cov3rt

i was wondering, is basic electrical tape, such as the one listed below, will it be safe and provide the insulation for gpu chipsets / cpu or any other chipset that has exposed capacitors / resistors on the package so that it can protect it against any conductive thermal paste that may get smeared onto the area by accident? i noticed some electrical tapes mention up to 105 degrees celsius for operating temperatures, but some only list up to 80 degrees. i'm not sure if this is true, but doesn't the package area ( not the die ) run lower in temperature than the die? if so, i don't think they would get higher than 80 degrees celsius, but obviously, the 105 degrees celsius one is safer. https://www.amazon.com/gp/offer-listing/B001EM4EHI/ref=dp_olp_new_mbc?ie=UTF8&condition=new also, here is a picture below of the types of chipsets i'd use the electrical tape on ( as you can see, there is exposed metal on the package ). https://i.stack.imgur.com/DaBq5.jpg

when you say the modified setupreg.hiv also enables biglba in the txt setup part, does that mean that it can allow manually creating partitions larger than 137 GB in setup, or does it just mean that after installing the operating system, the partition then shows up as more than 137 GB if the drive itself is larger than what the setup wizard can create or detect? because if it's only for option 2, then the manual method which i mentioned in "https://web.archive.org/web/20110904131427/http://www.48bitlba.com/enablebiglba.htm" would be more straightforward imo, at least, i haven't been able to find clear and straightforward step by step instructions on how to implement the setupreg.hiv method. although there are some unanswered stuff / unclear things, for now, i'm just gonna go with the manual method for allowing proper operation of lba 48 on windows 2000 sp4 after installation, as i did come across several threads many many years back where the simple manual method allowed their drives to be detected properly, i think the oldest reference was someone using a 2002 amd motherboard and was able to have their 300 GB hard drive have it's full capacity shown after using the manual method of enablebiglba.

so i've tried to research sources of people trying to get lba 48 working properly on windows 2000, specifically when trying to install with sp4, and i haven't been able to find out what exactly needs to be done and for what purpose. i do not wan't to include or integrated / slipstream unnecessary or excess things, i already have to test blackwingcat's unofficial acpi fix which is part of some package that a user here referenced to, but it's unclear whether this would fix that issue and what other things would need to be done afterwards, as it's something i see have to see after fully updating my system, since not everyone updates their system in the same exact way, i have to take this into account, which is why i mentioned the issue with integrating / slipstreaming stuff may cause conflicting problems and / or may cause me to have to find a workaround that may not be feasible. as for the lba 48 issue, some questions involved are getting the setup to detect and create a partition higher than 137 GB ( if possible ), and / or whether the partition can be expanded AFTER installing, and if so, how would this be done. OR be limited to 137 GB partition and call it a day BUT at least have proper working lba 48 for data protection, so that even if the real capacity of the HDD is higher than 137 GB, lets say 160 GB, it still wouldn't have any chance of data corruption, because a specified patch / setting would allow for lba 48 to work "properly" except with limiting the amount of usage space. really, i'm more concerned about a functional system, rather than having a higher capacity drive. one solution that was mentioned which is unclear whether it's the only thing that would require for proper lba 48 functioning is simply enabling / setting a registry key in the registry AFTER installing windows 2000 sp4, but what exactly this would do, i haven't gotten a clear understanding of. the link below has a dowload link for the program that checks whether or not you would need to set the key, and if it say's failed, you would click set enablebiglba, but other than this, what else would need to be done, is it as simple as that? manual method after installing windows 2000 sp4 - https://web.archive.org/web/20110904131427/http://www.48bitlba.com/enablebiglba.htm automatic method after installing windows 2000 sp4 - https://web.archive.org/web/20110903033314/http://www.48bitlba.com/enablebiglbatool.htm there is also another thread which indicates loading of some syssetup file or slipstreaming it perhaps? and another method which indicates the usp 5.1 having the fix, which then would allow for greater than 137 GB formatting during setup? however, this of course would conflict with what i mentioned earlier of using excess / unecessary packages. it seems like adding one file such as the sysetup part would be sufficient, but i haven't found an exact step by step source of how to do so and whether that would be solution for everything, so that the enablebiglba option wouldn't need to be done either? here is what i already am including or include in my windows 2000 sp4 installations: -slipstreamed blackwingcat's amd and intel drivers -extended core v16a package ( yet to be tested, but hopefully it doesn't require any changes AFTER os install, and if it does, very minor ), i only came across like two files that are different towards my fully updated systems, and that is the acpi and videoprt.sys i believe, which are modified and / or different after installing one of the other types of unofficial update rollup packages, ( i downloaded from sdfox's files, or whatever that guy's name is ).

ok so i have several questions related to software functionality and usage, specifically for use on windows 95 mainly. 1. are there any benefits that vlc player 0.8.6d player has over roytam's "TCPMP 0.72RC1 Mod 6" media player, or the other way around, what can roytam's media player do that vlc player 0.8.6d can't? for example, does vlc player 0.8.6d also support youtube video streaming or does it not? 2. are there any issues with roytam's media player build? 3. what would explain illegal operation related errors when using vlc player 0.8.6d on windows 95 with the inability for files to load and play on certain computer builds ( which don't seem to occur on 98SE )? 4. how exactly does one install the DivX codec packages and / or which version is the last compatible version for windows 95, as well as for 98SE? 5. does installing DivX codec package cause any problems in overall system functioning or conflict with files / software? for example, if someone were to install a divx codec package, as well as use vlc player 0.8.6d, would this cause problems with running the program itself or running files from it? 6. does installing DivX codecs help the system in any way directly or indirectly? for example, if installing Divx codec and running vlc player 0.8.6d, it would allow for slightly better support and functioning of files and videos? 7. is it necessary to install DivX codecs at all? if so, for what reason? does one need to install it when using vlc player or roytam's media player? 8. is flash player 7.0, specifically "flashplayer7r73_win", used in any way by vlc player or roytam's build? would it better to install it or would it cause more issues if installing it?

i've never tested amn refrigerator ( any version ) on 9x, but there are at least two other cpu usage viewing software you can test. i currently bookmarked the pages for now. as for cpu temperature monitoring, i have used hwmonitor 1.15 ( windows 98 version ) and / or speedfan 4.28 for 98SE. other choices you have for 98SE is DTemp 1.0.0.34 for checking storage device temperature only ( also works on windows 95 ). there is also Hmonitor 4.1.4.5 (October 2004) for possibly checking cpu temp or other system temps, tested working on both windows 95 and 98SE, then there is aida32 3.94.2 ( both windows 95 and 98SE tested working ) which is mostly for checking system information but might also show storage device temperature, then there is HDD Health (S.M.A.R.T. monitor) v. 2.1 beta, build 159 - for checking hdd temperatures and other info ( windows 95 and 98SE tested working ).

sorry, i got you confused with the other guys post saying they got camstudio 2.0 (not 2.7 ) working fine. as for whether or not it would work on windows me, i'd say most likely it would, as it mentions support for it, at least from a indirect source or two i found. but yeh, if you haven't already done so, you should give hypercam 2.13.01 and / or camstudio 2.0 a try on windows 9x / ME, i was hoping camstudio 2.0 would work "Better" as i actually removed hypercam 2.13.01 from my archived packages, since it has the watermark issue for one. the pentium m's shouldn't be too slow for camstudio 1.5 / 2.0 recording, so i was wondering if maybe the gpu and / or ram was the issue? do you remember which gpu's you tested and how much ram you had installed?

do you know what cpu and / or gpu / ram amount your using for camstudio 2.0? it would help to know, because i'm planning on removing hypercam 2.13.01 and replacing it with camstudio 2.0 for windows 95, 98SE, and NT4, because it seems that hypercam 2.13.01 ( being free, but unregistered ) shows a watermark for recorded videos, but camstudio 2.0 doesn't? although i have not confirmed if camstudio 2.0 does this or not, but hypercam 2.13.01 i believe does, as it mentioned it would show a watermark. though the focus on this thread was for windows 98, i'd like to add that hypercam 2.29 worldwide free version may or may not have the watermark issue ( it didn't give me the message after installing it ), and it should work on windows 2000 and newer, but probably will not install on non kernelex 98SE or older OS. i have only tested it on windows 7 briefly just to install it.

you can try hypercam 2.13.01, it works on windows 95 and should also work on windows 98. i noticed people mentioning camstudio 2.0 ( i have yet to test this on windows 9x ), however, one would think hypercam 2.13.01 may be better as it was released in May 20, 2006, where as camstudio 2.0 has a release history of April 6, 2004. one would have to look at change logs and see what exact differences are with each other and / or simply test them both. i don't like keeping two of the same programs usually, even if they are relatively small in size, unless there is something camstudio 2.0 can do better or functions better in some ways that hypercam 2.13.01 doesn't, then i'd rather just use hypercam 2.13.01 if it can do everything camstudio 2.0 can, however, after checking camstudio 2.0 really quick right now, i've decided to keep it too.

assuming the original thermal pad in the previous laptop i worked with was 1.5mm, using that reference and jaclaz's points, and factoring in thermal paste in between the chips and heatsink, it would roughly put it around 1.3mm or so which is within the specifications that jaclaz also mentioned, which likely would allow adequate functioning. the reason why i went with the copper shims vs thermal pads was because i couldn't find a good deal for any of thermal pads and / or felt that the padding in general would be less durable or more inferior than the shims. i got (20) 15mm x 15mm x 1.2mm copper shims from amazon for $8.54 shipped, though from amazon free prime trial offered the free 2-3 day shipping or else it would have been almost $13 with very slow shipping. the trial is only 1 week so i need to make sure to cancel it before that time ends.

anyone know the last official version of intel inf update for windows 2000? when i was checking my files, and even wayback machine when it was the last time it showed the older OS menus on intel's site, showing windows 2000, although i couldn't select it through waybackmachine, i was able to check what was already given for all OS by default, and it showed version 9.2.0.1030" as the newest i believe. however, just today i noticed blackwingcat mentioning in his blog of a version "9.3.0.1019" being official i believe, which i went and downloaded ( yet to test ), and now i also found even a newer package which mentions windows 2000 in the readme, it is version "9.4.0.1027". my problem isn't so much on whether they will work, it's more that i would like to know what exactly is the last official version or not. being that i would wan't to install blackwingcat's latest intel ahci / raid package via setup AFTER applying all previous intel inf updates to ensure proper updating of the system.

if i forget to mention, clearance issues makes thermal paste alone is not a option, as there would be a gap and would cause overheating. that's why i was asking what would be the better of the two, the copper shims + high quality thermal paste or thermal pad, and specifically factoring in for universal usage, thickness, etc.

i actually had a fully updated windows 2000 system with that update rollup v11 version i mentioned to check the acpi and videoprt.sys files and it shows the videoprt.sys as slightly newer at version "5.0.2195.6835", where as the one in the extended kernel v16a package was "5.0.2195.6833". now the question is whether the newer one i use is buggy. as for the acpi.sys version, the system with rollup v11 without extended kernel v16a shows as "5.0.2195.6920", where as the one in extended kernel v16a is "5.0.2195.6921". it's important to mention that the update rollup v11 i use gets installed late in the update process for the builds i work with as i have it specifically set up this way in the order of updates, tinkering with this may cause changes that are unfeasible to manage, since i have to find when or where exactly there are any breaks or errors and what to do about them which usually is caused by installing something too early or too late.

the instructions help, but the problem is, i don't even use the extended kernel v16a in my builds. i use the update rollup v11 with filename "Windows2000-UURollup-v11-d20141130-x86-ENU.7z". another issue is any complications that would occur by integrating these after os install, such as when updating the chipset through intel inf update or some amd / ati chipset package. and i was wondering if it's possible to just integrated only the acpi and videoprt.sys stuff as to not cause any other complications, i guess the same way or similar way to that xp reference i made earlier, or is that even more complicated or irrelevant?

as ragnargd mentioned, you have a q43 chipset which belongs to the ich10 family and if i'm not wrong, there should be working unofficial chipset drivers for them. i actually have the unofficial package but i never tested them myself, it includes ich10 in the inf, though requires a slipstream process. luckily at least you have a desktop motherboard, so assuming the chipset part gets taken care of, you should also have support for a decent video card through pci slot or vbemp basic display for the integrated gpu. i was gonna upload it here in case you wanted to test, but the package is about 45 KB zipped, and i can only upload max 10 KB. lonecrusader has a link for it to download though here "http://lonecrusader.x10host.com/files/INTELINF_9.4.0.1017_98SE.ZIP"

i was wondering if anyone knows if it's possible to substitute a thermal pad with a copper shim for gpu cooling if the thermal pad was used as the original part? since the shim would not compress, it would seem that i would need one considerably thinner than the thermal pad, but there are other issues. for one, a lot of manufacturers don't list the specs of the exact thickness of the thermal pads they use for the gpu or chipset, another thing that is unclear is whether or not i could use a copper shim if the heatsink already has copper surfacing for the gpu and / or chipset portions. what i mean by this is, would it be ok if i used thermal paste between gpu and chipset die, placed shims on top of them, and then thermal paste on top of the existing copper surface, or would this be an issue? the only experience i've had with using copper shims was on the dell latitude d630 where i replaced the thermal pads with shims for the gpu and chipset portions with thermal paste, and it ran a lot cooler, but the heatsink's surfaces weren't copper plated, just normal. the copper shims i used were 1.2mm in thickness. also, i was wondering if it's ok if the either the thermal padding or copper shims are larger than the die itself for the gpu or chipset? i mean, it would be impractical to have to buy separate sized shims or pads, just so they can fit only in the die portion. the reason why i would wan't to use the bigger ones is better universal compatibility and if possible, by using a shim or thermal pad that is a little bigger than the original ones, could possibly provide slightly better cooling on smaller dies, especially in the case of poorly designed systems that makes it harder to achieve decent temperatures. but to not forget, the 1.2mm shim may not work on let's say a system that originally used a 1mm thermal pad ( and i don't have a measure to check the thickness in millimetres ). i was wondering, with the thermal pads squishing, how much thinner does this make it? like if i were to buy a 1.5mm thermal pad, how much thinner can it get when bolted down by heatsink? because if i can get a more rough set measurement estimate, it may be possible to get a more "universal" copper shim with use of thermal paste, instead of thermal padding ( which i'd assume would be better than just using thermal pads ). so if originally the thermal pads were 1.5mm, and i buy the 1mm copper shims, then they should be fine, or would 1.2mm be a better "universal" size?

yeh, but what exactly do i do in detailed instructions with nlite and the acpi related stuff? the only thing i found through my own experience and understanding was a page listing a updated windows xp acpi.sys file, in which you would convert it to "acpi_sy_" file, then replace the old "acpi_sy_" file in the extracted directory ( to which you would use to inject into nlite ) with the new "acpi_sy_" file. so i was thinking i could simply replicate this but for a windows 2000 build and use that acpi.sys in that extended kernel v16a package you referenced to. but you also mentioned videoprt.sys, does it mean i should also convert that file to "videoprt_sy_", and replace the old one with it? you know, a lot of these things are confusing to me and it doesn't help that a lot of my questions often get unanswered, leaving me with no other choice but finding a way that works, but even if i manage to, i may not know why. as for the sata part, i was able to install with achi enabled in the laptop i was working with, i simply had the extracted windows 2000 sp4 operating system and integrated in nlite, then put both blackwingcat's amd sb600,700, and 800 chipset drivers in the multiple driver folder option after selecting the driver option, as well as the intel drivers, and made sure to select all and selected the textmode driver option, and also the bootable iso option from before. though i was browsing the forums and it appears that the driver i was using for amd was a little older than the newer one and may not be the most stable, so now i need to recompile with the newer ones and remove the older ones. but i also don't know what to do with the acpi related fixes and like i said, unanswered questions doesn't help and provide me with a practical and feasible solution. btw, below is the page for that windows xp updated acpi fix solution - https://retrosystemsrevival.blogspot.com/2019/10/updated-windows-xp-acpisys.html

i was not able to download the "acpi.cab" from that link, and waybackmachine didn't work either. however, i was able to download a package which apparently contains the patched acpi.sys from here "https://msfn.org/board/topic/181245-solved-problems-installing-windows-2000-sp4-on-a-thinkpad-t420/". however, i don't know what i need to do next. it seems i would need to integrate specifically the "acpi.sys" file or multiple files?..into nlite again to make a updated iso, but i was wondering if i could get the specific steps to do that and / or which exact files i need to be putting in the nlite integration. also, i was wondering if this will require me to do anything after installing the operating system, like for example, will any chipset inf updates affect the acpi.sys fix and require me to re update it?. i already asked if whether or not i need to use the setup of blackwingcat's intel package for the sata / ahci controllers AFTER all other inf update packages, but i havent got a response for that yet. also, for the amd builds, blackwingcat's package doesn't have any other stuff in it, other than the infs themselves / textmode file. it seems like the update process on amd systems may be more straightforward? however, from my experience, i've always had more trouble with getting systems to work properly with amd / ati chipsets, whether it was on windows 95, 98SE, or 2000. their chipset updating process is a bit unclear, like it always seems like there is some kind of chipset update i'm missing. i think it's also possible that i couldn't even get the gpu driver to load onto systems that specifically mentioned as having support for windows 2000 for that particular laptop, yet despite installing all missing drivers, etc, the gpu driver would not work or would cause a black screen.

so i went and made a nlite splipstreamed windows 2000 sp4 installation disc ( including both the amd package in the multiple folder option, as well as all of the intel drivers for the intel package ), and choose the textmode option. i put it in a 2011 or 2012 year laptop ( uses a amd a4 apu ) with sata interface to test, and i got to a point where it got stuck on the "setup is starting windows 2000". so i was wondering what could be the issue and solution in fixing this problem. i did a google search and the only two suggestions i got was to make sure EHCI is enabled in the BIOS and / or install with ACPI disabled by pressing f5 or f7 early in setup phase. so the laptop i was testing on did not have a EHCI option in the bios, and i didn't choose the f5 or f7 option yet. also, it's another person's laptop so i don't wan't to risk messing it up, and from what i know on ACPI, i don't think disabling it would be a viable option as some things may end up not being detected ( i only would do this mainly on windows 95 systems, and to a lesser extent, if at all on 98SE ). i also tested by pressing f6 one time and without pressing f6 from early in setup, but it still would hang for a while on the "Setup is starting windows 2000" portion, and so i just would shut it off. the only thing i can think of so far of why i get the hanging problem with the laptop i tested was simply that the sata controllers / chipset is not supported and is causing issues and can't be run without some other unofficial patches / integrated drivers, etc, and that there isn't anything wrong with the nlite part, or is there?

it's all fine and dandy if i can get through the setup phase with the slipstreamed drivers, but what about chipset drivers once you installed the operating system. what i mean by this is for example, if you run the intel inf update for let's say an ich8-m chipset, will it override or have a problematic effect on the blackwingcat's sata driver or would it not get affected and i won't need to do anything additional? i noticed that blackwingcat's modified intel sata driver package has a setup installer. it seems logical to run this after i installed all previous chipset inf updates so that all controllers or other devices are updated or is this not required? however, the amd package only contains the regular inf, .cat file, .sys and txtsetup.oem. amd chipset updates have usually been a little different than intel chipset inf updates. for the particular laptop i'm experimenting with, it only has a TPM driver listed on the official driver page under the chipset category. also, should i be setting the option in bios to AHCI if there is an option to do so, and if not, are there any issues if i don't, i understand that RAID is also another option listed in the readme of the intel packages at least, but i'm unlikely to ever select that particular option.

i had some odd problems with NT 4.0 and trying to get certain programs to install / run ( one reason why i lost interest in working with NT 4.0, other than being more complicated work with in general ). i tried asking about them on the forums, but no one was able to help with a solution, and google, we'll i tried but couldn't find a solution. specifically when trying to install odyssey client 4.56 ( i think ) or a version of wsc guard, and it would complain about not having permissions to install or something. i also remember having the same or similar problem when trying to install some amd / ati catalyst package that had a NT 4.0 driver, but wouldn't wan't to install due to having some kind of restriction / permission problem, regardless of whether i tried manually installing the driver or not. it sucks, because it doesn't seem like the programs themselves were unsupported, but some setting or problem in NT 4.0 wouldn't allow them to fully install, etc. i was never able to find a solution for any of this. i don't know if NT 4.0 supports any WPA2 wireless adapters. maybe if we were to assume middle ground, then wsc guard 4.0 with wpa / aes support would be ok, being able to connect if the router supports WPA and / or WPA2? if anyone knows exactly how this works, please correct me if i'm wrong on any of this. i do remember getting wifi working on NT 4.0 for i think the dell latitude d600 which seems to have near or full support for NT 4.0, but i don't remember exactly which wifi card i was using and / or wireless settings. it might have been some atheros card?

thanks for the info, it seems that nlite 1.4.9.3 would be good to start then? also, i was wondering, have you ever tested the uniATA textmode driver? also, can i combine all of them in the same slipstream or will there be conflicts? for uniATA, i'll wan't to use version 0.46.4.8, as it seems to be the last "Stable" version indicated from the site, though as an additional question, there are debug versions. would there be a problem using the debug versions as in affecting performance, and what benefits would they have? i posted a pic below of the driver package i have of uniATA btw, i'm assuming i need the two boxed files only?

it's been a little while since i last tried experimenting with trying to load / slipstream sata drivers or use specifically the nlite program. i've never been able to get it work and instrunctions elsewhere were always confusing. today i did some research again, for one, it's mentioned you need two files, the .sys file of the driver and the txtsetup.oem to be slipstreamed. i have searched the internet and found three sources in particular, blackwingcat's sata / ahci driver which works up to ich10 i think, and the amd version which supports sb600, sb700 and sb800 chipsets, and also the uniATA driver which seems to support all controllers, amd, intel, nforce, via, etc. now the problem is, i obviously do not know which exact steps / instructions to follow on how exactly to burn a slipstreamed windows 2000 sp4 cd with the included slipstreamed drivers. i currently have one page to follow which the name of the site article is "2 Ways to Integrate Floppy SATA RAID Driver into Windows XP CD". another issue is that there are all these different nLite versions. i know some support windows 2000, but is there a particular version that's "better" for burning ISOs for windows 2000 sp4 installations? for example, there was some post on the below thread here "https://www.vogons.org/viewtopic.php?t=27006", saying how later versions of nlite may break things, but there wasn't any more detailed explanation of what. this was in late 2010. so let's see i download nlite 1.01 final and wan't to integrate the uniATA driver files and / or intel / amd ones. will this work fine or would i need a newer version of nlite? also, would there be a conflict if i put all three the uniATA and intel / amd in the same slipstreamed cd? my goal is hopefully to create a custom windows 2000 sp4 installation disc with functional support for all sata controllers so that i can install windows 2000 on any laptop that has a sata controller only and can't revert to ide / legacy compatibility mode. i'm not worried too much on speed and performance, i mean ata 133 is still pretty fast, and random read performance is really more important, and if you have an ssd, even if limited to ata 133, you still will have a relatively fast system. even the most recent ssd's with the fastest random read 4k queue depth 1 such as the samsung 860 pro can only manage maybe around 40-45 MB/s.

i was wondering if anyone has any experience with the lenovo e440 and l440 laptops. i plan on using one or the other. i am concerned on the laptops feeding enough usable video memory from the integrated intel hd 4600 graphics. i noticed something mentioned for the e440 model, i am not sure if it also is said for the l440, but i found it for the e440 and it mentions "Graphic Intel® HD Graphics 4600 / NONGFF Expandable". does anyone know what NONGFF expandable means? i tried to google this and could not find any relevant things or abbreviations for it. i was worried that maybe it means that the laptop only comes with a fixed amount of video memory and can't be changed, regardless of how much system ram is installed, because often the case, if you were to install more system memory, the amount of usable video memory would go up, and a bad scenario would be that it would only have a max of something like 128 MB usable, regardless of application, which isn't enough for the applications i need it for, i would need 512 MB. now i have had experience with the t440p, i believe it allowed setting up to 512 MB of shared video memory in the bios. i no longer would use t440p models since they come with the gt 730m that just flat out sucked for me in performance compared to the intel hd 4600, which i found odd because according to passmark's score, it should do better, but it performed worse for the game i used it for, now for privacy reasons, i didn't want to mention the game. i also didn't like having the added gpu in the system, creating more heat and power usage, so that's why i am avoiding t440p's, and of course for the other reasons i mentioned. it simply doesn't make sense to have two gpus almost identical in performance, although ironically, this wasn't even the case with me, it did a lot more poorly for the gt 730m. t440 isn't an option because it doesn't have a optical drive ( not that important but i prefer one ), but the bigger issue is that the case design is in such a way that makes it very difficult and impractical to open up to replace the cpu thermal paste and do other stuff / cleaning, which is why im looking at the e440 / l440. i would be using 8 GB of ram for either the l440 or e440 if i get them. i understand that manually allocating video memory in bios may not be necessary in some laptops or systems, in that the gpu will use more video memory if possible, but of course, some manufacturers limit this ability and if there isn't a bios setting for it specifically, then it may also not give enough video memory when needed. also, here's the source for the site to which i got the info of someone else posting the information of the e440 -

i did test one of the palemoon versions with and without extended kernel a few days ago, i think it was the 26.5 version of palemoon, my problem was that there would be a black screen upon loading. i found a source mentioning to disable hardware acceleration, i did that for the gpu and no more black screen and i was able to load the browser fine, youtube also worked still at 360p or auto resolution, though the browser was a bit unstable / and did not work for everything / load everything, for example, hotmail would not completely load. it seems that this black screen issue might also be the same problem i had with kmeleon 74 on windows 98 and that if you complete disable hardware acceleration on the gpu, then the black screen would go away but i never went back to test this yet, though 98SE is a different story, i don't want to derail the topic from windows 2000 specific stuff. i can't comment on the specific script issue you mentioned. all i can say is that i didn't really have much of any issues with the latest kmeleon 76 version with extended kernel, although i did find a scripting error where the web browser would just freeze. i am not sure if any of incompatibilities / bugs have to do with the web browser more or maybe using sse2 cpu's on newer modern software / web browsers being the problem and not being able to handle stuff. for example, sometimes the cursor or selection point when i am typing such as now may jump randomly to another area of this post or at worse may completely go backwards for this entire page. honestly, i'd avoid using anything less than sse3 cpu, at least for web browsing. also back the hardware acceleration part, you may not need to completely disable, try going in increments as high or low to see which setting seems to work best / to your liking.

like i said, i'm not sure which version is / was compatible or gave that error message related to mfc42u.dll is missing. it's possible, that because the installer wouldn't properly install on windows 95, then it would also copy the wrong files or ask for the wrong files such as mfc42u.dll which i believe is only an NT file. but all this really doesn't matter anymore as i moved on from the experimenting. i did use version 1.1 for the installer on 98SE with unofficial service pack 3.64 without kernelex and it worked fine with wpa2 connection, i didn't bother with the newer versions. i am pretty sure there are no existing official wpa2 drivers / software for windows 95, unless i can see for myself, i'm not even sure on wpa.