Sage Client Crashes (Java experts needed!)

[Kirby]

I've been running into MANY java related (I assume) crashes of Sage the past month, and it doesnt appear to be JRE version related, as I have used 1.4.12, 1.6.0b99, and a few 1.5.x versions, all exhibit the same behavior.

The crashes never (that I can recall) occur while watching video. They always seem to occur while navigating through the menus, or if Sage is sleeping, or if its just sitting idle.

I am beginning to wonder if I have a hardware issue on my HTPC (maybe bad ram, or a memory controller going flakey or something), but I cant make sense of the Java crash reports. Attached are 3 sets of crashes, hs_err files renamed to correspond to the appropriate sageclient log file. If someone could please take a look at these and lend some advice, I would appreciate it.

I am trying to not over-burden the Support folks at Sage at this point, because this isnt a V6 beta issue, happened in V5 as well, and I know they are busy trying to work through V6 bugs.

[Kirby]

The last 2 logs here since you can only upload 5 files

[jreichen]

It looks like the garbage collector is crashing:
"VM_Operation (0x052cfb88): full generation collection, mode: safepoint, requested by thread 0x00ced210"

I only found two things about access violations in the Java forums on sun.com that might help you. They're not specifically about the gc crashing, though. Most questions about access violations go unanswered.

1. Try running the memory checker from http://oca.microsoft.com/en/windiag.asp.
2. 'You need to uninstall a windows update component named KB912945. Use add/remove programs, check the "show updates" checkbox on the top. The component is located at the bottom. Good luck.'

[Kirby]

Thanks for that info. I will burn that mem checker onto a CD and test when I get home. Dont have that component installed (KB912945).

Hopefully it will give me some useful results.

[Kirby]

Well I finally got around to testing my ram out with this tool. One of my 2 512MB dram's encountered 32 errors on the extended tests after running for 3 hours. So I will have to get some new ram, and hopefully that will solve the issue.

I am a bit surprised though, this ram had previously been torture tested when I ran this computer as an overclocked desktop, but maybe the higher voltage (2.8V) that I was pushing the ram with wore something out.

I think the biggest issue right now is that its in the first dram slot, which I would guess is being used up first by the system, so for a short term fix, possibly swapping the drams around would put the error prone chip in the 513-1024mb range, rather than in the first 512mb. But I'm not real sure how dual channel ram setup would affect that, if at all.