[NCLUG] July 1, 2003 talk - Linux System Performance

[jbass at dmsd.com]

some older data for grins ....

[jbass at dmsd dmsd2]$ cat mb.p60
mbench by John. L. Bass, DMS Design copyright 1985-1996
You are free to copy and use this program providing configuration info
and results are shared with the author by email to jbass at dmsd.com

Counts are per clock tick, presumed to be HZ=100
   1024 204552 100% 226343 100% |          |          |          |      *   @
   1536 204756 100% 235529 104% |                                       *
   2048 198793  97% 235765 104% |          |          |          |     *    |
   3072 203264  99% 226479 100% |                                       *   @
   4096 204697 100% 235670 104% |          |          |          |      *   |
   6144 203129  99% 211456  93% |                                      * @
   8192 202521  99% 215285  95% |          |          |          |     *  @ |
  12288  89845  43%  16969   7% |  @             *
  16384  61835  30%  16364   7% |  @       |*         |          |          |
  24576  39894  19%  16479   7% |  @    *
  32768  28214  13%  16479   7% |  @ *     |          |          |          |
  49152  22020  10%  16204   7% |  @*
  65536  18890   9%  16450   7% |  @*      |          |          |          |
  98304  16821   8%  16483   7% |  @
 131072  15181   7%  16467   7% |  @       |          |          |          |
 196608  14710   7%  16959   7% |  @
 262144  13926   6%  16993   7% |  @       |          |          |          |
 393216  10545   5%  16339   7% | *@
 524288   9243   4%  16290   7% | *@       |          |          |          |
 786432   8472   4%  16018   7% | *@
1048576   7943   3%  16278   7% | *@       |          |          |          |
1572864   7660   3%  16799   7% |* @
2097152   7408   3%  16816   7% |* @       |          |          |          |
3145728   7082   3%  16799   7% |* @
4194304   7204   3%  16076   7% |* @       |          |          |          |
[jbass at dmsd dmsd2]$ cat mb.p90.edo
mbench by John. L. Bass, DMS Design copyright 1985-1996
You are free to copy and use this program providing configuration info
and results are shared with the author by email to jbass at dmsd.com

Counts are per clock tick, presumed to be HZ=100
   1024 307029 100% 353796 100% |          |          |          |    *     @
   1536 307166 100% 348779  98% |                                     *    @
   2048 307257 100% 353627  99% |          |          |          |    *     @
   3072 307204 100% 346209  97% |                                     *    @
   4096 307464 100% 353747  99% |          |          |          |    *     @
   6144 300406  97% 328621  92% |                                    *   @
   8192 300665  97% 334078  94% |          |          |          |   *    @ |
  12288 181679  59%  38726  10% |    @                 *
  16384 135361  44%  38726  10% |    @     |     *    |          |          |
  24576  96377  31%  38765  10% |    @      *
  32768  73380  23%  38572  10% |    @   * |          |          |          |
  49152  59014  19%  37927  10% |    @ *
  65536  51988  16%  38041  10% |    @*    |          |          |          |
  98304  46455  15%  38194  10% |    @*
 131072  42428  13%  38042  10% |    @     |          |          |          |
 196608  41273  13%  38739  10% |    @
 262144  39700  12%  38700  10% |    @     |          |          |          |
 393216  24733   8%  37756  10% |  * @
 524288  20649   6%  37531  10% |  * @     |          |          |          |
 786432  17973   5%  37344  10% | *  @
1048576  16687   5%  37532  10% | *  @     |          |          |          |
1572864  15698   5%  37721  10% | *  @
2097152  15241   4%  38025  10% | *  @     |          |          |          |
3145728  14613   4%  38216  10% | *  @
4194304  14598   4%  37393  10% | *  @     |          |          |          |


notice memory subsystems are faster in raw counts, but are not keeping up
with processor technology.

2.xGHz numbers should be interesting, especially with rambus or faster
technologies.

John