Sfor

The simplest solution is to activate Windows 98 boot menu, instead of adding another option to the Windows 2000 menu. In the effect the Windows 98 boot menu would show after selecting it from Windows 2000 boot menu. In the MSDOS.SYS in the [OPTIONS] section there is a BootMenu=1 switch. I t should force the Windows 98 boot menu to show. Also it is possible to increase boot delay, when Windows 98 waits for F8 key. The standard setting is BootDelay=2. So hitting the F8 key in just 2 seconds time is very difficult. Also it is possible to create a Windows 98 boot menu through CONFIG.SYS and AUTOEXEC.BAT files.

The USB 2.0 is backward compatible with older USB standards. In windows 98 the USB 2.0 controller does not always work correctly. To be on the safe side it is possible to disable the USB 2.0 controller in the device manager. The remaining USB 1.x controllers should do the job, forcing the USB 2.0 devices to work slower.

I do suspect the mentioned device does not have a proper driver installed. Windows calls a "PCI Network Controller" every PCI network adapter it does not have drivers to. "Plug and play" means a plug and play and not USB. Since the name refers to the PCI bus, it is certain it has nothing to do with USB bus. I think you mistook the device conflict with unknown device without a proper driver installed.

I just do not get it. The MaxPhysPage and MaxFileCache should not affect a pure DOS application, when GUI is not runnig. As for the Windows 98 boot menu. Instalation of Windows 2000 does not mean there will be no windows 98 boot menu. In normal circumstances the Windows 98 boot menu appears after Windows 2000 boot menu. It just could be a bit difficult to hit F8 when necesary. In my case I do have Windows 98 menu forced through config.sys. But there is BootMenu=1 setting possibility in the msdos.sys, as well.

If the MaxFileCache is too big, the DOS related applications stop running. Yo should try to decrease the MaxFileCache. In my case on 1GB RAM the MaxFileCache setting above 800000 prevents DOS applications from running.

I'm thinking about testing older than 6.2 Ati Catalyst driver sets. But, I have serious problems with getting these. The Ati (AMD) pages do not offer such old driver sets, not to mention there is hardly anything about Windows 98. I can not get to the Windows 98 driver sets on Filehippo, as well. I found just XP or Vista related packages there.

The general speed difference between the two systems is not the topic here. I did research the problem, and the HDD performance on other computers is not particulary different between the systems in question. In some cases Windows 98 HDD performance appeared to be faster. Also I do not work with Windows 98 FE, at all. So, I'll be referring just to Windows 98 SE. As for the network speed difference, Windows 98 adds a significant lag to TCP/IP stack. The max connection speed suffers from that lag. So, in general Windows 2000 can get higher TCP/IP transfers, than Windows 98 does. On the other hand in certain circumstances Windows 98 seems to work significantly faster with NetBeui than Windows 2000 does. According to my experiences Windows 98 is faster with high quality video decoding. Too bad it can not handle newer codecs, thou.

When it comes to getting USB flash drives working in Safe Mode, it is possible to use DOS drivers, for such a task. But, it is a road full of thorns, as it is relatively difficult to get DOS level drivers to work correctly. The easiest thing would be to enable BIOS USB flash drive support, if available.

I'm using various DOS tools, quite fequently. According to my experience the AGP based ATI cards do have a tendency to hang when a full screen mode DOS application is launched, or switched to full screen mode. On the other hand the Nvidia AGP cards were much more reliable in this kind of tasks. So, when looking for a PCI-E Windows 98 compatible graphic, I started from Nvidia. The GeForce 6600 do seem to work reasonably well, but when it comes to DOS applications they gave me a headache. On the other hand ATI Radeon x600 seems to be working mych better than GeForce 6600, but it shows exactly the same symptoms, as older AGP ATI cards. Currently I'm trying to get a HP DC7100 i915 equipped with PCI-E DVI capable graphic card.

I do not think so. The HP DC7100 Intel 915 + 82801FB uses exactly the same BIOS naming convention. With both computer models in both Windows 98 and 2000 the "combined" mode presents OS with just one dual ATA controller with one PATA and one SATA. Both subcontrollers do show master and slave drives. It disables the third from the integrated controllers, as well. On the other hand, the "Separate" settings prevents Windows 98 from booting, while Windows 2000 detects additional controller. My observations are showing the same Windows 98 instalation ported to DC7100 works significantly faster with some disk related operations. In D530 i865 P4 3GHz ICH5 moving about 1GB of data by a DOS based application works several times slower, than on DC7100 i915 P4 2.8GHz ICH6 with a much slower HDD. It could be related to a fact the ICH5 82801EB is the first from the SATA capable 82801 series, so it could have some bugs in the BIOS "IDE compatibility mode" or driver for Windows 98.

I decided to leave Silicon Image controllers behind, when I started to play with cheap second hands HP computers with integrated Intel SATA controllers. Also, I had an opportunity to test two dual boot HP DC7100 with Intel 82801FB chip. I found no significant difference with HDD performane (W98 vs. W2000) in case of those two. But, both of them had just 40GB old SATA drives. So, the difference in tranfer speed could be bottlenecked by the drives poor performance. In any case, I'm planning to make some more detailed testing by moving the two huge Seagate dives to a HP DC7100. I should get a definitive answer, then. For the moment, the issue seems to be related just to the 82801EB chip or just to the HP D530.

I'm sorry, but you asked two different questions. So, I'm not entirely shure to what question the presumed answer applies to. In any case, both operating systems are working with exactly the same "Combined" BIOS setting. To be more precise the HP Compaq BIOS does not have "IDE emulation mode". However, the "Combined" setting does seem to be "IDE emulation mode", as opposed to "Separate" (probably AHCI).

Does your 250GB drive have a SATA interface? I do suspect the performance issue is somehow related to the SATA. As for my drivers versions: Intel® 82801EB Ultra ATA Storage Controllers CONFIGMG.vxd 4.10.2225 Primary IDE controller (dual fifo) ESDI_506.pdr 4.10.2230 IOS.vxd 4.10.2225

HP D530 with Intel 865 chipset and 82801EB controller. Two SATA Seagate 1.5TB and 2TB drives. HDD controller is set in BIOS to compatibility mode with multisector and UDMA enabled for both HDD drives. Dual boot Windows 98SE and Windows 2000 professional. File copy operation between drives (5.86GB): - Windows 2000 - 2m 13s - Windows 98 - 3m 15s The Windows 98 does have DMA transfers enabled for both HDD drives. I'm suspecting the W98 does not take advantage of the higher than UDMA 33 modes. I'm looking for the way to check what particular UDMA mode Windows 98 is using. Also, It would be good to force Windows 98 to perform as well as the Windows 2000 does.

According to my observations many Nvidia and Ati cards do have problems with DOS applications. Sometimes newer driver versions are more stable, but sometimes not. In case of AGP, the Nvidia seems to be more reliable, when Ati seems to be better with PCI-E. When it comes to plain old PCI cards with S3 chip were very reliable. I forgot most of my results with plain PCI, anyway.

I was lucky to observe a perfect system shutdown, when the ethernet cable was unplugged. Following through, I did confirm, the system shuts down correctly, when DHCP lease is released, just before the shutdown or reboot. The next step was to separate the Microsoft Networking client and sharing from TCP/IP, while leaving just the NetBeui for file sharing. TCP/IP stack is still available for Internet access, while file sharing is done just in the LAN through NetBeui. The system seems to power down and reboot correctly, now. Strange, but effective.

I do have a HP DC7100 convertible minitower computer. It does have i915 ICH6 chipset with i82801FB IO controller. The board does have 4 SATA and 1 IDE connector. The BIOS allows to set "combined" mode, in which the OS is presented with a standard Primary/Secondary dual IDE controller architecture. The Windows 98 seems to be working fine in such a situation, but it has an access to just 2 SATA channels and the IDE controller. The remaining 2 SATA channels are disabled. Since the HP/Compaq BIOS naming is quite different from other releases, I can only suspect the "combined" is equal to "compatibility" mode. The other from the "combined" BIOS setting adds a separate SATA controller. It appears, the Windows 98 fails to boot in such a case. It could be possible to go around the problem, by using DOS compatibility mode disk access, but it does not seem to be an efficient way. So, I would like to know, why Windows 98 behaves as it does. Also, is there a way of getting the full potential of the 4 SATA channels?

The GeForce 6800 is the oldest from the 6000 series. The next one was 6600 then the 6200 was released. As far as I remember, the 6200 support came in the Forceware 7x. There was no support in the Forceware 6x for GeForce 6200.

I have an update, as I was able to test a Gigabyte GV-NX66128DP (PCI-E GeForce 6600) with HP DC7100 (Intel 915). There is support for the GeForce 6800 series in Nvidia 61.76 driver package, however there is nothing about the 6600 series. After testing 77.72, 71.84 and 81.98 driver packages, it appears the system restart function works correctly in all of them. System shutdown works good except for 81.98. The computer ended with "it is safe to turn computer off now" like message, so I had to turn the power off by myself. In all three driver packages I did experience problems with running DOS applications. There was a dialog informing about an "illegal operation performed by the application" or complete loss of video function in full screen mode. The workaround is to change the screen color depth at least once after booting the system. The DOS applications are seeming to work correctly, after that.

Instead of a WDM hardware driver repackager / installer it could be easier to build a repository of already repacked drivers. However, there could be some legal issues with distribution of modified drivers.

Well. It appears there is yet another possibility to obtain drivers. WdmStub thread gives hope to use WDM 2k/XP drivers in W98. I looked in to the audio drivers for Asus EEE PC 900, and it was too much for me, sadly. The driver package consists of the Realtek drivers and Microsoft patch for Windows 2k/XP. The patch is Microsoft UAA Bus Driver for HD Audio. Apparently the entry in the device manager (PCI\VEN_8086&DEV_2668) is the Microsoft UAA Bus. The Realtek Audio drivers are to be installed after the bus driver. I have too little experience with WdmStub to try to port such a complicated driver package.

According to my research the FFDShow tryouts developers do not take active part in the Hi10P decoding development. Their intention was to include the Hi10P code, after it proves to be stable enough. So, it looks like the push to port the newest alpha FFDShow tryouts to Windows 98 is not a good course of action. The core of the Hi10P development seems to be centered around CCCP codec pack. However, the best results are reported from LAV Filters project.

As I have mentioned elsewhere some time ago, the MSVC2010 clsid builds from sourceforge can be made to work by replacing the kernel32 imported function HeapQueryInformation by SetLastError in ffdshow.ax. I am currently using the thus treated rev4257_20120115_clsid build without apparent problems. I haven't got a 10bit x264 file to test though but I don't see a reason why it shouldn't work if decoding of such streams has been implemented in those builds. I found the thread you mentioned: http://www.msfn.org/board/topic/153659-my-little-hex-hacks-to-complement-kernelex/ I tried to install ffdshow_rev4305_20120205_clsid.exe, but it does not seem to be working. Where should I put the patched MSVC2010 libraries? The decoder config dialogs are not working. The GOM PLayer does not find the FFDShow filters. The links to the decoder config are incorrect: C:\WINDOWS\SYSTEM\RUNDLL32.EXE"C:\WINDOWS\SYSTEM\ff_vfw.dll",configureVFW. I changed the link to: C:\WINDOWS\RUNDLL32.EXE "C:\WINDOWS\SYSTEM\ff_vfw.dll" ,configureVFW , but it did not help much. ---------------------------- I did the same with ffdshow_rev4257_20120115_clsid.exe. But, no effect, so far.

Well a few things happened. I did get a 3.2GHz/1Mb/800Mhz CPU. It works better, but it still freezes with some full HD videos. I had to switch to a different motherboard with Intel 915 chipset and Socket 775, in the process. I figured out how to time shift subtitles from the FFDShow directly. The trick was too simple, so I did not got it the first time. The keyboard shortuct are ment to be used with both Ctrl and Alt keys pressed. The 10bit x264 codec is more and more common. The output of such a video stream can not be treated as high quality, any more. The last stable FFDShow tryouts Beta7 does not solve the problem. Have anyone solved the problem of the 10bit codec on Windows 98? The newest FFDShow alpha releases are using MSVC2010 compiler. Will it work with the KernelEX? Have anyone been able to get it working?

I'll add some more informations. The HP d530 both Small Form Factor and Convertible Minitower versions do have the same motherboard. The Ultra-slim Desktop version does have a different one. The all three HP DC7100 versions are equipped with different motherboards, as far as I know. I'm using just the Convertible Minitower d530 ad DC7100. A side note: There seem to be not much sense in using standby function to shut down the system. I saw no possibility of using the Hibernate function in Windows 98. However it works fine in Windows 2000. The mentioned WOL functios are working great on both d530 and DC7100, if the BIOS settings are set correctly. It has nothing to do with the OS running, as far as I know. Also the WOL works correctly with capable PCI ethernet adapters (not just the integrated ethernet port) on d530. I had no opportunity to test everything on DC7100, so far.