Simen1

Member
  • Content count

    33
  • Joined

  • Last visited

Community Reputation

0 Neutral

About Simen1

  • Rank
    Member

Profile Information

  • Location
    Trondheim, Norway
  1. Simen1

    Samsung F series?!

    The news first started about 3. of June. Your post came 3 weeks later, the 25. of June. Now its been another 3 weeks and i have still not seen any reviews on the net. I guess real availability in more then a few exotic webshops are at least a month away. Hopefully tomorrow, but realisticly probably not before late august or september. Looking at the popularity on this topic on the net and comparing that with the fact that production normally ramps from low volumes to high volumes in a few months then it might be sold out fast and stay sold out for weeks before new supplies arrive. Samsung may well up the price if it the disc is extremly popular.
  2. Simen1

    Girl, 11, will be mother

    Its a tragic outcome for both parts. Both the girl, the boy and their familys. But i really hope the boy isnt charged for rape without reason. That would make it even more tragic. I hope and belive the police know how to see the difference between a rape and a voluntary stupid act by both of them. Stupid like in not using a condom. This stupidity may well come from not being mentaly mature enough to understand the risks and consequences.
  3. Simen1

    Samsung F series?!

    The samsung F1 will probably run in circles around the 7K1000 if the platter size is 334GB. 7K1000 only has 200 GB platters. Taking into account that Samsungs transferrate numbers are exaggerated with about 35-55% (like their 400 and 500 GB models) then the Samsung F1 still should have about 40% faster transfer rate then Hitachi 7K1000. I'll support the idea That will make up an unique marketing item - you never know who will start to fan it! Anyway, it's also technically a big waste to make 500GB and 750GB drives using 333GB platters... At the same time 333GB fits for 320GB pretty ok, although why not to offer +13GB bonus capacity also there. Just marketing it still as 320GB drive but with some bright extra sticker "13GB Bonus Capacity!" added on the drive label. Or just let it a hidden surprise for users (just like Hitachi did earlier - its 160GB drives were in fact 164GB). A 500 GB model with 2 platters and 3 heads would fit nice. But i also hope for a 666 GB (or 650 GB) model and two low profile models at 320 and 160GB (one platter and 1 og 2 heads). Maby also a 800 GB (3 platter 5 heads) model. Cutting down the capacity a bit is probably just done with making a slightly shorter stroke, meaning they disable the lowest performing part of the platter.
  4. To the question in the topic: Yes, i think mechanical storage (harddiscs) have a future. Flash does have a nice track of groth in space per $, better then harddrives have. Stipulating into infinity flash will pass harddrives in GB/$ sometime in the future, but things will of course not have the sema growthrate into infinity. Sometime both technologies will stop. Just like the clockspeed on CPU's that everybody belived to grow into infinity at the same growthrate as always, just a few years ago. So, how far will flash and harddrives keep the same growthrate as today? I dont know. But my guess is maby 5-15 years. That also means that when 0,85" and 1" harddrives meet hard competition by flash today, altso 1,8" drives will meet that competition in a few years. If 2,5" drives will meet the competition or not before growth slows down and maby stops, i dont know, but i'm convinced 3,5" drives will remain untoched. Another trend is that OS and programs dont increase in size as fast as other storage needs. That may split the market in a flashmarket for OS and programs, and a harddrivemarket for general storage. Write and read-speeds on flash is actually a far smaller problem then many thinks. Its pretty easy to RAID its way out of htat problem. The Raid-controller even dont have to be on the motherboard. It can easily be integrated into a part of a SSD-drive, and thus offer very fast transfer rates compared to harddrives. (Not only the more expensive OneNAND, but altso simple and cheap NAND)
  5. Simen1

    Flash vs Hard drive performance

    - Flash has an access time that is about 100 times as fast as harddrives. - Flash has a limited read/write bandwith at about 15-40 MB/s, but this can be solved by putting lots of (smaller) flashdrives in Raid0. So, 4 1GB-cards should not only outclass any harddrive on access times but altso offer some better bandwith. Note, flash keeps the same bandwith trough the whole media, but harddrives fall from a peek at the start normaly down to the half at the end of the media. Reliability may be a concern but if you keep an image of the OS-partition at a normal disk you should be pretty safe. I have heard of people running Windows XP from flash and most of them got problems within a year of usage. But this is about 2 years old information and i think the reliability have improved quite a bit since then.
  6. Simen1

    Magnets

    The magnetic field around theese magnets are strong compared to everyday-experience, but its actually very small compared to the extremly strong and consentrated macnetic field at the very "focal point" at the tip of the writing head. Magnetic field is reduced by distance^2 so a couple of mm away from the writehead its almost nothing left of it. The writehead is under 0,001 mm away from the platter. Since the magnetic field is extremly consentrated is will only affect one bit at a time, even though the neigbouring bit is just 0,0005mm away from it. This page is in norwegian, but it contains some nice pictures from when i disassembled my harddrive half a year ago: http://www.diskusjon.no/index.php?showtopic=527026
  7. Is the 160GB/platter-number a very hard fact-number? Did they stretch the number up to 167GB/platter to make a 3-platter, 500GB-model or is the 500GB-model 4 platters?
  8. Simen1

    Seagate announces 750GB Barracuda 7200.10

    Bingo13: I cant access your pictures. Have anyone seen any serious benchmarking on this yet?
  9. Simen1

    Any news on 600GB harddiscs?

    First: Thanks for the look into technical limitations. I just read the other day that demand for harddisc drives is increasing at the moment: Hard-Disk Shipments Break 100 Million The article points out that demand for drives to laptops, mp3-players, and digital video recorders are driving much of the sale. They altso point out that average price per drive is on its way up. Suggesting that more people buy high-end drives and less people buy low-end drives then before.
  10. Simen1

    Any news on 600GB harddiscs?

    SSD have a very long way yo go in terms of cost per GB. Currently a 250GB drive costs about 100$ and for the same price you can buy about 3GiB of flash memory. Thats a difference of 83 to catch up with. Even with continous exponential growth of doubling the capacity every 18 months, assuming hard disc drives cost the same all the time it will take 10 years to cover the gap. Realisticaly drive capacity will increase so it might take 15 years. Another factor that silicon chips is expected to reach technical limitations too in about 5-10 years from now. Currently it looks that 45nm is the smallest node that is practically and economically obtainable. This might be extended to 32 or 22nm, but any further seems extremly unrealisticaly. (Not by me, but by the industry it self). In other words it looks like SSD might never catch up with harddrives in costs per GB.
  11. Simen1

    Any news on 600GB harddiscs?

    There is no logical reason to why sizes like that should be followed. After what i have seen, sizes are always keept in rond nice numbers. Platter sizes are designed to match round numbers. i.e. 125GB platters that is used in 250GB drives doesnt get placed in 3 platter designs to make 375GB drives. 300GB is a more used size then 320GB. No discs of 480GB exists yet. Because of the technical limitations mentioned and marketing round numbers i think disks can increase in this order: 500 - 600 - 700 - 800 - 900 - 1000, or maby 500 - 600 - 800 - 1000.
  12. Simen1

    Any news on 600GB harddiscs?

    Harddisks have since mid-80s increased with about 50% per year. A pace that has lastet about 20 years in the industry. In my opinion a pace that is proved long term sustainable and economicaly. I understand an increase of a wopping 120% per year is too much to keep up with in the long term, but a dissapointing 0% since last february is a still standing. I dont expect en increase of 120% annually, but it would be nice if there was an increase at all. Even if 50% is much, i would be glad for a minor 30% increase. 30% is better then nothing. I regularly see people buying two og four 250GB harddiscs in stead of one or two 500GB. Isn't it much cheaper for the producers to just add a couple of platters to a 250GB-disk then to make a complete new drive?
  13. Is there any news on 600GB discs? It's over a year ago scince Hitachi broke the 500GB capacity limit for single harddiscs. Is 600GB discs just around the corner or do we have to wait another 6-12 months for it? When will 2-platter 300GB discs become the best value option? (Most GB for the buck) Why has disc capasity hardly been increasing the last years? About 5-10 years ago the size increased about 50% every year.
  14. Certainly. But it doent mean the drive probably would last longer. Contradictive? No, not if you look at the way MTBF realy are calculated. MTBF is NOT an estimation of how long a drive wil live. It's a estimate of the risk of failure within a certain period of time. Lets say 24 hours use since i dont know WDs practice on this. If MTBF is estimated to be 1,2 million hours, you could expect one in 1,2 million disks to die each hour the next 24 hours. In other words 24 out of 1,2 million drives wil probably die within 24 hours. The risk of drives dieing is at the first moments of "burn in" is very high. The disks with serious errors wil likely die within minutes. The risk drops as time goes. 8 hours of "burn in" is obvously enough for MTBF = 0,6 million hours. Further 16 hours of "burn in" makes a few more disks fail so the momentarly MTBF can be 1,2 million hours. Since many disks fail very early and the risk drops, the MTBF altså drops with more "burn in". Note: MTBF is a expectation of failure within a certain period of time, and not a estimation of how long a disk will last. After long usage the disk probably will get problems not accounted for in a 24 hours "burn in test". Like wear out of bearings, evaporation of oil in FDB, electromigration in the curciuits and so on. MTBF gives in my opinion a false expectation of disk lifetime. So dont expext a drive with MTBF = 1,2 million hours to last 1,2 million hours. (= 137 years). And dont even expect a 137 disk Raid0-array to last about one year. It's far to optimistic.
  15. Simen1

    3 Gb/sec or 300 MB/sec?

    I prefer 300 MB/s (of course meaning 300 000 000 Bytes per second) not "sec" since it's not SI-standard. I prefer this becaus its the most comparable number and Byte is more used then bit. But in the long run, i see the meaning of "byte" is vanishing. I dont want another counting method like "dosen". SI-standard is always using 10, 10^2, 10^3 and so on. Suddenly using an "8" would undermine the very nice SI-system. In the long run i wold prefer that the "byte" is degraded and "bit" becomes the dominating way of writing numbers.