Jump to content


Photo

Fujitsu MAU3147


  • You cannot start a new topic
  • Please log in to reply
43 replies to this topic

#1 Eugene

Eugene

    Administrator

  • Member
  • 1,057 posts

Posted 21 November 2004 - 03:15 PM

The Fujitsu MAS3735 has held the distinction as the "world's fastest drive" for nearly two years. As a result, expectations run high for the MAS's successor, the 147-gigabyte Fujitsu MAU3147. Join StorageReview as we examine how the MAU rates when compared to drives such as its predecessor and Seagate's new Cheetah 15K.4.

Fujitsu MAU3147 Review

#2 wattly

wattly

    Member

  • Member
  • 69 posts

Posted 21 November 2004 - 04:28 PM

I find it interesting that the new generation 15000 rpm drives are boosting single user scores so much, yet multi user scores are going down. I wonder if this was potential or not.

#3 chipstone

chipstone

    Member

  • Member
  • 150 posts

Posted 21 November 2004 - 04:43 PM

Those are some good number for single user workstation environment.

At last Fujitsu created a real exciting drive to save for: 146GB at 15K, I only imagine the cost; probably a bit more that a couple of bottles of Dom, and a lap dance.

#4 russofris

russofris

    Member

  • Member
  • 2,120 posts

Posted 21 November 2004 - 06:00 PM

Is it possible to see whether they have a "Server Mode / Desktop Mode" firmware switch like Seagate does? I'm having trouble finding docs.

Thank you fr your time,
Frank Russo

#5 Chewy509

Chewy509

Posted 21 November 2004 - 06:53 PM

Nice review...

These are my thoughts, on the loss of performance on the server side? Comments are welcome.

Since the intro of the Raptor, SCSI in general has taken a backward step on the desktop market. Could we be seeing the first evidence that manufacturers are trying to back back the workstation/desktop market from the Raptor? And since many corporations are sold on SCSI for servers, equal or slightly lower performance on the SCSI isn't going to be noticed by the larger corporate market, given that the newer models have a higher density. Lets face it, how many techs (in general) would understand the lower workings of their RAID setups, or even what drives are being used in their setups?

By re-enforcing SCSI superiority in the desktop field, also shows that SCSI is the only option for servers... (the PHB effect). So sales are NOT going to be lost on the corporate server market, by this move...

By showing that SCSI is still the best option for workstations, many companies won't switch from SCSI to Raptors on their next upgrades, saving lost sales to WDC and the Raptor.

Also I've been looking into many NAS boxes (for a future upgrade at work), and more and more models are utilising hotswap SATA drives instead on SCSI becuase of price/GB ratios. Also most NAS's focus on capacity rather than performance, esp when bottlenecked by either 100Mb or 1Gb LAN.

I'll now take off my tin-foil hat, and let the comments roll in...

#6 Jeff Poulin

Jeff Poulin

    Member

  • Member
  • 544 posts

Posted 21 November 2004 - 07:44 PM

Thanks as always for the nice reviews, Eugene. Just one correction that I could find:

"Even with segmentation settings switched to the default, server-oriented mode, however, the 15K.4's multi-user scores actually retreated from the levels set by the older 73-gigabyte flagships."

I think if you erase the first word, it would read better.

#7 isochar

isochar

    Member

  • Member
  • 227 posts

Posted 21 November 2004 - 08:11 PM

If you erase "even" then "however" should go too. ;) Only comment/complaint is that the Seagate drive wasn't shown with the desktop mode on in the desktop benchmarks.

#8 CityK

CityK

    Member

  • Member
  • 968 posts

Posted 21 November 2004 - 09:06 PM

However, even with segmentation settings switched to the default "server-oriented mode", the 15K.4's multi-user scores actually retreated from the levels obtained by the older 73-gigabyte flagships."

My work here is done.

Great Scott! What is that! Sentence fragment on Tech Report! This looks like a job for CityK! Up, Up and awwaaayyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy......

#9 lanolar

lanolar

    Member

  • Member
  • 4 posts

Posted 21 November 2004 - 09:28 PM

Anyone know how long until this hdd goes to market for sale? My HDD is got the click of death and I have been trying to hold out for this drive.

#10 isochar

isochar

    Member

  • Member
  • 227 posts

Posted 21 November 2004 - 10:37 PM

Might want to wait and see what the Atlas 15k II has to offer...

#11 KingGremlin

KingGremlin

    Member

  • Member
  • 368 posts

Posted 22 November 2004 - 12:15 AM

Maybe the problem with the server performance in the new generation of SCSI drives is the failure of the benchmark to accurately test it, and not the drives. IOMeter seems to be very limited in scope and has always struck me more as a glorified access time benchmark than a simulation of any real world usage. With the stagnation in access time with the new drives, it seams only natural that the IOMeter results basically stagnated as well and even slipped back a bit.

#12 Eugene

Eugene

    Administrator

  • Member
  • 1,057 posts

Posted 22 November 2004 - 01:22 AM

Only comment/complaint is that the Seagate drive wasn't shown with the desktop mode on in the desktop benchmarks.

This is a specific conscious decision that we made earlier when representing the Raptor in the Maxtor Atlas 10K V review. Should we include both TCQ and non-TCQ results since the latter does much better in the desktop area while the former does better in multi-user performance? We chose not to based on the fact that the review is that of the Atlas 10K V, not the Raptor and that the Raptor is marketed by WD not as an enthusiast-oriented drive but rather a server product. Interestingly, the decision generated some hoots and gloats at other places (Anandtech comes to mind... maybe KG recalls this exact instance) about how the 10K V "crushes" the Raptor. Its this case, in fact, that prompted the recent SR front page editorial.

Similarily, Seagate's concern remains the server market; desktop performance is incidental. In order to focus on the review drive at hand, Seagate's offerings, Savvio, Cheetah 10K, or Cheetah 15K, will include performance from only one setting, naturally the one that best optimizes the drive for its intended market.

#13 FaaR

FaaR

    Member

  • Member
  • 48 posts

Posted 22 November 2004 - 04:26 AM

Nice review as always Eugene, one thing that struck me tho - and this could be a HTML rendering issue - one of the bars in the performance graphs on the desktop performance page are incorrect on the two lowest graphs, specifically the bar second lowest in each of the graphs if I remember correctly. They extended all the way to the right, despite the score obviously wasn't matching even the drive just above, and much less the leader. :)

I use Firefox 1.0 on windows xp right now as I make this post.

#14 Eugene

Eugene

    Administrator

  • Member
  • 1,057 posts

Posted 22 November 2004 - 09:59 AM

Faar,

This is, unfortunately, a known issue with the gecko engine. Refreshing would have either fixed the problem or potentially extended out another bar instead. Its one of the reasons we're considering more standard graphs, but HTML-rendered graphs have so many ancillary benefits :(

#15 wattly

wattly

    Member

  • Member
  • 69 posts

Posted 22 November 2004 - 12:34 PM

Faar,

This is, unfortunately, a known issue with the gecko engine. Refreshing would have either fixed the problem or potentially extended out another bar instead. Its one of the reasons we're considering more standard graphs, but HTML-rendered graphs have so many ancillary benefits :(

Ah, I was curious about this too, but like you said a refresh will work, so I didn't think much about it.

#16 continuum

continuum

    Mod

  • Mod
  • 3,597 posts

Posted 22 November 2004 - 01:18 PM

Hmm. Interesting. I'll wait for the Atlas 15K II review...

Plus, 36GB is fine for me...

#17 HisMajestyTheKing

HisMajestyTheKing

    Member

  • Member
  • 5,000 posts

Posted 22 November 2004 - 04:49 PM

russofris = back?
PC is too old to brag with.

#18 Sivar

Sivar

    Moderator

  • Member
  • 3,465 posts

Posted 22 November 2004 - 06:05 PM

Faar,

This is, unfortunately, a known issue with the gecko engine. Refreshing would have either fixed the problem or potentially extended out another bar instead. Its one of the reasons we're considering more standard graphs, but HTML-rendered graphs have so many ancillary benefits :(

This is fixed in Gecko CVS, so will probably be in the next release of Firefox and Mozilla.
"No matter how far you've gone down the wrong road, TURN BACK."

#19 sdbardwick

sdbardwick

Posted 22 November 2004 - 07:54 PM

russofris = back?

russofris=BBH=occasional identity regression :)

#20 Eugene

Eugene

    Administrator

  • Member
  • 1,057 posts

Posted 23 November 2004 - 09:53 AM

My work here is done.

I've accepted City's submission ;) thanks!

#21 Maxtor storage

Maxtor storage

    Member

  • Member
  • 1,504 posts

Posted 23 November 2004 - 10:30 AM

Wow, the Single User graphs are just scary numbers. It seems this drive "is" the SCSI Drive of choise for those who like to use SCSI Hard Drives?

All though, I predict that Maxtor's AII15k will win all the STR Graphs and server performance. It probably has very high Desktop performance as well. I Hope it's the nest review to be shown!

#22 Serpent

Serpent

    Member

  • Member
  • 1 posts

Posted 23 November 2004 - 03:21 PM

I am currently looking to purchase 16 of the MAS3735NC's for a RAID array. Are these recommended over say the Seagate ST373453LC's? I am looking for performance as well as drive life. I don't have time to wait for the new drives to come out at this time.

Thank you,
Brad Dameron
SeaTab Software

#23 steve8

steve8

    Member

  • Member
  • 44 posts

Posted 24 November 2004 - 12:15 AM

I find it interesting that the new generation 15000 rpm drives are boosting single user scores so much, yet multi user scores are going down.  I wonder if this was potential or not.

i dont really see it as that purplexing.

multi-user apps are limited by access time
single user apps are limited by both access time and bandwidth


since these new drives dont improve on access time but do inprove on bandwidth....

we see single user performance increase while multi-user performance stagnate.

#24 Maxtor storage

Maxtor storage

    Member

  • Member
  • 1,504 posts

Posted 24 November 2004 - 12:26 AM

Yeah, but for some odd reason that can't apply completely. The newwer drives still hold the same access time numbers as their old ones... It could be the way they have been programed for these releases.

#25 wattly

wattly

    Member

  • Member
  • 69 posts

Posted 24 November 2004 - 12:56 AM

I find it interesting that the new generation 15000 rpm drives are boosting single user scores so much, yet multi user scores are going down.  I wonder if this was potential or not.

i dont really see it as that purplexing.

multi-user apps are limited by access time
single user apps are limited by both access time and bandwidth


since these new drives dont improve on access time but do inprove on bandwidth....

we see single user performance increase while multi-user performance stagnate.

The question is why did they improve on bandwidth and not the access time. I would have thought these drives would be marketed for multiuser server use, which as you say isn't helped as much by the improvments used in these drives. Infact, the multi user scores decreased, which would seem to hurt their main market. Although, it could just be as some have said, that it is related to the tests themselves. *shrugs*



0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users