Table 1 . Timing comparisons for triM and GridQTL.
| Number of Markers | GridQTL | triM | ||
|---|---|---|---|---|
| Number of Replicates Completed | Number Timed | Average Time (range) (hr:min) | Average Time (range) (sec) | |
| 1508 | 7 | 2 | 9:10 (7:20–11:00) | 45 (43–47) |
| 1400 | 6 | 3 | 4:39 (4:35–4:41) | |
| 1300 | 7 | 4 | 3:47 (3:13–4:16) | |
| 1200 | 9 | 9 | 2:33 (2:02–2:58) | |
Methods were tested on 10 replicates. Smaller datasets were produced by successively removing the last 100 markers. triM was able to analyze all replicates of the complete dataset with 1508 markers. GridQTL results are from version 1.4.1, running on the ECDF grid in March and April 2010. triM was run on a machine with a 2.66 GHz Intel Core i7 CPU. The current version of the cnF2freq codebase was used with OpenMP support disabled, compiled with gcc 4.2.1 and no specifically tuned optimization flags.