For those who are running any of the older releases, please reset the RALPH project to pick up the new bits.

This release contains all the fixes from the previous release plus some additional enhancements to track down the '1% bug.'

When a workunit is aborted additional information is going to be logged with the result when it is reported back to the server.

Here is what some of that additional information looks like. This additional information shows where the program was at and what it was doing at the time that you aborted it. With enough of these we should get a pretty good idea where the '1% bugs' are comming from.

The previous method for dealing with this bug involved using a debugger on a remote machine and creating a usermode dump file. For those who don't know what that is, it is a 350MB file that contains all the data the program was managing at the time the dump was taken. It proved to be to labor intensive to walk many people through the process and get the dump files over email. So with a little bit of tweaking last night I adjusted things so that I could get the information when you all abort the workunits.

So if you abort a workunit because of the '1% bug' on Ralf please log the result id here

Thanks in advance.

----- Rom

---

I am running Rosetta with RALPH 4.94. When Ralph is running, Rosetta is preempted, and Task Manager shows no CPU activity. However when Rosetta is running, RALPH shows CPU usage in Task Manager, but BOINC Manager shows "Preempted" and CPU time is stagnant.

Since we're all concentrating on the 1% bug; and the fact that Dr. Baker is helping create 5 more spots in the code to register the progress; I was wondering if you could update the progress bubble - especially for those of us running the client as a service. Would it be possible to add the following information to the percent done message: the current atom/building block we're working on (23 of 89) and the time since the last step was completed: (22 of 89 was finished 25 mins ago)

i.e. "2%" "Step 23 of 89" "25 mins on Step 23"

Which should make it easier to tell when we've run into the 1% bug - even if after the new code it is no longer being at 1% that the hangup is reported.

---

I tried to start 4.94 on this computer

Http://ralph.bakerlab.org/show_host_detail.php?hostid=2035

and ended up with this.

2006-03-23 19:09:45|ralph@home|Process creation failed: En enhet som är ansluten till datorn fungerar inte. (0x1f)
2006-03-23 19:09:45|ralph@home|Process creation failed: En enhet som är ansluten till datorn fungerar inte. (0x1f)
2006-03-23 19:09:46|ralph@home|Process creation failed: En enhet som är ansluten till datorn fungerar inte. (0x1f)
2006-03-23 19:09:47|ralph@home|Process creation failed: En enhet som är ansluten till datorn fungerar inte. (0x1f)
2006-03-23 19:09:47|ralph@home|Process creation failed: En enhet som är ansluten till datorn fungerar inte. (0x1f)
2006-03-23 19:09:48|ralph@home|Unrecoverable error for result HB_BARCODE_30_1bq9A_354_46_0 (CreateProcess() failed - En enhet som är ansluten till datorn fungerar inte. (0x1f))
2006-03-23 19:09:48|ralph@home|Deferring scheduler requests for 1 minutes and 0 seconds
2006-03-23 19:09:48||Rescheduling CPU: start failed

It comes up a windows message saying that KERNEL32.DLL does not exist.

Any way to fix it?

Anders n

---