SR has posted data in their Performance Database the benchmark runs on WD3200JD (They have also posted Hitachi's latest SCSI drives) Like Barracuda 7200.8, the Caviar 3200 lags its predecessor in performance. This is a disturbing trend. Maxtor has been able to buck the trend, but that is most likely due to the 16MB buffer, and these difficulties in getting performance out of 100GB+ platters may very well be the reason Maxtor opted to up the buffer to 16MB on DiamondMax10 and MaxLineIII. Hitachi and Samsung have 100GB and 125GB platter drives coming soon, respectively. It has been some time since the announcement of these drives, perhaps they are working on getting more performance out of them?
The latest iteration of SCSI drives, particularly 15K drives, had similar problems. However, they essentially broke even with their predecessors (Atlas 15K>15K II, Fujitsu MAS>MAU, and Cheetah 15K.3>15K.4) The 10K drives saw some pretty good sized gains, particularly Atlas 10K IV>10K V and Fujitsu MAP>MAT, but not on the scale of gains in the past. However, the latest crop of ATA drives are often slower than their predecessors. It is no secret that higher density requires more precise seeking, but in the past the makers have developed new techniques to improve seek speed while attaining the higher precision required for the higher density. However, they have not been able to this generation, at least not at first. Also, the Maxtor MaxLineIII review sample that SR had replaced because of its unusually high seek time serves to further reinforce this theory.
Kudos to Maxtor for delivering increased performance and increased density in the ATA market - they are the only maker that has been able to do both so far. Hitachi does have a 400GB drive, but it is a 5 platter drive using the older 80GB platters, so does not achieve a true increase in platter density - but also doesn't suffer from the apparent loss in performance by using denser platters. Seagate and WD need to go back to the drawing board and deliver a drive with performance at least equal to the last generation with the new platters, and I hope Hitachi and Samsung can at least hold the line on, ideally improve, performance with their new 100GB+/platter drives.
The latest iteration of SCSI drives, particularly 15K drives, had similar problems. However, they essentially broke even with their predecessors (Atlas 15K>15K II, Fujitsu MAS>MAU, and Cheetah 15K.3>15K.4) The 10K drives saw some pretty good sized gains, particularly Atlas 10K IV>10K V and Fujitsu MAP>MAT, but not on the scale of gains in the past. However, the latest crop of ATA drives are often slower than their predecessors. It is no secret that higher density requires more precise seeking, but in the past the makers have developed new techniques to improve seek speed while attaining the higher precision required for the higher density. However, they have not been able to this generation, at least not at first. Also, the Maxtor MaxLineIII review sample that SR had replaced because of its unusually high seek time serves to further reinforce this theory.
Kudos to Maxtor for delivering increased performance and increased density in the ATA market - they are the only maker that has been able to do both so far. Hitachi does have a 400GB drive, but it is a 5 platter drive using the older 80GB platters, so does not achieve a true increase in platter density - but also doesn't suffer from the apparent loss in performance by using denser platters. Seagate and WD need to go back to the drawing board and deliver a drive with performance at least equal to the last generation with the new platters, and I hope Hitachi and Samsung can at least hold the line on, ideally improve, performance with their new 100GB+/platter drives.