safely discard data.
|
|
#
?
Nov 29, 2012 16:58
|
|
|
My guess would be that that kind of hanging failure mode can be caused by multiple issues inside the drive so this MAY not mean the same problem extends to 000F, but that's just speculation.
|
|
#
?
Nov 29, 2012 17:00
|
|
|
Obsurveyor posted:What fix? Letting it since in the BIOS for 20 minutes? I've got a stock 000F M4 in my netbook and I'd like to know in case it does this. Also, did yours come with 000F or did you have to upgrade/downgrade it? My drive came with 000F on it. Yes, the fix was to just let the computer sit idle in the BIOS menu for about 20 minutes so it can initialize. It still didn't show up in my BIOS device manager, but after I rebooted, it finally got into Windows 8 again. I think this mainly has to do with UEFI BIOS compatibility, so you may not have to even worry about this issue with your netbook, but I could be wrong.
|
|
#
?
Nov 29, 2012 17:08
|
|
|
Just saw this posted on the crucial forumsquote:We are in the final test and validation phase for a new Crucial m4 SSD firmware (040H/04MH), expected in mid-December. Well here's hoping Crucial is only on the poo poo list for 2 more weeks.
|
|
#
?
Nov 29, 2012 17:31
|
|
|
Also, according to this post, some of the problems are caused by the drive mistaking a shutdown for a faulty power supply and turning off to try to protect itself. That's why leaving the system running for a while usually fixes the problem. I don't understand why they would even bother, since a faulty power supply would probably destroy the drive anyway.
|
|
#
?
Nov 29, 2012 18:08
|
|
|
Does anyone know how long the 128GB Mushkin Enhanced Atlas has been out of stock on Amazon? I have a ton of Amazon credit so want to get it there, but if it's one of these things where it's been out of stock for months that would be nice to know.
|
|
#
?
Nov 29, 2012 20:00
|
|
|
When I saw the M4 sticky I checked my firmware and saw it was 0309, then downloaded the 000F update from Crucial, but when I ran the update the program just said that no update was needed and quit. I re-read the sticky and saw that it said the 5000hr bug was in firmwares prior to 0309, so I figured I was okay. Assuming that Crucial doesn't put out a good firmware in a few weeks, should I be concerned at all about running 0309? Why won't their software update my drive?
|
|
#
?
Nov 29, 2012 20:24
|
|
|
Are you sure your drive is connected to an Intel SATA port set to AHCI mode? If so, please post a Crystal Disk Info screenshot.
|
|
#
?
Nov 29, 2012 20:33
|
|
|
0309 is one of the perfectly fine revisions, if you have it don't bother changing it.
|
|
#
?
Nov 29, 2012 21:21
|
|
|
Glen Goobersmooches posted:0309 is one of the perfectly fine revisions, if you have it don't bother changing it.
|
|
#
?
Nov 29, 2012 21:35
|
|
|
Alereon posted:Are you sure your drive is connected to an Intel SATA port set to AHCI mode? If so, please post a Crystal Disk Info screenshot. I don't even know what AHCI means so I probably left that to set to auto. I'll check it out later and post a shot.
|
|
#
?
Nov 29, 2012 21:39
|
|
|
Copper Vein posted:I think it's on a Marvel SATA port. I have a Asus Z68 board and it has two Intel 6gb ports and two Marvel 6gb ports. Does that matter? I have platter drives on the Intel ports. Never ever ever ever stick your SSD's on anything but Intel SATA ports.
|
|
#
?
Nov 29, 2012 21:47
|
|
|
Ok. Is the issue with them documented in the OP or summed up elsewhere on another forum? Does the issue extend to platter drives as well.
|
|
#
?
Nov 29, 2012 21:53
|
|
|
Papercut posted:Does anyone know how long the 128GB Mushkin Enhanced Atlas has been out of stock on Amazon? I have a ton of Amazon credit so want to get it there, but if it's one of these things where it's been out of stock for months that would be nice to know. I'm not sure about that particular drive, but I would avoid Mushkin. I just had one of their 128gb SSDs fail on me after about a week of use and when I did some research found that their failure rate is way above industry. It could just be the particular drive I had and not their whole lineup. Mine was the Chronos model. No idea how long they have been out of stock though.
|
|
#
?
Nov 29, 2012 21:57
|
|
|
Copper Vein posted:Ok. Is the issue with them documented in the OP or summed up elsewhere on another forum? Does the issue extend to platter drives as well.
|
|
#
?
Nov 29, 2012 22:25
|
|
|
Alereon posted:never use non-chipset-provided SATA controllers for your system drive (HDD or SSD), and it's strongly not recommended for storage drives. I hope you do get to update the OP because I'd really like to look into this further. Edit: Also, the term "controller" is being used a lot but the way I'm understanding it is that SATA controllers and SSD controllers are different things and that the Marvel SSD controller in my M4 may be allright but the Marvel SATA controller on my mobo is recommended against. Copper Vein fucked around with this message at 22:41 on Nov 29, 2012 |
|
#
?
Nov 29, 2012 22:38
|
|
|
sephiroth838 posted:I'm not sure about that particular drive, but I would avoid Mushkin. I just had one of their 128gb SSDs fail on me after about a week of use and when I did some research found that their failure rate is way above industry. It could just be the particular drive I had and not their whole lineup. Mine was the Chronos model.
|
|
#
?
Nov 29, 2012 22:41
|
|
|
What's the standard procedure for moving over the contents of an old small ssd install of windows to a new larger drive? I feel like I've asked this before in this thread, but I wasn't able to find it. Thanks again Alereon for the deal you posted over the weekend on the 240GB drive. edit: I see somebody in this thread used Macrium Reflect Free to do this. Barring other suggestions I'll probably go this route. MixMasterMalaria fucked around with this message at 23:12 on Nov 29, 2012 |
|
#
?
Nov 29, 2012 22:53
|
|
|
Copper Vein posted:Well that sucks. That would cut my available 6gb ports down to 2 when I've got 3 6gb drives installed now and a fourth that I just got cheap from Newegg last week. They're both controllers; the SSD controller is responsible for controlling/interfacing with the flash memory, and turning that into a SATA interface. A SATA controller sits on some bus and interfaces SATA devices with the host system, usually as a PCI device (to software anyway). And yeah, right now for most consumers, you only get 2 6Gb/s ports out of the chipset. You'll need a larger/beefier HBA like something based on the LSI SAS2008 to get more than those, and I have no idea about TRIM support or similar on those.
|
|
#
?
Nov 29, 2012 23:02
|
|
|
Is this article still accurate regrading how drives do garbage collection? It seems like a drive could simply read the NTFS free space bitmap and use that to do GC. ext2/3/4, UFS, HPFS, and HPFS+ work similarly, and FAT is simple enough to read and figure out which clusters are free. Obviously building in per-OS support isn't ideal but if you hit the major FSs that could be worth while. Does anyone have GC that works this way?
|
|
#
?
Nov 29, 2012 23:50
|
|
|
Ninja Rope posted:Is this article still accurate regrading how drives do garbage collection? It seems like a drive could simply read the NTFS free space bitmap and use that to do GC. ext2/3/4, UFS, HPFS, and HPFS+ work similarly, and FAT is simple enough to read and figure out which clusters are free. Obviously building in per-OS support isn't ideal but if you hit the major FSs that could be worth while. Does anyone have GC that works this way?
|
|
#
?
Nov 30, 2012 00:04
|
|
|
HalloKitty posted:Nope, Samsung 830 does have idle-time garbage collection AnandTech doesn't seem to like the "idle-time" part of that. http://www.anandtech.com/show/4863/the-samsung-ssd-830-review/6 quote:My biggest issue with Samsung SSDs in the past has been their extremely poor performance over time. Samsung doesn't do a lot of active garbage collection while writing in order to maintain ultra high write speeds, instead it prefers to clean up the drive during periods of little to no IO activity. Unfortunately this approach can result in pretty poor performance over time. quote:Performance drops significantly, down to as low as 50MB/s for the earliest LBAs. Given enough idle time the 830 should correct much of this and obviously TRIMing those LBAs will restore full performance (as you'll see below) but the point is that by delaying the bulk of garbage collection the Samsung SSD 830 is able to drop in performance by a degree that I'm not super comfortable with. This phenomenon isn't exclusive to Samsung, you'll remember that we've complained about it with Crucial drives as well. Other than SandForce and Intel most controller manufacturers tend to follow a similar clean up the mess later approach to firmware design. In my opinion I'd much rather see lower peak performance and get higher worst case scenario performance as it tends to impact the user experience less. The SandForce drives may not have the performance of the Samsung, but average performance keeps up nicely due to active garbage-collection. If you are 101% sure that you're going to be in an OS that does its own TRIM correctly, the Samsung is of course more than fine. That is easily most of the population. But what if you want a cheaper drive for your quirky OS of choice? Space Gopher posted:Out of curiosity - how does this work? I'd think that the drive would need to be able to read the filesystem to see what blocks are marked as safe to wipe. Adding that for common filesystems like NTFS, HFS+, FAT, and so forth probably wouldn't be too hard, but without either understanding the filesystem or the OS sending TRIM hints, I don't see how it would be possible to safely discard data. It's transparent to the OS. The drive monitors writes at a block level as part of the wear leveling. The SSD hardware doesn't know if an OS is updating your file or marking it as "deleted". It simply knows it just received a write request to an area already occupied. The wear-leveling won't write to the same area twice, so it just writes the change to another area. It keeps track of the areas have had the multiple write requests made, and once an entire "block" has had its write redirected somewhere else, the garbage collection is then performed on that area. By the time that block is used again, performance has already been restored. The SSD basically does a lot of logical remapping that the OS doesn't know about. Some graphics and info here: http://thessdreview.com/daily-news/latest-buzz/garbage-collection-and-trim-in-ssds-explained-an-ssd-primer/ Xenomorph fucked around with this message at 00:47 on Nov 30, 2012 |
|
#
?
Nov 30, 2012 00:45
|
|
|
So if I'm putting a Sandforce SSD into my MacBook Pro, is there any point in running the TRIM Enabler or is it safer to just not bother with something that could potentially make the system unstable / get disabled in an update and let the SSD handle things itself?
|
|
#
?
Nov 30, 2012 01:55
|
|
|
Caged posted:So if I'm putting a Sandforce SSD into my MacBook Pro, is there any point in running the TRIM Enabler or is it safer to just not bother with something that could potentially make the system unstable / get disabled in an update and let the SSD handle things itself? OWC says: don't enable TRIM, let the drive handle things itself. Someone from SandForce (the bottom link in my post) says: built-in garbage collection in addition to OS TRIM is best. Me: I have no idea. I *think* I have TRIM enabler set up on my Mac. When upgrading to 10.8 from 10.7, it was simply disabled. I just turned it back "on" (I believe it just does a simple byte swap on a system file, a 1 to a zero or something). It just needs to be re-enabled with each OS update. I haven't done any real tests to see how well it works.
|
|
#
?
Nov 30, 2012 02:10
|
|
|
Am I reading the OP correctly, that if I'm installing fresh Ubuntu 12.04 on an ext4 partition, I won't need to mess with any settings? I ordered that Sandisk256g drive that Alereon posted during the black Friday sales.
|
|
#
?
Nov 30, 2012 03:02
|
|
|
Alereon posted:I'll see if I can find some place to put details int he OP, but yes, you can never use non-chipset-provided SATA controllers for your system drive (HDD or SSD), and it's strongly not recommended for storage drives. The only reason they exist is to allow for eSATA without using some of the limited chipset SATA ports. As a follow up, I have an ASRock Z77 Extreme4 that I'm about to install. I'm planning on permanently running 2x HDDs for storage, and 1x SSD for system. In addition, I've got a 3rd HDD I'm going to connect temporarily for file transfer before I wipe both it and its backup drive and get rid of them. Am I correct in reading that this board has 6x SATA ports provided by the chipset (2x 6.0 Gbp/s, 4x 3.0Gbp/s) and 2 additional that are non-chipset provided, since they share a controller with the rear eSATA port? As a followup, I'm planning on putting my SSD on the 6.0 port and the 2 permanent HDDs on the 3.0, just to eliminate any I/O difference between the storage drives - is that the best option?
|
|
#
?
Nov 30, 2012 03:12
|
|
|
Shooting Blanks posted:Am I correct in reading that this board has 6x SATA ports provided by the chipset (2x 6.0 Gbp/s, 4x 3.0Gbp/s) and 2 additional that are non-chipset provided, since they share a controller with the rear eSATA port? quote:As a followup, I'm planning on putting my SSD on the 6.0 port and the 2 permanent HDDs on the 3.0, just to eliminate any I/O difference between the storage drives - is that the best option?
|
|
#
?
Nov 30, 2012 03:23
|
|
|
MixMasterMalaria posted:What's the standard procedure for moving over the contents of an old small ssd install of windows to a new larger drive? I feel like I've asked this before in this thread, but I wasn't able to find it. Thanks again Alereon for the deal you posted over the weekend on the 240GB drive. For those wondering, it worked like a charm and was very painless.
|
|
#
?
Nov 30, 2012 03:26
|
|
|
I'm trying to minimize costs and I'm looking for a secondary drive (to put games on) that has at least 128 gigs. I think I found a Transcend drive for 100 and a Samsung for 100 USD. Transcend SSD Samsung SSD. These look relatively inexpensive with the current deals, probably around 73 cents per gig. Is there a massively better deal out there or a more reliable model?
|
|
#
?
Nov 30, 2012 03:51
|
|
|
Both of those models are on the "do not buy" list because of questionable reliability (error-correction disabled, unproven memory and firmware, respectively). Check out this ADATA S510 120GB for $89.99-$10 MIR=$79.99 at Newegg. It would seem to be roughly identical to the Transcend but with error-correction enabled (hence the 120GB instead of 128GB capacity).
|
|
#
?
Nov 30, 2012 03:58
|
|
|
|
| # ? Jun 8, 2024 00:45 |
|
|
Alereon posted:Both of those models are on the "do not buy" list because of questionable reliability (error-correction disabled, unproven memory and firmware, respectively). Check out this ADATA S510 120GB for $89.99-$10 MIR=$79.99 at Newegg. It would seem to be roughly identical to the Transcend but with error-correction enabled (hence the 120GB instead of 128GB capacity). That looks great.  Steam finally allows installation to secondary drives, so this is a good way to make some more room for 5 or 6 laaaaaarge games + scads of small ones.
|
|
#
?
Nov 30, 2012 04:00
|
|