betamax

OK will do. I set the option to create a complete dump, and set the min and max of my page file to 2096MB (as windows reports me as having 2046MB-- I thought 2gigs was 2048, but I guess not according to Microsoft Math). I also uninstalled the Intellimouse drivers and restarted, letting windows detect the mouse again and install drivers for it. So far no BSOD, but if it happens again I should have a complete dump for it. Thanks for the help.

Hello, Recently, I have been getting the PAGE_FAULT_IN_NONPAGED_AREA (0x50) when starting up my XP system. If I reset the computer, it starts up fine. However, if I have been running the computer for a few days, shut it down, then start it back up, I will get the error again on occasion. Windows generated a "minidump" file (see attached) but I have no idea how to view it. I looked at it with a text editor and it was nothing but garbage. Is there any other info I can provide to help determine the problem? I googled around for this error message and found a lot of hits, however none of them matched my situation. It seems this error can be thrown under various scenarios, but none seemed to cover it being thrown during boot. Minidump file attached. For some reason the board won't let me upload the .dmp file as is, so I placed it in a .rar and uploaded it. Thanks in advance for your help! -Paul EDIT: uploaded a copy of the same file in .zip format EDIT2: I should also mention that this started happening after I installed a new USB mouse. It is the exact same make and model mouse (Microsoft Optical IntilliMouse USB), just new (the buttons on the old one were starting to stick). All I did was shut down my PC, unplug the old mouse, plug in the new mouse, start the computer back up, and there was the BSOD. I don't have the old mouse anymore, perhaps I should try starting the PC w/o the new mouse attached and see if it still occurs. Mini060308_01.rar Mini060308_01.zip

Well it looks like I might have answered my own question. So for those with the same problem, here is what I did: I poked around and found a file located at c:\windows\inf\sysoc.inf. One section in particular to note is: Games=ocgen.dll,OcEntry,games.inf,HIDE,7 AccessUtil=ocgen.dll,OcEntry,accessor.inf,,7 CommApps=ocgen.dll,OcEntry,communic.inf,HIDE,7 MultiM=ocgen.dll,OcEntry,multimed.inf,HIDE,7 AccessOpt=ocgen.dll,OcEntry,optional.inf,HIDE,7 Pinball=ocgen.dll,OcEntry,pinball.inf,HIDE,7 MSWordPad=ocgen.dll,OcEntry,wordpad.inf,HIDE,7 ZoneGames=zoneoc.dll,ZoneSetupProc,igames.inf,,7 I commented out the "Games=" line by putting a semicolon in front of it. Then tried again. This time it complained about pinball.inf. So I did the same thing to the pinball line. It also complained about igames.inf, so I commented out the igames line as well. That section now looks like this: ;Games=ocgen.dll,OcEntry,games.inf,HIDE,7 AccessUtil=ocgen.dll,OcEntry,accessor.inf,,7 CommApps=ocgen.dll,OcEntry,communic.inf,HIDE,7 MultiM=ocgen.dll,OcEntry,multimed.inf,HIDE,7 AccessOpt=ocgen.dll,OcEntry,optional.inf,HIDE,7 ;Pinball=ocgen.dll,OcEntry,pinball.inf,HIDE,7 MSWordPad=ocgen.dll,OcEntry,wordpad.inf,HIDE,7 ;ZoneGames=zoneoc.dll,ZoneSetupProc,igames.inf,,7 And now I am able to access the add/remove windows components program (incidentally the program can be executed from the command line as sysocmgr.exe.)

I made an nLite'd CD a while back and chose the option to omit games from the install. Later on down the road I tried to add or remove a windows component, and am greeted with a nice message. It seems that file was needed after all. It is not in my nLite'd CD. So my questions are: Why is it looking for games.inf? Can I stop it from looking for games.inf via an ini file or registry entry? I googled this error message before posting here. I actually found a lot of hits. However, the one thing in common among all the hits was that nobody replied with any useful information other than copying the games.inf file from the windows cd to c:\windows\inf. But my cd doesn't have that file because I created an nLite'd with the option to remove games. Will this thread join the ranks of the countless others like it? Only time will tell...

bumped... I REGRET NOTHING!!!

bumped without shame or remorse

But the problem is: the copy destination is determined by the SourcePath entry under [ComputerSettings]. However, when running factory -winpe, the SourcePath entry isn't read (see image above). How then, is factory.exe supposed to know where to copy the source files?

shameless bump

It starts off by saying this: Ok, fine. So we use winnt32 from WinPE. But then it says: Alright so... when do we get to run winnt32? What happened to running winnt32? Just wait.. it gets better: Ok so if the Factory tool just used winbom.ini to install windows, why would we run winnt32? Now take a look at this.. remember according to the quotes above, we are running Factory -winpe. But take a look at the highlighted section of the table. You see that? First it says SourcePath under [ComputerSettings] is used to determine the location of the i386 folder. But if you look at that table, in winpe mode, we don't process that entry. Anyone know what REALLY happens when you run factory -winpe?

It's a shame nobody replied to this. As far as I know, the WinPE (that comes on your OPK CD) environment should have generic drivers for your ethernet card.

That was my initial idea as well. The only reason why I considered T-39 was that I wanted to run the setup.exe *before* windows detected the card and installed any drivers for it. This is because, when invoked in regular mode, the setup program will recommend that you uninstall the current drivers and restart. Is there a way to prevent windows from installing drivers for a particular device?

A few questions: When you start an install with syspart and makelocalsource, does it leave the local source files on the drive after setup has completed? Where does it leave them? Does it set the registry values to point to the local source as the SourcePath, or does it still refer to the CD? After you run winnt32.exe from the PE environment and reboot (to start text setup from the HDD), do you need the CD anymore? Or can it be removed from the cd rom drive?

Wait wait wait wait wait wait! Wait........ Wait. How did you load the cd as a ramdisk? Can you post your startnet.cmd?

What if it's always going to run on a machine with a sound max card? I have a standard retail copy of windows for installation on one machine. I also have the CD-Key in the answer file as well, so it's really only meant for my pc and my pc only. Are there any other concerns about installing at T-39? Filesystem wise? I was just worried if it was too early in the install stage to run that setup program at T-39, because I thought something might get overwritten later on. Which stage are regular applications usually installed at? T-12 with RunOnceEx? The reason why I'm not using RunOnceEx or GuiRunOnce is because I'd like to have the software installed before even the first login. Let me know if this is doable. Thanks

It might be wroth beraking the .reg script up into multiple files. I'd suggest using the REG ADD command line utility in a .cmd script, but you have so many registry changes it would be painful to rewrite them all.

That's what I originally did. It works, but there's a problem. Installing by specifying the .inf file works for installing the *drivers*. But I also want it to install the control panel that let's me access otherwise inaccessible options (mic filtering specifically). The only way that works is to run the installshield setup.exe provided with the driver pack. Is there a way to integrate this as well?

I would like to have this program silently executed at T-39, since this is before windows tries to install it's own drivers for devices. Therefore I was wondering if it's safe to execute it from DetachedPrograms with Arguments = "-s" (for silent install). The setup won't restart after it's done, so that's not a problem. The rest of the windows installation should proceed normally. Does this sound safe?

QWCTF CounterStrike RogueSpear Dark Age of Camelot (BEFORE ToA Expansion)

The start.cmd file you create yourself. It's only if you don't want to put commands directly in the cmdlines.txt file. You can specify .cmd files to run from cmdlines.txt. You can put start.cmd wherever you want. Most people put it in $OEM$\install on their install CD. Note that anything in the $OEM$ folder of the install CD will be copied to your root drive. So, when referencing this file from your cmdlines.txt file, it would look like [COMMANDS] .\install\start.cmd Since cmdlines.txt is parsed from the root directory, you can use a dot (.) in your path. You start.cmd would look something like below: Start "Title" /wait %systemdrive%\install\some_application\setup.exe /switch /anotherswitch You don't need to put anything in winnt.sif. People put a reference to start.cmd in [GuiRunOnce] of winnt.sif only if they're using GuiRunOnce which fires after the first login to windows. Cmdlines.txt is scanned for automatically by Setup. There is no need to tell windows to look for it. Cmdlines.txt is parsed at the T-12 phase of setup. Note, for your needs, you can probably just put the command directly in the cmdlines.txt file. You won't need start.cmd in that case. You WILL however need to have OEMPreinstall set to YES. This tells Setup to copy the contents of the $OEM$ folder from your cd drive to the root of the drive windows is being installed on (commonly referred to as %systemdrive% NOT TO BE CONFUSED WITH %systemroot% which is actually %systemdrive%\windows). Also, use REGEDT32.EXE instead of regedit: [COMMANDS] "REGEDT32.EXE /S regtweaks.reg" From http://unattended.msfn.org/unattended.xp/view/web/30/ Also check out http://unattended.msfn.org/unattended.xp/view/web/14/

soundmax recommends that you use the setup program so that their control panel is installed... this allows you to set otherwise inaccessible options. their setup.exe has a silent install switch (setup.exe -s)... my question, however, is can this be integrated somehow as a software install? where do you normally put installer programs that have a silent option? can they go in with the hotfixes/updates, or does it have to go in guirunonce (i'd rather it not go there)?

So how is the new system working now that you got your Conroe? 3.6GHz? WOW... you must be at about 1600MHz FSB. Your RAM can take a FSB of 2132 if it is in dual channel. Just gotta push your mobo's clock from 400MHz to 533... it's only another 133MHz. must.... go..... faster.....

When you change all of your hardware like that, it can put windows into "shock". I had a similar problem. First thing: make sure you have the latest BIOS for your mobo. Question: Are you using SATA RAID, or just plain old SATA? If you are not using RAID, you don't need to hit F6. WinxpSP2 should detect the drive. You may have to check your BIOS to make sure that SATA is enabled. Some boards need you to enable both SATA and RAID, even if you aren't using RAID. Make sure you upgrade the BIOS first. If you are using RAID, check out this thread on using nLite to slipstream the latest service pack / hotfixes and burn a new cd. It allows you to install drivers ahead of time: I highly suggest you use nLite to create a custom CD. Within nLite (I think it's the Unattended tab). There is an option to have Setup suggest a Repair install. Make sure that is enabled when you burn your cd. Boot off your new CD. When it asks to press R to try to Repair the system with the Automatic System Recovery Console, say NO. On the next page it should find your current windows install and ask you if you want to Repair it. Choose [R]epair at this screen. Doing a repair install is your best bet. If it still doesn't work (and it didn't for me). Ya gotta reformat. Use BartPE to boot from and setup a network share so you can backup all of your stuff to another computer (or partition if you have multiple drives/partitions). From what I gather, most RAID driver floppies won't work at the F6 screen, because they are not WHQL certified (GG Microsoft).

Here's the part that confuses me: At 1066MHz FSB, your transfer rate will be 8.528MB/s (1066 instructions per second x 64 bits per instruction x 8 bits per byte) DDR2-533 has a transfer rate of only half of that: 4.264MB/s. The core speeds are matched (266MHz). But the RAM will still be the bottleneck. This is because intel chips are quad pumped, sending 4 instructions in one cycle (hence the effective 1066 FSB). If one were to get DDR2-1066RAM, the core clock speed of the RAM bus would be 533MHz which is faster than the mobo's 266MHz, but the data transfer rates would be equal. Would the core speed of the ram bus slow down to 266MHz? Or will it remain at 533MHz? If it remained at 533MHz, it would send two signals every cycle. Once at the high, once at the low. Does the mobo still fetch at the quad pumped rate? Once at the high, once at the first zero, once at the low, once at the second zero. If so, will the instructions at the high and low of the twice as fast 533MHz line up with the instructions at the high, 0, low, 0 of the slower 266MHz? That would maintain the 8.528MB/s transfer rate. Is this even electronically possible? It's a little rusty in my EE courses, but I thought the frequency has to be the same throughout the entire circuit. However if this is possible, I suppose it would look like: 533 MHz Hi Lo Hi Lo Hi Lo etc. 1066MHz Hi 0 Lo 0 Hi 0 etc. I suppose this is why new chipsets for Intel CPUs support dual channel. DDR2-533 RAM in a dual channel configuration would run at twice the transfer rate. In that case, but the core clocks would be matched, as well as the transfer rates. Discuss.

So here's the deal. I was completely clueless about the new DDR2 RAM, so I decided to educate myself. I read some articles about picking the right ram speed etc. And here's what I gathered. Apparently DDR2 RAM uses the same speed chips, but just double pumps them at the cost of latency time. The clock speed on the RAM stick's bus is twice that of the chip. DDR2-800 RAM uses 200MHz chips. They are double pumped to 400MHz on the RAM bus. At double data rate, this is an effective speed of 800. However, when an Intel CPU has a Front Side Bus of 1066, that is really only a 266MHz signal on the mobo bus itself (the "system clock" speed, or "core clock"). Since the core clock of the mobo is only 266MHz, the bus speed of the RAM must slow down to 266MHz at the RAM bus to match the mobo's clock. So your DDR2-800 is only running at an effective speed of 533. In this case, DDR2-533 would be the optimal choice. The only reason for getting DDR2-800 would be if you are overclocking your core clock speed of your mobo to 400MHz, to match the 400MHz bus speed of the RAM. Am I understanding this correctly? Or is there more to it? I read another article that didn't talk about matching core speeds, it just discussed matching the bandwidth (Gb/s) of the FSB with the bandwidth of the RAM. So which is it? Synchronize the core signals, or match the throughputs?

No pics, there's nothing pretty about it as it's just a closed case. Here are the specs: ASUS P5N32-E SLI SoundMax HD sound card onboard Supports DTS, Array Mic, 5.1, etc. Dual Gigabit Ethernet adapters. Nifty little heat pipes. and all the other stuff the board has. Core2 Duo 6600 (2.4GHz) 2x 1GB sticks of Kingston DDR2-533 in dual channel mode. WD Raptor SATA drive. 150GB, 10,000rpm. SATA DVD Burner/reader combo drive. eVGA GeForce 8800 GTS (640MB DDR3 RAM. Not sure of the GPU specs). This was a huge upgrade from my very old 6+ year old AMD system (A7V266 with a 1.4GHz Athlon XP first generation). I went from playing new games at 20FPS with a low of 5-10FPS during heavy chaos, to 60+FPS with a low of 40FPS during heavy chaos. I don't care about overclocking... I'm just stoked I can play games with eyecandy AND performance! It was a long overdue upgrade.