• Content Count

  • Joined

  • Last visited

Everything posted by rugger

  1. It is fine. The media wearout indicator still shows a lot of life left. When that reaches 0, and then maybe sometime a long time afterwards .... the ssd may finally be worn out. The old 50nm NAND can take a hell of a beating.
  2. Are your partitions aligned on the SSDs? Have you formatted your partitions with the correct raid parameters (mkfs,ext3 -stripe_size or whatever it is)? I can't tell that from the partition table information you have printed.
  3. Your performance is quite normal in my experience with the 480gig Sandforce drives. The fastest sandforce drives are definitely the 240gig drives. Movies are already highly compressed, and will always show the worst case write performance for the sandforce drives. Your programs and settings will usually be in a more compressible format, allowing for faster performance. Not as fast as the ATTO results, but not as slow as copying movie files. The compression that sandforce drives use is typically better for enterprise workloads like databases though, where the compressible nature of many of the writes greatly increases performance and lifetime of the drives. To answer your other thread, over-provisioning won't generally help you. Just keep the drive less then 80% full and you won't see the any performance problems.
  4. rugger

    Guts of a 16GB Transcend SLC SSD

    Yes it does have overprovisioning Based on LBA count of 31,277,056 sectors the drive has 16,013,852,672 bytes of user accessable storage. 16GB of NAND is 17,179,869,184 bytes. That means that the drive has 1,166,016,512 bytes of overprovisoning, (or about 6.7%) This is plenty for a small SLC based SSD. However, the JMF602 controller is quite awful, I wouldn't have bothered buying it.
  5. rugger

    ARE SSDs

    First of all, in the field, there have been plenty of firmware problems with the 320, so it isn't really the shining beacon of reliability that people expected. In any case, you are much better off having a good backup then you are trying to choose a perfectly reliable SSD. Because none of them are perfectly reliable, and it only takes a bit of bad luck and you lose your data. Hard drives are also vulnerable.
  6. rugger

    Vertex 4 on PCI-E slow

    First of all, you are limited to PCI-e version 1 on that motherboard for the 1x slots, so your Silverstone card is currently Limited to 160megabytes per second. Even on a PCI-e version 2 motherboard, you would still be limited to just a bit over 300meg per second. Thirdly, the Marvell chipsets, which the Silverstone card is based on, are pretty slow compared to the built in Intel SATA. So basicly, what I am saying, is that the Sivlerstone card is useless for what you want, and just to plug your SSD into the motherboard. When you get some money, upgrade to a new SATA 6 motherboard. Also make sure you use AHCI mode.
  7. rugger

    How failure prone are SSDs?

    While your argument has some merit, you really do overexaggerate .... If SSDs were as bad as you are suggesting, then I certainly would have seen something by now. Yes, I've seen a few firmware stuffups, and sandforce should be beaten with a metal poll for the firmware they release, but hard drives have the same sorts of stupid firmware issues too. Certainly some SSDs are worse then others in this regard. Just like hard drives, there are some SSDs that you need to stay the hell away from (eg OCZ Petrol models) It isn't because SSDs are unreliable ... it is because all data storage is inherently unreliable. Anyone who has a clue about PCs and data storage should already know this. Doesn't matter if you use hard drives, SSDs, optical disks or even data chisseled into stone ... you need a backup or one day you will lose data. I, so far, have seen pretty damn good data reliability from SSDs, but any device can up and insta-die on you. Funny enough, wheels still fall off cars .... but we accept it anyway as a price of having affordable transport. Of course, SSDs cannot be nearly as inexpensive as they are now if you want perfection out of the box. I'm prepared to live with the status quo, of letting early adaptors find the problems and get them fixed before deploying an SSD, and get better prices doing so. I'll even be an early adaptor occasionally!
  8. rugger

    Is it time to replace my SSD?

    You probably just need to secure erase the drive and reinstall from scratch. Note that secure erase is different from a normal format! If that doesn't correct it, then maybe a new drive is a consideration.
  9. At this point, I think you will be out of luck. Whatever deleted your files would have triggered TRIM to recover the deleted space. The sectors would have been garbage collected and no accessable trace left. I don't think it is a hardware problem, because your windows directories would have been affected too, causing the computer to show many errors, or even refuse to boot.
  10. This really sounds like one of the many viruses that hides files/programs from the user .... there is a utility called unhide.exe that can reverse this.
  11. Face it, until people put their money where their mouth is, and pay for all this extra testing and validation that they claim they want, it is all a bunch of hot air and what people really want are fast and cheap SSDs. Edit: you can always buy from intel, which spent ages validating their sandforce drives (but still didn't pick up the malfunctioning AES engine)
  12. The drive would be completely unreadable and unaccessable, as the controller chip includes the SATA interface and NAND interface hardware and software. You may even let out some magic smoke out of other parts, but I think that is not likely to happen.
  13. This TRIM issue seems to be a mountain out of a molehill issue. I've punched almost 700TB of data through a sandisk extreme 120gig drive (one of the drives supposedly affected by the TRIM bug) without having the drive die or preformance degrade. Possibly my drive will benchmark faster when updated to 5.0.3 (if it doesn't esucumb to NAND wear), but the current firmware is not unusably slow, and write endurance doesn't seem badly affected.
  14. Just be aware that the retention gets worse as a drive ages. As each flash cell wears from erase and program cycles, the floating gate loses it's ability to retain a charge over a long period of time. So when you have a fresh out the box SSD, retention will be very good, possibly even 10 years or more. However, apply 3000 or 5000 write cycles to the NAND and retention will drop quite a lot ... possibly as low as a year (but probably a fair bit longer still) This is why enterprise specifications only call for 3 months retention at end of life. This means you can use the flash cells for longer before they go out of spec, allowing for higher write applications.
  15. Page 26: For client drives, minimum 1 year for powered off data retention.
  16. Actually, retention on consumer drives should be at least a year .... at least while the drive hasn't used all it's erase cycles. A fresh out of the box drive may however be good for 5 years or more retention. For SSDs which are powered on regularly, the controller should be scrubbing the NAND, finding weak cells, and rewriting them elsewhere, so retention should be nearly infinite.
  17. Strange, under linux, I can rebuild a 4 drive software RAID 5 array of cheap WD 2TB drives in well less then a day, while operating under normal workload conditions. (I accidently caused my array to go degraded a few times due to dodgy SAS controller firmware and my Intel Motherboard weirdness!)
  18. If these drives are going to be as expensive as the RE2 RAID drives, then count me completely uninterested!
  19. rugger

    Pata SSD options?

    I think you are putting too much effort into a laptop that would now be very very old. Old laptops are flaky .... SSDs in laptops can be flaky .... IDE to SATA converters can be flaky .... A SSD in an old laptop using an IDE to SATA converter is just asking for so much trouble Find an old 7200rpm PATA spinner to put in the laptop, or buy a new cheap laptop with SATA.
  20. rugger

    Intel 510 250 over modern drives?

    Heh, I got a brand new intel 510 250gig for $300 from a local computer store clearing stock. At that price, it isn't a bad deal. I could probably get an M4 slightly cheaper then this, but waiting ages for international delivery is hard .... vs instant gratification.
  21. rugger

    Intel SSD 330 Review Discussion

    The Corsair force gt 120gb would be a great SSD to compare the intel 330 120gb to. They both have sync NAND, so we can easily see the difference between intel 330 and "normal" sandforce with sync NAND.
  22. rugger

    Intel SSD 330 Review Discussion

    Intel would have to use the SF2282 rather then the SF2281 to use 16GB dies. Probably just 2x8 GB dies per package.
  23. rugger

    Intel SSD 330 Review Discussion

    Can't you compare the 120gb Intel 330 against a 120gb Intel 520 or some other 120gb SSD with synchronous NAND? That way we can see how it performs in comparison to other 120gb sandforce based drives. Of course a 240gb drive will beat the tar out of a 120gb. It has more NAND banks per channel and can interleave more write/erase commands to achieve greater performance, so it isn't really a fair test, and all we can say is that a 240gb drive beats a 120gb drive.
  24. rugger

    SSD Wear Leveling question...

    No, SSDs will rewrite the data. As long as you are not restoring every day, wear would not be a significant problem though. Even if you do restore a full drive every day, you will still get at least a decade from a modern SSD before the NAND is worn.
  25. The SMART in the drive can show this information for many drives. Sandforce based drives expose this information in all their drives (TiB read/written) Intel 520/320 drives have a special SMART counter you can use to measure wear and writes. Crucial M4 has a erase cycle counter, which is crude, but helps you track SSD life. You can view smart information with CrystalDiskInfo