LLXX

With it disabled, the HDD operates in write-through cache mode, in which all data that gets written to the drive is immediately written to the disks and also stored in the cache. Writes are not cached, but reads are. When the option is enabled, the HDD operates in write-back cache mode, in which all the data that gets written to the drive is first stored in the cache, and then later written to the disk. Both writes and reads are cached in this case. I recommend disabling it, to reduce chances of data corruption.

Nothing especially malicious in the HJT log, but you might want to clean up these missing-file entries: O23 - Service: iPodService - Unknown owner - C:\Program Files\iPod\bin\iPodService.exe (file missing) O23 - Service: Maya 7.0 Documentation Server (maya70docserver) - Unknown owner - C:\Program Files\Alias\Maya7.0\docs\wrapper.exe" -s "C:\Program Files\Alias\Maya7.0\docs\Wrapper.conf (file missing) O23 - Service: Remote Packet Capture Protocol v.0 (experimental) (rpcapd) - Unknown owner - %ProgramFiles%\WinPcap\rpcapd.exe" -d -f "%ProgramFiles%\WinPcap\rpcapd.ini (file missing) Does this problem still happen whenever you open a .mov file in Quicktime? And is it only one file that's causing this problem? Does opening .mov in another program (try notepad, it obviously won't play but it'll read the file from the HDD) produce this error?

I decided to do a stress test on one of my older machines, a Celeron 2.4 that's been running at 3.2GHz for the past few months. I started the Prime95 stress test with the temperature monitor on and approximately 10 minutes into the test, the machine started to sing. At first it was a high-pitched intermittent warbling sound, which then developed into a rather loud squealing - like a "pig being tortured". The temperatures were still normal: CPU was at 42°C and CPU fan was still running at ~3600RPM, Vcore at ~1.575. I immediately pulled off the side cover, but everything looked alright. There was no obvious source of the squealing - the whole case seemed to be emitting this rather terrifying sound. All the fans were operating normally, none of the chipset heatsinks were hot, but the voltage regulators slightly warm. After a few minutes the squealing became a siren-like whine, increasing and decreasing in pitch. The CPU continued to heat - it was at 50°C now. Then, the whining turned into squealing again and bck to the warbling sound it started with, and finally disappeared when the CPU was at 55°C. For the rest of the 4 hours the CPU remained around 60°C and the sound was not heard from again. I restarted the test after the CPU cooled back down to idle (30°C), and again 10 minutes in the sound reappeared. It seems to be emanating from the case itself, but the case does not seem to be vibrating. All the other components are also silent. Stopping the test, letting the CPU cool, then starting it again returns the sound. I've also tried other CPU-stress-testing programs and they also exhibit this behavior - sound starts ~40° temperature and disappears at around 55°. System: Celeron 2.4 Northwood @ 3.2 256Mb DDR400 Gigabyte GA8S655FX mobo What exactly is going on here?

A spill of pure water will probably be recoverable, just wait a while for it to *completely* dry. It may help to remove the entire keyboard assy. from the case and check to see if any moisture remains on the mobo below. Let it dry in a warm area (not HOT!) for at least 2 days. Did you mention it was spilled while the lid was closed? Check for moisture on the mobo. The keyb itself is probably not at fault here.

Is it the MBR or just the partition bootsector? If it's the MBR, it's going to be easier, but if you understand structure of partition table, you can reconstruct the MBR manually with a disk editor. I've done this several times, when automated recovery just won't work. If it's the partition bootsector, it'll be slightly more difficult as there are more values to fill in. Normally I'd put in what I think are the correct values, reboot and test, and if the disk is not recognised, change the values and try again. As suggested above, a sector-by-sector image of the drive is ESSENTIAL. The OS may attempt to write to a partially-recovered filesystem and corrupt it more.

See Application Installs section...

You will need to extract the original Win32k.sys from the fileset.

FlashGet is excellent download manager. I have v1.50 which seems to have been the best. Tools -> Options -> Proxy Edit the proxy options to match those displayed in Internet Connection Properties -> Connection.

I think that this "warning" is moot now. It was most likely formulated at a time when some manufacturer's drives had write-back caches, and old software caches e.g. SMARTDRV had to be manually "flushed" before powering down. Some people had problems with corruption, so they reported it to M$ and this is the result. But I don't think this problem occurs anymore now. A good way to test if your drive is write-through or write-back is to perform read-speed and write-speed testing (bypassing the OS's disk cache). Almost all drives will "spike" on read speed as the data is being read from the cache, then decrease to a steady rate which indicates the transfer rate from the platters themselves. On write, if the write speed spikes up and then goes down to a steady rate after a long time, it's probably write-back. If the write speed remains constant from start to finish, it's write-through. All the drives I've tested exhibit the latter characteristic.

It must depend on the exact mobo and CPU design... 2Gb with borrowed RAM for 3 months with no problem here I'm going to see if I can borrow another gig and try 3GB, since that's what my mobo claims to support. The addressing limit of every CPU since the i386 was 4Gb, but the design of the PC architecture limits it to approximately 3Gb of physical memory. The upper 1 Gb is reserved for I/O ports and other miscellaneous items.You don't need the unofficial service pack to work with lots of RAM - it's all in the MaxFileCache setting. A stock install of 98se will work as long as you set MaxFileCache=524288 or less.

i mention it because if MSBackup is installed removing these files (assuming you're not using tape backup) will greatly increase 98SE boot up times. INcrease? As in make faster or make slower?

Can you see any process taking a high percentage of the CPU in the task manager?

What did you do to cause this error?

Noise is an inherent quality of the image sensors used in webcams. If you want better quality, get a proper digital camera.

/f must be used to apply fixes...

Intel... of course. They have thermal limiter which halves and then quarters the clock speed if the temperature gets too high. Your CPU is running at 1/2 or even 1/4 of its normal speed, which is why everything slows down and CPU usage goes to 100% since it's much slower now.

Don't scanners and cameras come with their own drivers already?

Hmm... really?Now I'm starting to suspect an overheating chipset.

What's the CPU model and what temperature is it running at?

Then it's safe to assume that your memory and CPU are working correctly. The problem is either in some other hardware or the software. What about posting HiJackThis log and doing a virus scan?

Intel CPUs don't have anything to short out on the top of the package. Unless you really did put so much that it ran down the sides and onto the mobo, there should be no problem.All those random bugchecks do point to a hardware problem though... software problems usually don't cause such a wide variety of errors.

I'll let him decide, seeing as how I recovered his entire script, complete with comments and formatting, inside the compiled EXE

I'm not sure about their PSUs, but they make some good CPU coolers, like this one: http://www.msfn.org/board/index.php?showtopic=66599

Corrupted program files are *not* going to be easily repairable. The damaged area is probably at least several hundred bytes long, and you have no idea what those bytes used to be. Try searching the Internet to see if you can find a download.

According to the spec link you gave it's 180 watts? That's a bit on the small side... even small PCs should have at least 200W.