J.Ray.Scott@SEI.CMU.EDU.UUCP (06/24/86)
Introduction We have conducted a series of timings to gauge the relative performance of various PostScript printers. The three printers tested were the Apple LaserWriter and LaserWriter++ and the Dataproducts LZR 2660 that we purchased from Apollo as the DOMAIN/Laser-26. A description of the tests and comments follows the tables. The times in the tables are minutes:seconds to complete the print. Printer Timings: 6/1/86 External times (stopwatch -- timed from command to last page) name LaserWriter LaserWriter++ 2660 pages horizontal boxes 00:22 00:19 00:24 1 (1) vertical boxes 00:32 00:27 00:12 1 (1) c code 06:04 04:44 01:29 34 (2) c code (2 col, land.) 03:58 02:55 01:57 15 (2) TeX document 05:20 03:26 03:17 15 (3) Scribe manuscript 10:47 09:53 04:18 54 (2) Scribe document 20:30 13:32 08:10 71 (3) (1) Document was a single page hand-coded PostScript program (2) Document was a single-font document (3) Document was a multiple-font document Printer Timings: 5/28/86 Internal times (PostScript "usertime" command) name LaserWriter LaserWriter++ 2660 size (bytes) horizontal boxes 00:09 00:04 00:16 773 vertical boxes 00:19 00:14 00:07 772 c code 02:51 04:52 02:05 69008 c code (2 col, land.) 02:19 02:58 02:01 68025 TeX document 03:47 04:18 04:20 158747 Scribe manuscript 05:32 09:41 04:04 233800 Scribe document 13:38 13:24 08:03 362445 Timing methods: The first set of timings was done using a stopwatch, and timing the document from the time the document was sent to the printer to the time the last page was printed. The file was copied directly to the printer device line, using the unix cat facility. The second set of timings was done using the "usertime" function of PostScript. This time returns the execution time of the PostScript interpreter. The time is returned in milliseconds, however, anything less than a second was dropped for our purposes. The unix lpr facility was used to send the files to the printer. All files were already in PostScript form before printing. Document types: There were 3 types of documents that we used for our timings: single font documents, multiple font documents, and simple graphics. Single font document. These documents give us an indication of how fast raw text files could be printed. There is no overhead in these files for printing graphical characters or differing fonts (bold, italics, etc) Multiple font document. These documents give us an indication of how fast 'production quality' documents can be printed. These documents include multiple special fonts and differing font sizes, as well as special graphical characters. Simple graphics. These 2 files were originally created for a different purpose, but we decided to include them in our timings. Each file draws 21 rectangles of varying shades of gray, one drawing each horizontally across the page and the other drawing them vertically across the page. So if the colors ranged from 1 to 9, with 1 being the darkest color, the pages would look like the following. horizontal boxes vertical boxes ----------------- ----------------- | | | | | 111111111 | | 123456789 | | 222222222 | | 123456789 | | 333333333 | | 123456789 | | ... | | ... | | 999999999 | | 123456789 | | | | | | | | | | | | | ----------------- ----------------- Comment The first two tests with the simple graphics surprised us. In the original program we drew 30,000 lines instead of boxes. In this case the 2660 was 15 times slower than the LaserWriter. Since the 2660 feeds the long edge of the paper first while the LaserWriters feed the short edge first, we created the second test with vertical boxes and saw results more in line with what you would expect. TM's LZR-2660 is a trademark of Dataproducts Corporation DOMAIN/Laser-26 is a trademark of Apollo Computer Inc LaserWriter and LaserWriter Plus are trademarks of Apple Computer, Inc. PostScript is a trademark of Adobe Systems, Inc. Scribe is a trademark of Unilogic, Inc. TeX is a trademark of the American Mathematical Society