Indra Performance Relative to Memory InstalledBob Pentecost10/5/98
Summary
These tests were run to demonstrate the effect of adding memory to the printer when printing complex documents. The results show, as expected, that performance improves with more memory. The most complex documents show the most improvement (175–215% were the best cases). However, once a significant performance gain was made, there was little advantage to adding more memory; in fact, a very slight performance degradation was noted in several cases.
Given this limited testing, it is difficult to draw a conclusion as to how much memory is enough. For performance, 128M appears to be enough for even very complex, large format documents. For all except the most complex letter size documents, 64M seems to be adequate.
Test Method
The time measured includes the time to transfer the print data to the printer, format the data, print the sheet(s) and deliver them to the output bin. The time for the client computer to generate the print data was not included.
The documents used were as follows:
Document
Application
Pages, Format
Content
Print file size (Megabytes)
psalq.tif
Corel PhotoPaint
1, Letter, Simplex
Single page raster graphic.
94
rgb102.ps
None (raw PostScript)
Demo page, mixed text and graphics.
0.29
porsche.pub
MS Publisher
2, Letter, Duplex
Newsletter, mixed text and graphics.
10
fleet.ppt
MS PowerPoint
30, Letter, Duplex
Presentation, mixed text and graphics.
18
Tips and Tricks.qxd
Quark Express
8, 11x17, Duplex
Booklet, mixed text and graphics.
25
success4.qxd
1, 12x18, Simplex
Single page, some text but mostly graphics.
73
Test Results
Windows 95
Print Time
Performance Improvement
32M
64M
128M
256M
32M to 64M
64M to 128M
128M to 256M
0:14:51
0:12:27
0:11:02
0:10:58
119%
113%
101%
0:02:57
0:02:16
130%
100%
0:05:47
0:03:10
0:03:03
0:03:04
183%
104%
99%
0:13:14
0:11:59
0:11:10
0:11:51
110%
107%
94%
0:25:15
0:18:01
0:18:04
0:17:30
140%
103%
out of memory
0:17:22
0:08:04
scerbaty