Lots Of CoreStatus = 66 (102) Errors Lately

[geokilla]

I've been getting a lot of CoreStatus = 66 (102) errors lately with the Notfred client. Resetting the client seems to fix it. It is however a PITA as the error usually occurs at around the same part of the WU.

[13:27:15] [email protected] Gromacs SMP Core
[13:27:15] Version 2.10 (Sun Aug 30 03:43:28 CEST 2009)
[13:27:15] 
[13:27:15] Preparing to commence simulation
[13:27:15] - Ensuring status. Please wait.
[13:27:25] - Assembly optimizations manually forced on.
[13:27:25] - Not checking prior termination.
[13:27:27] - Expanded 4839120 -> 23985345 (decompressed 495.6 percent)
[13:27:28] Called DecompressByteArray: compressed_data_size=4839120 data_size=23985345, decompressed_data_size=23985345 diff=0
[13:27:28] - Digital signature verified
[13:27:28] 
[13:27:28] Project: 2669 (Run 17, Clone 76, Gen 80)
[13:27:28] 
[13:27:28] Assembly optimizations on if available.
[13:27:28] Entering M.D.
[13:27:36] Multi-core optimizations on
[13:27:43] Completed 0 out of 250000 steps  (0%)
[13:47:36] Completed 2500 out of 250000 steps  (1%)
[14:07:10] Completed 5000 out of 250000 steps  (2%)
[14:26:21] Completed 7500 out of 250000 steps  (3%)
[14:45:45] Completed 10000 out of 250000 steps  (4%)
[15:05:25] Completed 12500 out of 250000 steps  (5%)
[15:25:20] Completed 15000 out of 250000 steps  (6%)
[15:48:00] Completed 17500 out of 250000 steps  (7%)
[16:08:00] Completed 20000 out of 250000 steps  (8%)
[16:27:53] Completed 22500 out of 250000 steps  (9%)
[16:45:26] 
[16:45:26] [email protected] Core Shutdown: INTERRUPTED
[16:45:30] CoreStatus = 66 (102)
[16:45:30] + Shutdown requested by user. Exiting.
[email protected] Client Shutdown.

[Dead Things]

It's a mysterious error that - while still being unresolved - tends to be associated with bad WUs. Is it always on P2669-R17-C76-G80 that you get this error, or does it occur on other WUs as well?

[geokilla]

Haven't noticed about that. I think I get them on WU too, but not sure since the log file resets when I reset the client.

Also, another error.

[17:40:30] Completed 30000 out of 250000 steps  (12%)
[17:59:56] Completed 32500 out of 250000 steps  (13%)
[18:19:17] Completed 35000 out of 250000 steps  (14%)
[18:21:07] CoreStatus = 0 (0)
[18:21:07] Client-core communications error: ERROR 0x0
[18:21:07] Deleting current work unit & continuing...
[18:21:21] - Preparing to get new work unit...
[18:21:21] + Attempting to get work packet
[18:21:21] - Connecting to assignment server
[18:21:21] - Successful: assigned to (171.64.65.56).
[18:21:21] + News From [email protected]: Welcome to [email protected]

[Dead Things]

CS0 errors are a result of an unstable overclock - and the 0x0 ones in particular are almost always a result of unstable RAM. Keep in mind, you can pass memtest for 52 weeks straight and still not be stable enough for FAH.