thewizard75

First of all, did you check your bios configuration and make sure that ACPI is enabled, and that PnP OS is set to No? (the latter setting **shouldn't** matter, but it's how my system is configured). Beyond that, there are several other possibilities: 1) Automatic power off can be done through APM too. Try going to device manager, "view hidden devices" and you should see one entry that's disabled, called NT/APM Legacy APM (or something similar). Enable that, reboot, and your computer should at least power itself off. 2) Windows 2000/XP follow the ACPI standard much more closely; as such, there are numerous motherboards on which Windows 98 will use ACPI but that 2000/XP will not (because in fact these boards have buggy ACPI implementations). In short, your motherboard could be one of these; check for a bios update to correct this.

This could be a function of the 3com drivers for the network card. Or it **Could** be due to the RAID function of the Promise card (what is the raid configuration, etc.)? Note though that even 450MB/min = ~8MB/sec sustained, which shouldn't tax even one of those hard drives. The 125MB/min = ~2.2MB/sec you are getting makes it seem like it is NOT a disk bottleneck. Also, I assume all of the machines above have plenty of RAM?

I've seen this too, but never found a solution...

Are you on a network?

Or bootpart ... do a search for it online and you'll find it... it lets you use the NT/2000/XP bootloader to load linux (as long as you make sure during linux setup to install the boot loader on the first sector of the partition, NOT the MBR). I installed XP first, RedHat second, with zero problems using bootpart ...

Is the drive on the same chain as another hard drive or cd-rom drive? Do you have the jumper settings on the HDD set correctly (master, slave, etc.)? The latter can easily cause lots of these problems...

I'm assuming you are using the builtin w32time service to synchronize with an NTP server? If so, knowledge base article Q223184 desribes the registry entries that can be used to control the service (this applies to Windows 2000 and Windows XP). Under the registry key: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\W32Time\Parameters One of the values is Period : REG_DWORD or REG_SZ Used to control how often the time service synchronizes. If a string value is specified, it must be one of special ones listed below. 0 = once a day 65535, "BiDaily" = once every 2 days 65534, "Tridaily" = once every 3 days 65533, "Weekly" = once every week (7 days) 65532, "SpecialSkew" = once every 45 minutes until 3 good synchronizations occur, then once every 8 hours (3 per day) [default] 65531, "DailySpecialSkew" = once every 45 minutes until 1 good synchronization occurs, then once every day freq = freq times per day Just change this value to how often. Look and the knowledge base article or post here if this doesn't make sense.

Keep in mind that for two CPUs to be used in SMP mode, both should be the same stepping (you MIGHT be able to run with different steppings, but it's not recommended). Other than that, I know of one machine here based on the same VIA chipset that also had problems... replacing the motherboard fixed that. But that could have just been a faulty mb (and no fault of VIA's). How long do you wait to say it's locking up? Did you change from MPC to ACPI or vice versa?

I've seen this once before, on a similar system. Never did figure out what caused it -- just moved to a SCSI Cd drive. Maybe it's a power thing? There is an awful lot of stuff in that computer... The other possibility (slim) is that the IDE drives you are trying do not support multisession disks, and all of the disks you tried are multisession (unlikely).

NT based OS's like to have a 2MB pagefile on the system drive (c: in your case?) to deal with crashes and the like (it's necessary because of how crash dump works, etc). I wouldn't worry about it, just set it to 2MB on c: and a large 1024MB on your other partition.

What you are observing is the enhanced file system management in Windows NT/2000/XP. Large (compared to w9x) file caches (esp. write caches) are used under these operating systems to improve performance and responsiveness. As such, if the computer unexpectantly crashes, data can be lost. The file system driver is designed in such a way as to make sure that the data on the disk is always in a consistant state, but that means that if the computer crashes, some data can be lost. Now, if you have not already done so, I would recommend converting your filesystems to NTFS (a far superior fs compared to FAT12/16/32). If you want to prevent this loss of data from occuring on 2000/XP (with a performance penality, but I suspect the system will still be faster than under 9x), as administrator: Start -> Settings -> Control Panel -> System -> Device Manager Make sure "view devices by type" is selected. One of the tabs should be "disk drives" Open that. For each hard drive in the list, right click, select properties, and then go to the disk properties tab, and UNCHECK "write cache enabled". Then reboot the computer. This should prevent the loss of data due to write caching, but again there will be a performance penality (but no more than with 9x in theory). On a side note, I am surprised that you are having problems with crashes -- every 2000/XP machine I work with on a regular basis (10+) has never crashed (except once when I installed a faulty device driver -- rolling back to the old one solved that). Let me know if this solves your problem....

I assume you mean a choice between two 7,200 RPM drives in RAID 0 configuration vs. 1 10,000 RPM drive. Many factors affect hard drive performance, and what is reasonably required. For serious video editing or other large dataset applications, the striped drives (2 in RAID 0) would likely be the best candidate (best throughput). However, the latency will likely be higher than with a single 10,000 RPM drive. As such, if your need is very dependent on how fast (relatively) small pieces of data can be read from the disk, the 10,000 RPM drive is the better choice. If you need to read a large amount of data to/from the disk, the RAID 0 pair is probably better. Note that in a RAID 0 configuration, if one HDD fails, you lose all of the data, so data is slightly less reliable there. If I were building a machine today, I would go with a single 10,000 RPM hdd, but again this depends on your specific application.