Return to recnv.n CVS log

Up to [CSRG BSD Unix] / 43BSDTahoe / cci / pt

Annotation of 43BSDTahoe/cci/pt/recnv.n, revision 1.1.1.1

1.1       root        1: .1C
                      2: .nr PO 1i
                      3: .nr LL 6.5i
                      4: .bp
                      5: .TO  "DISTRIBUTION"
                      6: 
                      7: .FR  "Jim Walker"
                      8: 
                      9: .SB  "Effects of Record Compaction on PT Download"
                     10: 
                     11: .LP
                     12: A tool was developed which rebuilds the downloadable file for the PT
                     13: terminal.  Effectively it reduces the size of the  file by approximately
                     14: 24 percent (14000 bytes).  Logically what happens is that the data field
                     15: which is transmitted goes from 32 bytes to 256 bytes.  This appears to be
                     16: the point of diminishing return performance wise.
                     17: .LP
                     18: This tool (recnv) allows us to modify only the hexfile, without having to
                     19: change the EPROM, dlpt, pt.init or any other files.  The performance
                     20: gain varies with baud rate and whether or not reliable mode is used when
                     21: downloading.
                     22: .LP
                     23: Below I list the performance data gathered so far.  The first set of data
                     24: is using dlpt with the standard 3.02 download file.  The second set of
                     25: data is using the converted hexfile.  The third column were the downloads
                     26: using an CPU port (no percentages taken).  The fourth column is the performance
                     27: gain in percentage of Recnv over Standard.  The measurement were taken on
                     28: an Oneida runing PERPOS-P 2.01, while connected to a PVIOC-X port (all times arein minutes:seconds).
                     29: .LP
                     30: .TS
                     31: box, center;
                     32: c s s
                     33: c|c|c|c|c
                     34: l|n|n|n|n.
                     35: Recnv Performance Analysis
                     36: =
                     37: Condition      Standard        Recnv   CPU     Gain
                     38: _
                     39: 19200 -r       2:36    1:12    1:19    .54
                     40: _
                     41: 9600 -r        2:31    1:30            .40
                     42: _
                     43: 4800 -r        3:37    2:21    2:32    .35
                     44: _
                     45: 2400 -r        6:56    3:59    4:14    .43
                     46: _
                     47: 1200 -r        11:35   7:10    7:20    .38
                     48: _
                     49: 19200  1:28    0:40    1:05    .55
                     50: _
                     51: 9600   1:22    0:58            .29
                     52: _
                     53: 4800   2:29    1:44    1:48    .30
                     54: _
                     55: 2400   4:56    3:23    3:27    .31
                     56: _
                     57: 1200   9:03    6:30    6:46    .28
                     58: .TE
                     59: .SL
                     60: .CC "R. Privitera"
                     61: .br
                     62: B. Morgan
                     63: .br
                     64: J. West
                     65: .br
                     66: V. Shamapant
                     67: .br
                     68: M. Eadie
                     69: .br
                     70: D. Sheehan
                     71: .br
                     72: K. Karhuse

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.