6  Speed Matching Performance 6.1  Introduction This section shows the benefits of speed matching in reducing the number of repositions and improving throughput performance.  The comparison is done using the same IBM Ultrium 2 drive by executing the test case twice.  One test is run with the default of speed matching enabled.   The other test is run with speed matching disabled, and with the drive set to its highest velocity.   The tests are executed using a benchmarking tool that writes fixed blocks from memory.  For this test a 1:1 compressible data pattern was used. Data transfer rates and number of reposition events were collected at net host data rates between 1-40 MB/sec in 1 MB/sec increments.  At each host data rate, 6000 MB of data was written to the tape device. 6.2  Speed Matching Performance Summary The following summarizes the performance results with the IBM Ultrium 2 with speed matching enabled when compared to speed matching disabled.  Refer to Section 6.4 for detailed results. The IBM Ultrium 2 demonstrates the following performance improvements due to speed matching. y Between 17-32 MB/sec the average throughput performance benefit is 13% y Between 1-32 MB/sec the average reposition reduction benefit is 68% 6.3  Test Environment The description of the test environment is as follows. AIX 5.1.0.25 Operating Systems: FC 6203: Dual-Channel Ultra3 SCSI (Type 4-Y) Placement:  64 bit 3.3V Slot at 66 MHz Adapter: 7025-6F1 pSeries 620 Processor: Hardware and Software Environment Page 15 of 20 p1   p2   p3   p4   p5   p6   p7   p8   p9   p10   p11   p12
p13   p14   p15   p16   p17   p18   p19   p20