user57

how you actually write your vista driver ? when you have the code open source we might actually can make a new driver you didnt say the other part so clear, did you make a vista radeon driver of a new version ? like 2024+ ?

well the one-core-api has this function, it could be added to that msvcrt.dll there some ways to do so but it useally dont end with just one function but since chrome is open source why not just adding an exception handler at the place where chrome wants to have one it might can be ignored either, if that problem apears the app useally crash anyway - even with an exception handler

that thread function just set attributes if i see it directly without going into detail i would suspect it´s predecessor is CreateThread and createthread has higher tier function after that called ZwCreateThread or NtCreateThread // attributes : InitializeObjectAttributes(&ObjectAttributes, NULL, 0, NULL, SecurityDescriptor); // the thread function : Status = ZwCreateThread(&Handle, THREAD_ALL_ACCESS, &ObjectAttributes, ProcessHandle, &ThreadCid, &Context, &InitialTeb, CreateSuspended); that _except_handler4_common is a exception handler an exception handler apears when a bad memory access has happend (what not must even happen) any exception handler (and xp has this so do win98 or win2000) can be used for that we dont need that certain version ... for processes it´s Createprocess here the same thing happens we actually dont need the vista structure what maybe has 1 difference (startupinfo to startupinfoex) those 1 extra field is for misc stuff, not for its purpose (just to start a new app/process) - that can be done without that too chrome probaly use it to create new open "tabs" , nor it need that function InitializeProcThreadAttributeList to function that is probaly just a check if that function was set , then returning a error - that certainly can be reprogrammed again this extra info is not needed to start a such tab

well according to microsoft more then 4 GB ram is possible in 32 bit: https://devblogs.microsoft.com/oldnewthing/20090706-00/?p=17623 the microsoft guy even says how however it seems the other links are deleted from microsoft for whatever reason - is there something to hide ? he says CreateFileMapping can create 4 GB fragments (he isnt saying segments what is kinda correct) he also mention the 2GB to 3 GB extension, but 3 GB is a bit small when you actually want to more then 4 GB CS:EIP is a combination CS has 16 bits and the EIP (called the code segment(cs) and instruction pointer (IP)) for data there would be a way to use an extra DS (data segment) , that would give 2 (segments or with the ms guy´s words to fragments of 4 gb address room) the EIP pharse the instrutions , 4 GB for instructions is actually a lot DATA is unspecific but you guys can tell what that is , a RGB buffer, a FILE, a GAMES 3d object, a VIDEO - those are not instructions those are just memory or "DATA" DS is actually in use but even then why not ES (extra segment) or maybe making a better use of the GS (global segment) i tryed to ask if the 128 GB patch actually really works, i think it do - but actually i dont seen a proof the other already mention 2 approaches where the PAE and PSE not to forget the PDBR/CR3 (that is what points to a different app) and then the PTE/PDE ect. can point to different ram - so over that also more then 4 GB of ram would be possible the PDBR/CR3 solution is app limited, but having unlimited amounts of 4 GB app´s is not that bad

well x64 needs more ram then 32 bit but there more things involved that play a role - i dont know all either maybe it is time for a little discussion ... the instruction code is a such example, when you ask for a offset that is greater or in a higher area of 4 GB that questions is hit for example a other thing is coding a int 32 is 32 bit big (and so stored like this in memory) a int 64 need 64 bit memory thats a important reason that the offsets need to store instead of 32 bit 64 bits this "offset size, int size, register size" question follow you in many times i made a screenshot of a x64 app (regeedit.exe) , os is win7 x64 the app is shown as x64 in pe editor as in taskmanager BUT! it use x32 bit instruction sets ! as shown in IDA so in reality thats rather an 32 bit app when i compiled some apps up i saw a increase in file size not 100 % but like 50 % and if i rememberright (in some older x64 apps i meant to see they tryed to avoid this problem in only using 34 bits, that probaly dont work all the time)

you actually may not need a fast grafic card for an old game either, some just give a old card they no longer need - because often they get thrown into the trash bag even the prices are often like very low maybe 5 dollars, thats affordable

well unlike nvidia the radeon card´s have a different driver on the radeon´s website actually dont find any information what the latest drivers for certain OS´s are and what grafic cards they actually support that is better solved with nvidia, there at least for the moment you find that information i think the right status about that radeon driver is not having a well solution at the moment but again you might buy a nvidia card if you need windows xp as operating system radeon lost a possible customer that would pay them money ... the other only quick way would be to find the lastest available radeon driver, and buy a older radeon card (but thats not a good solution in my opinion) here is a such discussion https://www.majorgeeks.com/files/details/amd_radeon_video_card_drivers_for_windows_xp.html if you just missing a driver for XP for that card the kernelex patches actually dont solve all the problems (the one core api also has some) therefore your question would actually change, you dont have a OS problem anymore you have a missing driver problem, due a unsupported grafic card some games actually dont work on certain operating systems, i think Resident Evil 1 needed windows 95, and the dos c&c 1 version needed a nativ DOS bootet, and did not work on windows xp either in that case the one core api dont fix the missing functions either - coming from an ask that a newer grafic driver just installs on xp the current kernelex dont have enough functions for a newer driver ---- that with the driver signs i had actually that problem too (win7 actually also has different versions with 2 different driver sigs) my usb drive was blocked and i wondered why it did that with with an upgraded win7, while the non upgraded win7 did work no problem the solution is simple while boot up you bring up the boot screen via holding F8 when the first menu apears you hit F8 again , and there you can disable that driver signature after that my usb driver installed without any problems ... ( i still use a digital camera, useally not smartphones) you should have been told that the most games actually dont run not having a grafic card driver, makes sence after all

well it is actually hard to say if your problem is a driver problem for xp we have this one that support new grafic cards: https://msfn.org/board/topic/181454-official-driver-supporting-gtx-970-and-up-found/ your card seems to be a desktop card, there should be drivers ---- in lap-tops some companys use a trick to hide the grafic card they use for example they use a different name for their card like instead of "gtx 970" they use "gtx 970 hps" this make the common driver that would work not install but you actually can add that driver into the .inf file (then the driver finds thecard) in the driver installation you also can choose a driver that you want to install for this card you can force to install that driver, that also works but thats for lap-tops using that "trick" ----- what games are these actually ? xp has a well balance between backwards and forwards compatibility that win7 + (use xp modus) is useally not working often a placebo that trys to make people think win7 can do what xp can do (for marketing like "oh win7 can do xp, buy win7, buy win10") (what in reality it can 99 % it can not) to fix your problem we should take a look on the game or game(s) itself beginning with the question what game you first wanna use then we should know if that games or games work in a nativ xp machine or not (xp dont have 100 % backwards compatibility either - xp might have a lot here but not the absolution) the "one core api" is not for old games, it actually is for new applications and games that have functions that xp useally dont have the one core api therefore is for new games, therefore it dont make sence to use one core api on a old game if you have a dos game it might dont work, dos games often need a nativ 98 machine bootet into dos commander and conquer 1 (dos version) would be a such example, there are windows versions too but ---------------------------------- to the 8 GB problem i can say this there are page table entrys (that leads to the physical ram) , those can addresse more then 8 GB however for example the EIP 32 bit for example can only address 4 GB then however that EIP - can with a different application (games are applications) address different memory - therefore having multiple applications/games/executables/processes (is always the same thing) can theoretical address more RAM but here is the next problem the OS not always seems to use that (xp dont i heared, with the 128 gb ram patch it suppose to do it) a other thing i have to point out are segments, and maybe the PDE/PTE´s a segment is a selector for a specific amount of ram since we have 32 bit (what are 4 gb ram) every segment actually could point to 4 GB ram but applications dont use segments (what you can see with a disassembler every common app dont use segments) these segments are in 32 bit protected mode 16 bit wide therefore 16 bits = 65535 and that times 4 GB = 65535 * 4 GB ram https://en.wikipedia.org/wiki/X86_memory_segmentation a other problem can be the hardware lets say the CPU only got 32 wires , then the software could use more then 4 GB ram but the cpu actually can not use more then 4 GB ram but here kick in the segment question again the CPU actually can translate that segment to the next segment (the first segment is 4 GB ram) - again with 65535 possibilites (4 gb * 16 bit (65535)) so the 32 bit protected mode actually has not 32 bit, the 32 bit protected mode theoretical got 48 bit (32 bit + 16 bit) but as i said, neither the operating system the software (such as segment selectors) (applications/games/executables/processes) or maybe even the hardware (such as cpu wires, or even the "manufactor logic" on the chip itself) dont/cant do that what the hardware can do is translate a segment/page table entry/page directory entry just into a next access somewhere into more of the 4 gb ram as said the os, software, and hardware have to do so ... (in 8 bit and 16 bit times segments where very common instead of 65kb ram (16 bit) you could then use 1mb (20 bit))

i had some time to look at it his function is bugged up basicly he write it himself "// hacky but works: small number is command id, large is submenu (a pointer)"" that indicates he tryed to trick it and had to do "hacky solutions" he use a endless while true loop with ends escapes and continues, also goto´s,calls to the same functions in a function, and then seperators for his menu to fix how he called it "bad seperators" this ended up in a os specific solution you useally dont do that because the operating system can might change the relevant codes over time (such as 7 to 10, or 98 to nt) - something like this already happend in the past but the solution to this problem is possible as the author use multiple fixes all over around (probaly to compensate the problems) it is easy to add that fix - anyways it actually was a bit weird to find it out in win10 its just a flag, that was found like very quickly - but no in xp it didnt do that and gave very weird results even turned off xp was working ok while on 10 it did what it suppose to do - it was turned off as the code suppose to was changed to let me know if it works now! the .avif file (and heic) format dont work for me, neither on xp nor on 10 nor on 7 (both unchanged from the official download page and the changed version) https://www.file-upload.net/download-15277606/SumatraPDF_WINXP_3.5.2_3.zip.html https://www.mediafire.com/file/a4mtyf33ozs6q63/SumatraPDF_WINXP_3.5.2_3.zip/file

https://www.file-upload.net/download-15277606/SumatraPDF_WINXP_3.5.2_3.zip.html https://www.mediafire.com/file/a4mtyf33ozs6q63/SumatraPDF_WINXP_3.5.2_3.zip/file

i wrote a private message with a changed version it might fix the annotation bug for the choose the other bug i try to look later

well you might make the test then heic vs jpeg xl you have to look a high resolution base file then create the 2 formats from that high resolution base file in this case jpeg xl and heic and compare if fine details are vanishing or are no longer shown the compression with more remaining pixels wins (a similiar file size should be choosen)

a good idea is to look what the driver useally is called often a lap-top/desktop comes with a driver disc, that might also give that information a other idea would be to look a pre installation with a working operating system (like if its a nativ vista computer you gonna look the device manager) google might also help what type is related to that computer/motherboard in the past there where no onboard lan drivers, they where actually cards (those can still be buyed if the motherboard dont support the operating system) the same goes for the sound card, those still exits to buy too for hdd´s/ssd´s harddrives whatsoever they most likely where onboard, but actually you could also buy a card if you wanted some more harddrives then the motherboard actually supports but why im trying to say is that you need the right names for the drivers

well here is a experimental version of sumatra pdf for windows xp i had to change lots of things this time, so it might have errors but we should test it out version 3.5.2 is lastest version up to current date 21.02.2024 https://www.file-upload.net/download-15277606/SumatraPDF_WINXP_3.5.2_3.zip.html https://www.mediafire.com/file/a4mtyf33ozs6q63/SumatraPDF_WINXP_3.5.2_3.zip/file edit to fix some problems

good ideas are to tell the hardware what you actually did (HP Pavilion dv7-1127c (the last number is important because there many dv7 types, these are very different) a other good thing is where exactly the problem apeared like while the installation, or after windows bootet up, did it find the HDD, did the boot cd/dvd/blu ray start up a next good idea is a picture with a smartphone or camera the net give us different informations can that be possible ? (AMD Turion 64 X2 Mobile), that cpu is kinda old (2005) the grafic card (AMD Radeon HD 3200) is like 2008 specs say these types HP Pavilion dv7-1127c was released in like 2009 you may enter your lap-top number here to get the right information: https://support.hp.com/rs-en/product/details/hp-pavilion-dv7-1100-entertainment-notebook-pc-series/model/3802944 a problem can be the chipset of the motherboard but 2005-2009 sounds rather XP like, the specs for this lap-top actually say that this lap-top is a nativ windows vista lap-top cant find anything related to the chipset for now

well hard to say this player is not very self-contained this also is seeable when you compile it up(like dibya did), it need this exact version of phyton , the next exact version yasm, the right vs versionthe right SDK version, the the LAV version (only to name a few) that more or less says it break if any of these versions (i always call them engines) changed from like 0.0.0.0.1 to 0.0.0.0.2 if one of these decided - no i do not longer want xp - its like a directly "bugged up" that raise questions .... if that player is a good choose or not and hardware dependencys ? if we always need a new grafic card for a codec - i dont think thats what we want - what we want is a working codec (or more precise a decoder for certain formats) not always a new hardware (aka a new grafic card that can do this) can somebody tell me why new programmers like to choose this kind of path ? instead of going for a "decoder to RGB" (display) solution this problem you see kinda often today - you actually didnt in the past sure actually this player is getting probaly getting fixed up, but always like that ?

Skype 8.75.0.140 + it has worked having an existing login that it remembered but not if you install a new installation installs but you neither can register an email nor the login works the login screen looks bugged up and the buttons do not react (existing user valid login tryed)

it would be possible to instead of the gpu doing that decompression maybe "gpu acceleration" to create a own decoder the decoder is very less cpu intense but it would for me only for the h.265 codec, then the others still would be missing (what actually is a big disadvantage) not to say that the code then has to done the way so its fits into that mpc-hc player the reason why a lot of these use SSE2 is that SSE2 has a huge range of CPU´s that can use that, its like a common available hardware acceleration some can understand this when using that "windows xp heic" en/decoder the decoder is like instand the encoder can take severial minutes only using x86 commands the encoder using mmx-avx512 can speed to this like 20-100 times faster a other big problem even if you have windows 8 the decompression might not be available still because the grafic card dont support that then it would be a GTX700+ to have the "gpu hardware acceleration", probaly why some have choosen SSE2 instead - relativ compatible - many cpu´s that can do that - significant speed boost so going for a "all codec solution" end up in making a own video player (and all the codecs) ... thats some work to do FFMPEG can be used also as video player - that raise exactly this question FranceBB actually know about this since he is doing things in that direction, but he hasnt said anything here yet

well might be a idea, do 3.5.2 + has useful changes ?

well it wont be that simple with just a why question but what we can hear is that already 3 things are there that xp actually dont have dx10/11 was called out/dxva2 lav dll´s where called out (that actually use vista functions) srw locks (vista api´s) (probaly not the last ones, makes already 3 why´s) this is the "engine problem" i described they take in those engines all the time one being that LAV engine that should solve the things like "here is your mpg file do it" that actually dont decode a codec it use a engine to do so lets say that LAV is not public it already ends here - because if its not published you dont have insight a very common problem of this "engine, script like, copy/paste like coding" problem if LAV is public we had some luck the other alternativ would be to copy a real decoder (that actually has to be applied to that MPC-HC player) that LAV "engine" seems to work to some extend on vista the hardware acceleration can either come from a certain engine too like DX or SSL/other variant of that registers that x265 can do a lot more registers then SSL i dont think MPC-HC is self sufficient if it has to call engines that requie certain operating systems going a such way often end up in "now not being supported" what we certainly actually can say is that the MPC-HC player is not very self independent special with the mpc-hc author decided to choose to use engines having the source code published certainly makes possible to find the related problems it´s a entire player but, something to read and then collectiong all of its problems but here your question is also answered - mpc-hc is not very self independent it was luck that the mpc-hc "engines" still worked on vista what they do not on already on xp it might dont use that win10 engine - probaly it vista dont have the heic image format - but it do use engines

well it might be a guess the problem might resides in that LAV files (what are again something like a engine) (those seem to use vista functions) (actually to find the precise answer i would need to read the entire code of this programm) i did say quite the opsite i said that this "codec engine" is split up into windows 10 (through severial engines/dll´s (what actually made it hard to split it from windows)) not that the math and logic cant work on xp / nor vista i saided the oposite here - and thats what i did with the winxp heic en/decoder a cpu might can take a software calculation hardware based (in fact that is very common because software solutions are useally not that fast the 4:2:0 transfer for example is made via hardware it also can be done by software) the directx (d3d12.dll) or (opengl) do also hardware processing https://www.youtube.com/watch?v=fm-nGCXRZqI as we can see the pixels also look different those GPU can make a billinear upsampling (but this also can be done via software and via that SSE registers) with SSE registers we have to dig a little before there used to be like 90-120 opcodes what solved all kind of possible math questions what you actually could do instead of c / cpp ect.you could write your code in assembly that speeded up code like at least 2 times up to 50 times (and smaller file sizes) the second idea was a new technolegy for speed those then because extra registers and firstly where called MMX MMX up to AVX512 are the same thing but they are extra registers called the XMM0-ZMM31 registers SSE and AVX where just extensions and MMX got named to SSE those are made for speed and have different hardware logic and progress a lot faster (and actually windows 10/11 try to make use of these because the win10/11 engine is very slow) and even tho this are far faster registers win10/11 is still very slow a other thing we should not forget is that vista is a child of win10 there was a big change from kernel version 5.2 to 6.0 (win7 is 6.1), win 8.1 (rather 6.2 (maybe 6.201), and win10 is 8.1 maybe a few new functions and at least in my opinion spyware components vista has those SRW locks win10 still use them , they are often a gap between xp and win10 why the "win10 programm" actually not work win7 was the only win version that tryed to make something better - probaly made its success

well a problem todays programmers follow is that they use engines, very script like code, very much of copy paste often multiple engines like a engine for a engine and a next engine having weird names like "STD" or "dwrite" or "grafic engine" when i looked that hevc (h.265) public code i reconized like countless engines and scripts and this weird c++ 17 c++20 c++23 (that in reality do nothing useful like at all) then you see them copying these codes all around later on i actually saw that h.265 not even use a own encoder/decoder it loads external functions and dll´s via a script like "erm do me C:\123.mpg to C:\123.hevc" that was actually the reason why xp cant decode/encode that codec it is slit up into a windows engine from at least win10 ? maybe 7 i dont care and this dll what also is a engine (that calls a next dll -> and that 1 is doing the real codec) so after seeing that you have like 5 script engines before even 1 things happens (and those trigger like 100 of unneccesary functions older os´s dont have) so digging deeper into the rabit hole what do the real codec is not windows either or ffmpeg its that x265 codec that is doing the real codec the entire real math, logic is there https://x265.readthedocs.io/en/master/releasenotes.html#version-3-5 the math algo functions are in plain c - and in fact dont care about about some weird windows functions that not being done the heic encoder/decoder i made dont need anything of these files - because it do what a encoder should do it use the en/decoder relativ directly a illness we hopefully dont face too much

the forgotten chipset cooler question ? it either go for "chipset cooler" "southbridge cooler" "northbridge cooler" HB-802 Northbriddge Noctua NC-U6 Swiftech BC-MCX159CU Enzotech SLF-40mm https://www.amazon.in/Heatpipes-Radiator-Motherboard-Northbridge-Southbridge/dp/B08NBCFJPT https://de.aliexpress.com/item/4001133383805.html it seems quite hard to find these today, ebay gave very bad search results i think the chipset cooler got quite forgotten, Saxon made a good point if the chipset is getting to hot just get a chipset cooler as Saxon mentioned there are chipsets that can handle more then 1333 mhz/FSB like 1600 or even higher , a 1600 mhz/FSB chipset clocked to 1333 is running underclocked - what actually means it should keep cool

well i looked around his (Intel Pentium E6800 @ 3.33GHz) has a benchmark of: - 1151 your old motherboard was a intel LGA/Socket 775 MB this is a relativ complete list of the 4 core cpus for LGA/Socket 775: arranged by performance test from cpubenchmark.net: for the prices i have looked ebay: Intel Core2 Quad Q9650 @ 3.00GHz (45 nm) - 2456 - price ~ 33 € Intel Core2 Extreme X9650 @ 3.00GHz (45 nm) - 2394 - price ~ 75 € Intel Core2 Extreme X9750 @ 3.16GHz (?) - 2348 - price ? Intel Core2 Quad Q9550 @ 2.83GHz (45 nm) - 2333 - price ~ 15 € Intel Core2 Extreme X9770 @ 3.20GHz (45 nm) - 2312 - price ? Intel Core2 Extreme Q6850 @ 3.00GHz (65 nm) - 2304 - price ~ 49 € Intel Core2 Quad Q9500 @ 2.83GHz (45 nm) - 2215 - price ~ 15 € Intel Core2 Quad Q9450 @ 2.66GHz (45 nm) - 2178 - price ~ 24 € Intel Core2 Extreme Q6800 @ 2.93GHz (65 nm) - 2145 - price ~ 85 € Intel Core2 Quad Q9400 @ 2.66GHz (45 nm) - 2111 - price ~ 22 € Intel Core2 Quad Q6700 @ 2.66GHz (65 nm) - 2097 - price ~ 6 € Intel Core2 Quad Q9505 @ 2.83GHz (45 nm) - 2084 - price ~ 21 € Intel Core2 Quad Q8400 @ 2.66GHz (45 nm) - 2076 - price ~ 6 € Intel Core2 Quad Q9300 @ 2.50GHz (45 nm) - 1965 - price ~ 6 € Intel Core2 Quad Q8300 @ 2.50GHz (45 nm) - 1885 - price ~ 4 € Intel Core2 Quad Q6600 @ 2.40GHz (65 nm) - 1811 - price ~ 8 € Intel Core2 Quad Q8200 @ 2.33GHz (45 nm) - 1774 - price ~ 3 € Intel Pentium E6800 @ 3.33GHz (his) (45 nm) - 1151 - price ~ 12 € unknown charts, but they should be similiar to the the chart list above Core 2 Extreme QX6700 Core 2 Quad Q9700 Core 2 Quad Q6400 Core 2 Quad Q8200s Core 2 Quad Q9450S Core 2 Quad Q9505s Core 2 Quad Q9550s the gap between these cpu´s is not that big and they are cheap to get and have higher benchmarks then yours even in cheap range of 3-33 € ----------------------------------------------- for a grafic card he should go to: https://www.videocardbenchmark.net/ and look for a cpu you want to have for the right money then you should take a look if for that card is a working windows xp or not this is done here: https://www.nvidia.com/download/index.aspx --- coolers i already gave a nice list before - "socket 775" --- ram for those Core 2 Quad CPU´s useally are 1333 mhz (the most ram can underclock to this speed) if he keeps that old cpu (Intel Pentium E6800 @ 3.33GHz) he can upgrade his ram speed (he runs underclocked) the right terms is : DDR2-1066 PC2-8500, DDR2 1066, PC2-8500 for ddr3 right term is : DDR3 PC 1066 the most DDR3 rams can be underclocked to 1333 MHZ (PC 1333) im 1066 can make problems in this case he can buy any DDR3 ram speed what is above 1333 mhz

well: aeroadmin but actually i think we have to understand something very important in the past there where a mp3 player called winamp it made like countless upgrades said to make everything better - but actually it was always the same quality of sound at the very best a few bug fixed where needed and for misc maybe a nice gui (what is not neccesary requied its just a view question) what we have to consider if that "new versions" even brought some useful new functions (what they often do not - often even they decrease in quality like more cpu useage and such) it always like "here is a new version we must have this" - what actually might not be needed - it can be needed - but not always and on the other hand we should consider that we might dont have benefits or less benefits (such as just a new grafic interface) if we could we might should ask teamviewer why they even do this like that - they dont seems to have benefits to me just getting their software less compatible (what actually means less possible customers)