gerwin

Back to check. No, there is no case fan or any other fan blowing at the card. There is only the CPU fan (stock intel) and PSU fan.

It is a passive GV-N710D5SL-2GL. Details in post no.3 of this topic. Big heatsink. The Lian-Li case can actually take a full size ATX PSU. But I don't like how that blocks airflow, so I offset the ATX PSU to protrude 50mm at the backside. I may have added a small case fan near the GPU. I am not sure, and cannot check now...

That is the case I am using, yes. Graphics card goes at the bottom. Heatsink pointing down. Back then, when I assembled it, I checked the temperature by touch. It was within reason. 90 degrees as you mentioned, would not be acceptable for a build of mine. I don't know why you are getting these high temperatures.

An image of the Lian-li mini-ITX case is om page 2 of this topic. Don’t remember whether or not I added a small fan. Not home to check now…

I only have time for a brief test this week. I get 53 degrees C. After 20 minutes on Desktop with GPU-Z render test running that whole time. It is a fanless GT 710 in a small case. Light duty GPU temperature was 25 degrees C. See Image. Edit for completeness: OS Windows XP SP3 x86. Driver 344.75. The card is a Gigabyte one with 2GB GDDR5 memory.

"unlike GT 710" with a newer driver you mean? Tiny bump in OpenGL version or PhysX is irrelevant unless you have a specific use for it. IMO Practical observed stability goes above any trivial features or small benchmark gains. There is no "GPU health" as such. How would that be measured? Only thing I can imagine is obviously overheated cards and/or cards which show artifacts on screen. Or of course, 'dead' cards. If you are worried, buy a spare, that is what I do. I mean, I am not worried, but I still bought a spare. You could test GPU-Z display of "Bus interface". I found on older motherboards this may get stuck in a certain power saving mode, after warm reboots. Which then decreases GPU performance a bit, until it is 'woken up again'.

Epson All-in-One ET-16650 printer and scanner, A3 format. As sold new 03-2024. The currently offered drivers still work fine in Windows XP SP3 x86, for both printing and scanning. Tested through Ethernet connection. No tweaks or hacks necessary. Unbelievable...

I don't see how hardware PC health could be negatively affected. Not in general and not with this particular modification. If some essential part of the driver would be a mismatch with the hardware, then it would hang the system as soon as that mismatched code path was ran. Also, if people would have dying hardware because of such, it would be reported. My GT 710 system still runs fine.

Surely people at Vogons have tried similar things. I just did a quick search with the terms: ""site:vogons.org "windows xp" 486 cmpxchg8b"" See for example post from KCompRoom2000 here: https://www.vogons.org/viewtopic.php?t=82914 EDIT: Also the link to the POD tests at winhistory.de here: https://www.vogons.org/viewtopic.php?t=75778 PS, this is my 486 system, With DOS and Windows 95: https://www.vogons.org/viewtopic.php?p=1117089#p1117089

AssImp - The Asset-Importer Library Viewer and converter for many 3D Model formats (and their textures). Had some variants based on v3 core from around 2014 to 2017. x86 / XP compatible. Which work quite well. But the v3 FBX-format parser failed on a few model files. A moment ago I stumbled upon a 2023 build of the Viewer/converter and improved v5 core. v5.2.5.0 SVN of 01-2023 to be exact. That is neat! https://github.com/assimp/assimp/issues/4972 https://archive.org/details/assimp-master-build-21-jan-2023-winxp For completeness: Here is an older version based on AssImp v3 core. It is in the SDK installer and requires a little PE header patching to remove the NT 6 check. https://github.com/assimp/assimp/releases/tag/v3.3.1 Here is variation also based on AssImp v3, but with OpenGL-rendering instead of Direct3D 9. It does use .NET though. But this one works on XP, unlike other similar builds. https://archive.org/details/open3mod_1_1_setup

Good to know. May do that when the times comes and projects require it. (I use KernelEx tricks sometimes, but I find them a solution of last resort. It makes the kernelEx-ed software XP only, the software crashes when starting them from Windows 7 in a multiboot)

Bit of necro, So I just found VLC 3(.0.3) crashes when trying to play a H265 encoded video. libx265_plugin.dll contains NT6-specific kernel32.dll calls. The program crashes whenever that plugin dll is loaded on demand. Here is a great summery by gho421 on this particular issue, with XP API-tests for the other plugins as well. Showing these for version 2.2.2 up to 3.0.10. https://code.videolan.org/videolan/vlc/-/issues/24858 Now I can imagine tricks like: plugin re-compiles, kernel-ex wrappers, mix-and-match of files of different VLC versions. But for the time being the safest bet is to stick with VLC v2.2.1 on Windows XP. Yes it plays H265 without crashing. EDIT: below a copy of gho421's report on v.3.0.10

When I see some interesting hobby projects on github, it is fun to try and build a Windows XP compatible version from the sources. Preferably by building on Windows XP itself. For that one needs a somewhat recent version of CMake. https://cmake.org/ Official CMAKE packages dropped windows XP compatibility after this one: cmake-3.13.5-win32-x86.zip from 2019-05-14, last in 3.13 serie. ReactOS Build Environment package RosBE-2.2.1.exe includes an XP compatible CMake: cmake-3.17.2 (without GUI) By comparing and selecting MSYS2 packages, I assembled an XP compatible version of CMake v3.20 (08-07-2021). Available on my website in the back-ports section. http://www.gb-homepage.nl/index.htm

At work I need to put Windows 10 on some older Hardware. Z68 Chipset based. BIOS and MBR config. SATA SSD 500 or 1000 GB. 8 to 16GB RAM. I already had a test setup for it, from 2020 already, with Windows 10 installed from CD with some unfinished SysPrep intentions. Decided to connect it to the internet and let Windows Update run. Unfortunately, The 21H2 Feature update repeatedly failed. It is detailed as error 0xc1900200, Which may mean that the system requirements are not met. Now I suspect it is because back in 2020 I did not let it create a "System Reserved Partition". Thinking; Such dependency would just complicate the backup and imaging of the OS. I am not sure though if that is the sole reason. Windows recovery "bootcfg" failed too, saying "access denied". So instead I made a 22H2 install media on USB stick, and installed it like that. This works and runs, but the partitions are unchanged afterwards. so again no System Reserved Partition. No complaints from Windows setup about that (just the usual Windows 10 silly questions and remarks, styled as if the computer is a person now). On one hand I think I should retry and make such a partition in some way. On the other hand I read here that 22H2 is the last feature update, in which case I don't really need to ever do a feature update again, and I can just as well leave it alone.

Just tested a rapidgator download, and it works fine here. The captcha is to fill in a bunch of words. It passes. Roytam's Newmoon Version 28/10.6a1 (32-bit) (2023-01-13) Useragent switcher fake ID; Firefox 105 (Win10 64-bit) Ublock origin running