yeti

Member
  • Content Count

    32
  • Joined

  • Last visited

Posts posted by yeti


  1. Yes it's possible, you know why ?

    214167[/snapback]

    We probably see here the same thing that hit HGST earlier: a larger number of platters make the assembly wobble more, hence making it harder to keep the heads on track with the high track density...

    so the first drives coming out with the high track density tend to have fewer platters.

    Yéti


  2. About the 8 vs 16 MB cache:

    I had the occasion to test the caching capabilities of the Toshiba drives with 16MB cache. They acted very much as two independent caches, one for reading, one for writing, and not even that great at it. Mind you this was on a previous model.

    I would thus ignore the 16MB "advantage" as it did not show, and go for the proven HGST.

    Yéti


  3. Anyway, I think, I am not sure though, that the WD drive is made up of 80GB-platters! So the 320GB model should have 4x80GB-platters.

    WD does not have a four-platter platform. The 320 GB drive they recently announced is based on their existing three-platter platform, so it's ~107 GB/platter. (I think Eugene even mentioned this on the main SR page in the news section.) If they *had* a four-platter platform, I would think they'd try to get a 400 GB product out.

    195207[/snapback]

    I would think that this is in deed a four platters platform as the maximum sequential transfer rate announced is identical to that of the WD250.

    If it had a three platter for this capacity it would have had to maintain the bit rate, which means have taken all the capacity increase by squeezing tracks closer to each other, which in turn makes little sense -but Seagate did similar things in earlier times...

    So, my bet is that this is a four platter for the time being until they can increase their areal density (and sequential transfer rate), and if WD is true to form the three platter version will come unannouced.

    Yéti


  4. arga,

    The easiest way to find what kind of drive you have would be to test the maximum sequential transfer rate: a 7200.7 (100GB/p) gets you almost 65MB/s where the 7200.8 (133 GB/p) should get you 68MB/s or more. Average transfer rates or burst values would not tell you much.

    If you only get 65 MB/s, you probably have hit a three platter 100GB/p drive which is the equivalent of the latest Maxtor Maxline whatever, albeit with a smaller cache.

    Keep us in the know!

    This looks great, Seagate is moving ahead in the Gigabyte/platter race.

    Yéti


  5. This appeared AFAIK yesterday, July 1st, with an interesting twist: The gaming score is extraordinarily high, entirely due to an outlandish success in "Black and White 1.1".

    Could SR's decision to limit the testbench memory to 128 MB warp these results?

    In my opinion, yes, I cannot trust the general applicability of this gaming benchmark, if an otherwise decent drive suddenly shows up 10 times faster on it.

    Yéti


  6. Both USB2 high speed and e-sata should do well time wise, even if e-sata is much faster than USB.

    Yes, keep to cool drives, you don't need fast expensive ones.

    And think of making sure you have high quality connectors. Many do not like to be plugged/unplugged too many times, and will give up on you long before you are ready to change the drive.

    Yéti


  7. We've had the opportunity to test both drives at work.

    The 8MB cache Momentus (it comes in 2MB and 8MB) is faster in WinBench Biz, by over 10% and HE by 2% (not a significant difference here) (oops, sorry Eugene, WinBench is not to be mentioned here) than the HGST one, but 6% slower in STR.

    Random reads are faster with Momentus, while random writes are faster with HGST. This is probably due to the much larger "average seek time" of the HGST drive.

    Yéti


  8. Cache size:

    ATA manufacturers think mostly about costs. Memory costs continue to go down fast. The cache size should continue to increase, from the current "standard" of 8MB, to 16 or more probably 32 MB within at most a couple years.

    Areal Density / Capacity per platter:

    I've read somewhere that there were 3 manufacturers planning on introducing a 100 GB/p 3.5" ATA drive. This should make Seagate (who said it would), Maxtor (who rumored it would) and... WD?

    The next level is expected at 120 GB/p, and should hit (my bet only) late '04.

    Spin speeds:

    This is costly but gets performance. I'd see it coming for desktops only if performance is really required (and with cache sizes increasing, the pressure may be off). Raptor' claims to success notwithstanding, I'd guess all other manufacturers are checking the sales numbers closely before introducing their own competitors to Raptor. For servers (SCSI et al) the story is different. 20K should eventually (say within a couple years) be in the offing.

    Yéti


  9. WD has not been very forthcoming in giving this type of info...

    but if you have the drive, you usually can tell which platter capacity you have by checking the sequential transfer rate (almost any benchmark will do).

    A WD 40 GB/platter should be close to 49 MB/s, while a WD 20 GB/platter will only get you about 34 MB/s. As a comparison, the 80 GB/platter (modern) drive gets you over 55 MB/s

    If you can't test the drive, you have to ask a WD insider

    Yéti


  10. The biggest ones I've actually seen in stock and available are 250 GB, from both WD and Maxtor. Maxtor lists a 300 GB drive on their site, but I haven't seen it in stock anywhere.

    A Google search shows some available at least at DABS in the UK... but you're right not much in the US.

    As the thread hints there must be yield issues on those fat heads!

    Yéti


  11. LOST6200,

    SR has had the scores only of the WD250 available since the middle of last week on their database (very similar to those of the 200)... even if there is no full review, the hard numbers are there, despite your worries, Eugene does a decent job!

    Yéti


  12. Does anyone know about the expected shipping date of the IBM/Hitachi/HGST 2.5" Travelstar 80GN?

    80GN Product page

    This drive has been announced in autumn 2002, if I remember correctly, with volume shipments coming in Jan 2003.

    I need to upgrade in the next month or so, and the 80GN looks very attractive, but so far it's only a paper product.

    I thought I remembered shipments anounced for March '03 (probably by Hitachi last month), which would be consistent with IBM's position in this market.

    Yéti


  13. For some reason that I do not fathom, only 128 GB of a WD1800BB is accessible when the drive is connected via a Firewire bridge (ADS w/Oxford 911).  The odd part is that 160GB DM D540X and 250GB DM 16 drives function at the full capacity on the same controller.  There is nothing wrong with the individual WD1800BB; two exhibit the same behavior.  The drives function normally on an ATA-133 controller.  The OS is Win2k if that makes any difference.

    If I remember correctly, the earlier versions of the 911 firmware were limited to 28 bit LBA addressing, hence the 128 GB limitation. You should probably check with ADS and make sure they give you the latest firmware, which is updated to handle 48 bit LBA addresses (a few TB at least)

    Yéti


  14. I think that Eugene's suggestion for the WD120 / 200 issue (see that thread) applies well here, and that Maxtor has different transfer rates for the different capacities per platter (which does not seem to be the case for WD between 40 GB/p and 60 GB/p).

    If I'm not confused it should be:

    - a 60 GB/p if the max STR is about 50 MB/s,

    - a 80 GB/p if the max STR is 57 MB/s or just above.

    Yéti