G4 MDD kernel panic diagnosis

Question

Hi,
I have a single cpu 1.25ghz MDD G4. This machine has been having some kernel panic problems that seem to be very much tied to sustained heavy CPU load. It will run for days on end doing simple tasks with occasional high CPU processes. If I start doing something intensive (like encoding a big batch of audio files) it will certainly crash after some time.
I thought maybe this was tied to temperature somehow so I've been monitoring the cpu temp. It does seem to never panic when the temp is between 25 and 40 C, but when it gets to the mid 40s it is only a matter of time. This doesn't seem hot enough to me to be a problem. I reapplied some thermal paste to the CPU and made sure the heat sink was seated well, with no improvement. The main fan and the PSU's fans seem to be functioning and the system is free of dust.
I'm at a loss here, but I'm definitely not an expert. Maybe this temperature issue is coincidental but it seems to truly indicate an impending panic. Could the PSU be overheating? Is this just a sign of a failing CPU?
Also, the backtrace in all of the panics is the same. What is the significance of this? Below is the system information and a couple of panic logs.
Thanks in advance for any assistance.
Model Name: Power Mac G4
Model Identifier: PowerMac3,6
Processor Name: PowerPC G4 (3.3)
Processor Speed: 1.25 GHz
Number Of CPUs: 1
L2 Cache (per CPU): 256 KB
L3 Cache (per CPU): 1 MB
Memory: 1.5 GB
Bus Speed: 167 MHz
Boot ROM Version: 4.4.8f2
Sun Sep 14 16:38:30 2008
panic(cpu 0 caller 0x000AED04): "Uncorrectable machine check: pc = 0000000009AD1FCC, msr = 000000004200D930, dsisr = 42000000, dar = 00000000049CD000\n" " AsyncSrc = 0000000000000000, CoreFIR = 0000000000000000\n" " L2FIR = 0000000000000000, BusFir = 00000000ff000000\n"@/SourceCache/xnu/xnu-1228.5.20/osfmk/ppc/trap.c:975
Latest stack backtrace for cpu 0:
Backtrace:
0x0009B498 0x0009BE3C 0x00029DD8 0x000AED04 0x000AF514 0x000B2A78
Proceeding back via exception chain:
Exception state (sv=0x29f91c80)
PC=0x09AD1FCC; MSR=0x4200D930; DAR=0x049CD000; DSISR=0x42000000; LR=0x09AD101C; R1=0xBFFFDF20; XCP=0x00000008 (0x200 - Machine check)
BSD process name corresponding to current thread: Camino
Mac OS version:
9E17
Kernel version:
Darwin Kernel Version 9.4.0: Mon Jun 9 19:36:17 PDT 2008; root:xnu-1228.5.20~1/RELEASE_PPC
System model name: PowerMac3,6
Tue Sep 2 05:44:07 2008
panic(cpu 0 caller 0x000AED04): "Uncorrectable machine check: pc = 000000009340A360, msr = 000000004200F030, dsisr = 42000000, dar = 0000000004E7F000\n" " AsyncSrc = 0000000000000000, CoreFIR = 0000000000000000\n" " L2FIR = 0000000000000000, BusFir = 00000000ffffffff\n"@/SourceCache/xnu/xnu-1228.5.20/osfmk/ppc/trap.c:975
Latest stack backtrace for cpu 0:
Backtrace:
0x0009B498 0x0009BE3C 0x00029DD8 0x000AED04 0x000AF514 0x000B2A78
Proceeding back via exception chain:
Exception state (sv=0x306f0780)
PC=0x9340A360; MSR=0x4200F030; DAR=0x04E7F000; DSISR=0x42000000; LR=0x9340A328; R1=0xBFFFEE30; XCP=0x00000008 (0x200 - Machine check)
BSD process name corresponding to current thread: WindowServer
Mac OS version:
9E17
Kernel version:
Darwin Kernel Version 9.4.0: Mon Jun 9 19:36:17 PDT 2008; root:xnu-1228.5.20~1/RELEASE_PPC
System model name: PowerMac3,6
Sat Sep 13 12:30:00 2008
panic(cpu 0 caller 0x000AED04): "Uncorrectable machine check: pc = 000000000670D360, msr = 000000004200F030, dsisr = 42000000, dar = 0000000003B83000\n" " AsyncSrc = 0000000000000000, CoreFIR = 0000000000000000\n" " L2FIR = 0000000000000000, BusFir = 00000000ffffffff\n"@/SourceCache/xnu/xnu-1228.5.20/osfmk/ppc/trap.c:975
Latest stack backtrace for cpu 0:
Backtrace:
0x0009B498 0x0009BE3C 0x00029DD8 0x000AED04 0x000AF514 0x000B2A78
Proceeding back via exception chain:
Exception state (sv=0x3013c500)
PC=0x0670D360; MSR=0x4200F030; DAR=0x03B83000; DSISR=0x42000000; LR=0x0670535C; R1=0xF032C0F0; XCP=0x00000008 (0x200 - Machine check)
BSD process name corresponding to current thread: XLD
Mac OS version:
9E17
Kernel version:
Darwin Kernel Version 9.4.0: Mon Jun 9 19:36:17 PDT 2008; root:xnu-1228.5.20~1/RELEASE_PPC
System model name: PowerMac3,6

Answer 1

japamac wrote:
You said that you replaced the thermal compound. Depending on the type of compound, curing time can require 200-300 hours before full thermal conductivity is reached. Arctic Silver MX-2 is the best compound, and requires no curing time.
I used Arctic Silver 5
What type of heatsink is in your machine? There is a thick finned aluminum heatsink, a thin finned AL heatsink (much better than the thick finned one) and a copper heatsink used in MDD's. Is the heatsink free from dust?
It has the thin finned aluminum heatsink and I've used an entire can of air blowing all the dust I could find out of this machine.
Does the 120mm case fan turn properly? Is it free from dust and are the blades clean? A new case fan (120mm x 38mm ≧ 88cfm) may help.
It seems to be clean and functioning, but I wonder if cpu temperature is truly the issue.
A PSU that is clogged with dust, or one that has fans that aren't performing will overheat. An overly hot PSU cannot produce reliable voltage, which will cause a system to freeze.
Use compressed air to blow out the PSU.
I will get some more canned air and try this tonight. I don't like that the PSU intakes air from inside the top of the case, either. It seems like it may be kind of warm up there.
See G4 System Reviews and Related Articles for advice on MDD cooling.
Thanks for this!