| 1 |
edhill |
1.1 |
|
| 2 |
|
|
|
| 3 |
|
|
The directories here contain shell scripts used to automate testing on |
| 4 |
|
|
various systems. The directory names approximately match the machine |
| 5 |
|
|
or cluster "nicknames" for which they are intended. For a description |
| 6 |
|
|
of the machines and the latest testing output, please see: |
| 7 |
|
|
|
| 8 |
|
|
http://mitgcm.org/testing.html |
| 9 |
|
|
|
| 10 |
|
|
|
| 11 |
|
|
|
| 12 |
|
|
====================================================================== |
| 13 |
|
|
|
| 14 |
|
|
NOTE: mpack problems |
| 15 |
|
|
|
| 16 |
|
|
The mpack utility is used by testreport to mime-encode and email the |
| 17 |
|
|
testing results back to the mitgcm.org server. There appears to be a |
| 18 |
|
|
problem with the mpack utility on some 64-bit systems. Notably, |
| 19 |
|
|
emails sent from the following systems: |
| 20 |
|
|
|
| 21 |
|
|
SGI Altix |
| 22 |
|
|
AMD Opteron |
| 23 |
|
|
|
| 24 |
|
|
will consistently result in errors ("MIME file corrupted in transit"). |
| 25 |
|
|
Perhaps the mpack utility is not "64-bit clean" and this is |
| 26 |
|
|
interpreted by munpack as an error during transmission? |
| 27 |
|
|
|