I was able to run a few quick tests on the Western Digital 100GB "Special Edition" hard drive with 8MB cache noted in Monday's news. I compared performance to the OEM IBM 60GB drive in a G4/733 Digital Audio system (overclocked to 800Mhz). I ran tests with the 100GB WD drive partitioned into 3 volumes and with the drive as a single partition with a copy of all the files installed on the OEM IBM 60GB drive. (Both drives were optimized with Speeddisk before the tests.)
This isn't as detailed as my normal full drive reviews, but hopefully is of some use. I've run some quick real world tests on the 100GB drive but would like to do more before returning the system . I'll update this page later with those results.
Macbench's CPU test has rightfully been criticized for being too cache sensitive, not MP or Altivec aware, etc., but the MacBench disk tests are one of the few Mac disk benchmarks that does more than sequential I/O. (This is why I've still used it, and some other sites have now started to again as well.)
The first graph below shows results of the full disk test series for the OEM 60GB IBM drive vs the 100GB WD drive (8MB cache model) on the same system with the same disk files. Both disks were optimized before the tests.
To see how well the WD drive performs across the platter, I then partitioned the drive into 3 volumes (each appx 31GB) and ran the test series on partition 1, 2 and 3. (There's some run/run variation possible, which is why on one test the 1st partion is a few % lower scoring than partition 2 and on one 512byte test, partition3 scored a few% higher. Also remember for small file I/O, the large cache of the drive makes up for the partition location.) The main thing shown in these tests is that the drive has good rates across the entire platter.
To repeat what I've said about this drive in another article, with an 8MB cache, the ATTO Tools benchmark maximum 8MB test sizes means it's not a good benchmark for this drive. (Since the max test file size is not larger than the drive's cache - unless the cache is segmented for Read/Write.) I did run ATTO's benchmark however just to see what the rates/curves looked like.
The read curve looks like it has a cache spike (up to about 512KB), but look at the flat write curve. Some IDE drives I've seen in Macs test in ATTO like there's no write cache enabled - but clearly not this drive.
Another ATTO test series run on the 100GB WD drive with 3 partitions, all blank. Note even with the 8MB cache drive and 8MB file size limit of this benchmark, you can still see a slight drop in read rates on the inner (3rd) partition, but the write caching shows no loss of performance.
Another 8MB Cache WD 100GB drive owner sent benchmark tests of the drive with a Promax ATA/66 card in a Quicksilver.
Western Digital Links Since these drives ship with no documentation or jumper settings info here's some links you may find useful. [Note: WD seems to have totally redone their site after this article was posted - all the old links below had to be changed in Nov 2001.]
After a phone call to WD, they say this drive marking does represent the 8MB model (despite a mfr date of July 18th, 2001 - several months ago)
Model No: WD1000BB-75CHE0
I do not have any model numbers from the 2MB cache model to verify if the above is unique to the 8GB model.
I welcome other owner reports on this drive. There was a previous page where with some WD 100GB (2MB cache model) IDE RAID performance reports. PC site StorageReview recently reviewed this drive (tested in PCs of course). (I wish there were Mac versions of the wide array of disk test utilities on the PC.)
Other Storage Related Articles/Reviews: For previous reviews/tests/guides on Storage (IDE, Firewire or SCSI) see the topics page links below or in the page header above.