Subject: ZIP Drive Benchmark Enclosed From: cam@skyview.bison.mb.ca (Cam Giesbrecht) I just purchased a ZIP drive today, so the following is a benchmark of the Zip drive compared to an internal and external hard drive, as well as the standard 1.44 MB floppy disk. ---------------------------- [Internal] Apple (Sony) 1.44 MB FDHD Floppy Disk Drive ---------------------------- Volume: untitled . . . . . . . . . . . . . . . . . Size: 1415K test size = 512K (using a temporary contiguous file of size 512K) Pass 1: Latency = 50.00 ms (600 RPM) Ave. Seek = 43.08 ms, (access = 93.08 ms) Max. Seek = 273.33 ms, (access = 323.33 ms) Write transfer rate = 61.6 KBytes/Sec. Read transfer rate = 78.6 KBytes/Sec. Simulated "Typical" rate = 56 KBytes/Sec. ---------------------------- [External] Iomega Zip SCSI Drive (w/100 MB Cartridge) * Formatted w/Zip Tools v 4.2 * ---------------------------- Volume: Zip Disk 1 (w/Zip Tools) . . . . . . . . . Size: 96507K test size = 512K (using a temporary contiguous file of size 512K) Pass 1: Latency = 10.16 ms (2952 RPM) Ave. Seek = 34.09 ms, (access = 44.25 ms) Max. Seek = 70.50 ms, (access = 80.66 ms) Write transfer rate = 1084 KBytes/Sec. Read transfer rate = 1123 KBytes/Sec. Simulated "Typical" rate = 148 KBytes/Sec. ---------------------------- [Internal] Apple Quantum SCSI (160 MB LPS) ******************************************************************* * NOTE: This volume is partitioned into 2 partitions of: * * - 120 MB startup partition * * - 40 MB data partition * * The test was performed on the larger startup partition. * ******************************************************************* * Formatted w/La CIE's SilverLining v 5.4/14 * ---------------------------- Volume: Mac OS* 7.5.1 . . . . . . . . . . . . . . Size: 118872K test size = 512K (using a temporary contiguous file of size 512K) Pass 1: Latency = 8.41 ms (3567 RPM) Ave. Seek = 14.00 ms, (access = 22.41 ms) Max. Seek = 15.59 ms, (access = 24.00 ms) Write transfer rate = 1497 KBytes/Sec. Read transfer rate = 1850 KBytes/Sec. Simulated "Typical" rate = 189 KBytes/Sec. ---------------------------- [External] APS Quantum SCSI (160 MB ELS) * Formatted w/FWB Hard Disk Toolkit Personal Edition v 1.7 * ---------------------------- Volume: Quantum HD . . . . . . . . . . . . . . . . Size: 164778K test size = 512K (using a temporary contiguous file of size 512K) Pass 1: Latency = 8.16 ms (3676 RPM) Ave. Seek = 20.75 ms, (access = 28.91 ms) Max. Seek = 23.50 ms, (access = 31.66 ms) Write transfer rate = 1367 KBytes/Sec. Read transfer rate = 1367 KBytes/Sec. Simulated "Typical" rate = 206 KBytes/Sec. ---------------------------- NOTE: These tests were performed on a Macintosh Quadra 605 with 8 MB of RAM (16 w/RAMDoubler), 32 K Cache, File Sharing Off, about 50 Extensions (basically a real-world test), and no apps running in the background. Test results were obtained using TimeDrive 1.3. TimeDrive 1.3 creates a temporary contiguous scratch file. It then executes single large block reads and writes, measuring the time and converting to Kilobytes per Second. Seeks are restricted to a single volume (partition) on a drive. A volume that is smaller than the total drive will yield noticeably smaller seek times than the drive specification. (This is another advantage of partitioning hard drives.) NOTE: Total time for a disk transfer is: T = Seek Time + Latency + Data size/transfer speed Latency is the time for data to rotate under the heads. (On average, this is 1/2 the disk revolution time). Average seek times are measured by reading 200 random locations within the selected volume. Maximum seek times are measured by reading the first and last location within the selected volume. Data Transfer Speed measures how fast data is transferred between your drive and the Mac once the drive gets there. (transfer times vary depending on the drive/Mac combination) Simulated System measures how fast a sequence of randomly selected, yet typical, disk transfers take. The transfers invlove a small number of single block requests (system resources) folowed by a number of 1-32 block requests (application requests). The number reported, the "effective transfer rate," is lower than read/write rates since seek and latency times are included in the calculation. BONUS: The data read during Read Transfer rate measurements is checked against the known random data written. This thoroughly tests the reliability of storing and retrieving data on the selected volume. Program by Roger D. Bates. (C)1990-1991. La Cie, Ltd. All rights reserved.