Computurd

Member
  • Content Count

    80
  • Joined

  • Last visited

Everything posted by Computurd

  1. Good analysis, IBM seems to be limping off the field at this point.
  2. I agree, they definitely need to get more involved. There are so many benefits of SSD that to stay with HDD is just a waste, especially from a power perspective.
  3. Its important to remember that SSDs only account for 3% of the storage mix in the datacenter, possibly 4% if you are being generous. There are periods of time now where the fabs are placing orders on hold, simply because they cannot physically make more flash. We are currently having periodic NAND shortages every few months just trying to populate this tiny sliver of the datacenter capacity, with more severe flash shortages in the future being predicted by nearly all analysts. There is no way we can even scratch the capacity required to replace HDD with flash.
  4. I agree absolutely that those without Fabs are going the way of the dodo. They will have to differentiate on software offerings, but even that will wane as the companies with valuable software tech are vacuumed up by the fabs. HDDs imo are kind of like tape, they will just progress down the storage stack into lower and slower tiers. No doubt the performance tier is going to flash almost immediately, but I'm not sure about the capacity tier, at least for a longer period of time. Flash Forward (SanDisk/Tosh) have 2 fabs, Samsung has 3 with one being built, Micron has 3 and Hynix has 2. That is a long way from being able to support the EB capacity required. I personally think we will be on to ReRAM or some other memory tech before flash can supplant HDD.
  5. The issue is that in order to supply the amount of capacity to unseat hard drives it would require 75 more fabs. These weigh in at 10 billion apiece. There is simply no where near the capacity production to unseat HDDs, now or anytime within the forseeable future. Just one fab is a significant investment, and we do not see those popping up, except when necessary to replace older process fabs. They aren't actually increasing the number of fabs pumping out NAND, they are just replacing the old ones.
  6. The best advice I have ever received is to clone all of the drives in the RAID array before attempting a rebuild. This way, if your rebuild fails you can try again. Sorry such general advice...but i have no specific advice for that controller. Sounds like a freak occurrence (the worst kind!).
  7. Thanks for the detailed response. The level of granularity that you provide is excellent, and the idea of operating outside of file systems and their caching is a good way to eliminate possible performance variances. I enjoy your articles and the preconditioning graphs illustrate very well that you do a great job of getting the drives into steady state
  8. By utilizing the full LBA space, are you indicating that there is a file (or just data of some sort) written to the full LBA range? Or is the test just ran against the full LBA range while it is empty? Basically the gist of the question is whether there is a 100% fill level of the device (with data of some sort) when tested for the evaluation?
  9. Do you test with 100% fill on the SSD?
  10. Most SSDs for consumer use will retain data for at least five years, even though the spec for MLC requires 1 year of data retention without power. For SLC the spec requires 10 years. For enterprise the retention period has been shortened to 3 months.
  11. Computurd

    LSI-9211 and 9260-8i

    just wondering if anyone here has either the 9211 or 9260-8i, just wanted to compare notes, etc.
  12. Yups has worked for many others. funny thing is that was what an intel feller advised on their forums. however, that has been many moons ago and i do not have the link. however, in storage 'circles' that is pretty widely used.
  13. little known fact: you can run winsat disk to force the OS to recognize an ssd. you had to do this with older ssd's that werent automatically detected by winders. elevate a cmd prompt to admin status type in: winsat disk this is the same utility used for wei scores, but more indepth. it is what it runs on installation. if it detects ssd, it will make the change
  14. unfortunately if it is like the Intel cards that are also LSI rebrands, it might be slow with firmware updates, and might not hit the performance of the LSI branded controllers. the intels have always had this problem, even though they are the exact same thing. that is why many users would crossflash the firmware, before they stopped allowing it.
  15. you cloned the drive is why. you need to do a new install. Win7 is going to optimize your system for SSD if it is detected during setup. if not, no dice. also, make sure you are running achi. ssd arent that great on ide. you can try to force win7 to kick over into 'ssd mode' by running winsat elevate cmd prompt type in: winsat disk this tests your disk using the same parameters as WEI, and is how the os identifies the SSD in the first place. one more thing of note. most laptops still are using sata 1. there is a place in there to unlock it to sata 2 in some laptops. in others you are bonered. there are even manufacturers who arent allowing sata 2 unless you buy the laptop with an SSD already installed! it was initially for the purpose of power saving, to not use sata 2 btw. you can google and find people complaining about sata1 on lappies all over. there are even some 'hacked' bios out there allowing sata 2 for certain lappies.
  16. Computurd

    Do SSDs help game consoles?

    yeah i dont think it would be worth it, for a system with only one use such as that the price v performance ratio is horrible. consoles suck anyway
  17. that PERC will nbot even begin to handle the power of those SSD. it will hold them back more than anything. the fastest raid card available for your needs is a 9260-8i with Fastpath key. 160K IOPS @ 4k random. it is UBER. get a good SSD with good GC, and say screw the trim. you will be fine without it. my 9260 with 8 generation 1 crappy 30 gb vertex beat the pants off the Fusion I/O, not only in max read speed (i hit 1.875 gb/s, top is 2.8 read) but in IOPS and real world testing as well. imagine if you get 8 C300 on there. here are results of this controller with several different arrays: intels, c300's, vertexes, and acards. this is by far the fastest solution on the planet right now FASTPATH check out post 83 for nice summary
  18. well, long story very condensed here.... i purchased a ST1500W PSU, and on first boot it fried 3 of my SSD, all 3 were on the same power line. they are vertex gen1 30 gb drives. One of the SSD actually began smoking very badly, and was apparently the cause of the issue. I contacted silverstone and they did some testing of the unit, as a matter of fact VERY detailed testing. i will post their results, they tested teh PSU, and the wire the devices were on, even taking photographs of the entire testing process. so, since the drives are not going to be RMA'd i did some testing myself..... Upon power on to the SSD from another PSU, same result. the SSD starts smoking badly. so i dismantled the drives. the two that were not smoking show some very slight scorching on a particular part, but nothing near what the one that fired up shows. funny thing is that the damage is not near the power socket of the drive. it is near the SATA port and the side of the pcb.
  19. Computurd

    Anatomy of a OCZ smoked SSD drive

    yeah they agreed to look at them, so i am happy, at least they are gonna check them out. RMA is denied on grounds that i had opened the devices.
  20. pestilence! i vote for pestilence!
  21. the only thing that concerns me with this approach is that intelligent caching is going to use algorithms. and over time, usage patterns vary, esp with certain apps like games/etc. so if you are using different games, and different things, the algorithm is naturally going to adapt, thus caching different elements at different times, ad they fluctuate in usage. This means alot of movement of files, as they change in priority etc, they will be moved back and forth from HDD to SSD and vice versa. and that means erasing, lots of erasing! SSD don't necessarily like tons of random writes. they kill them, very quickly and most of your 'hot' files will be of the small random variety, esp with OS usage
  22. Computurd

    Anatomy of a OCZ smoked SSD drive

    At least SilverStone exercised some due diligence in their testing, trying to see if it was their problem, they even tested the same sata line it happened on. Silverstone is saying that it is an overvoltage caused by a bad drive on the line.
  23. Computurd

    Anatomy of a OCZ smoked SSD drive

    one of the other drives that was ruined, but didnt smoke, here is a pic for comparisons sake
  24. i have had the 1680 and the 1880. hands down 1880 is faster. next gen roc, plus x8 pci-e 2.0 = WIN