Jump to content

Tux

Ultra Members
  • Posts

    1,056
  • Joined

  • Last visited

  • Days Won

    219

Posts posted by Tux

  1. The 7z code in raine is very old, 11 years already, and the distributed code changed quite a lot in all these years...
    Since the performance of the 7z archiver didn't change a lot in the same time, I didn't expect a big improvement for the performance, but I was curious to try to update this anyway...
    In the end it was long, complicated, and for almost no improvement for the performance, quite disappointing !
    I'll keep the update anyway, maybe I missed something in it, and it's probably better to have some more recent code.

    Things to notice : by default 7z archives are solid, but for some big rom archive, making it solid gains about 1% only in size, almost nothing. The time to decompress is almost the same for solid and non solid archives, but if the archive is not solid you get a real progress bar when loading the rom. If it's solid, then the whole archive is processed for the 1st file, and then the rest goes so fast that the progress bar becomes totally useless ! So it's probably better to create non solid archives (-ms=off when using the 7z or 7zz command line in linux).

    • Like 1
  2. 6 hours ago, mer-curious said:

    Great news Tux!

    I agree with you, I think it's better to switch again to the faster DLL once the bug is now fixed and everything works normally.

    Thank you again for your work.

    It's smaller, but probably not faster, mpg123 is very concerned about speed too, it's full of assembler code inside so I doubt it's slower !

    But I reverted to the old mp3 driver anyway... ! (I kept the mpg123 code commented in case it's useful again one day !)

    • Like 1
  3. Bang, Iccuslus suddenly exited from hibernation and fixed sdl_sound... ! See there : https://github.com/icculus/SDL_sound/commits/main almost 1 month later, but better now than never of course...

    Now I hesitate : is it better to remove the new mp3 dll and return to the old decoder and just forget all that and stay with what we have ?
    Well luckily only 1 version used the new dll, and it was included in the raine archive, so removing it now will have no consequence.
    So I guess it's probably the thing to do, the mp3 decoder used in sdl_sound is still smaller and maybe faster...

    Oh at least I'll have to try it first. Apparently the bug was that when the decoder returned something less than the expected bytes to read (which happens for the last block), it was understood as eof (end of file), and so the last block was not handled. Oh well... He finally closed the bug thread on github with this fix.

    edit : yeah, the fix works !

    • Like 2
  4. 1 hour ago, mer-curious said:

    Hey Tux! I'm sorry for the late reply. Don't worry about the shaders, it was just a suggestion that came to my mind again when I found this pack from Emu-France.

    By the way, have your motivation to add the Konami drivers for TMNT Turtles in Time and Sunset Riders increased somehow lately or are these drivers really too complicated to support?

    Anyway, I was doing some game-play tests with KOF99 in the FBNeo Retroarch core and stumbled upon this very weird Unibios error:

    https://imgur.com/a/Kv5GxCY

    I never expected the Unibios to hang the game like that. It has never happened to me in Raine fortunately.

    Should I report this error to the FBNeo core maintainer or is this expected when using the Unibios?

    Thank you in advance for your attention.

    No it's not a unibios problem, it just displays a nice screen instead of just crashing here, the error shows the pc has just gone out of bounds, it should not have this value, it's an emulation problem, which is a little surprising since fbneo is based on mame, I didn't think they touched the emulation part very much, but I don't know it (no linux version !)

    For tmnt, I didn't even look ! It's probably not an easy one considering konami reputation, but no idea.
    For now I am looking at other things for a little while, I don't spend all my time in emulation...

    • Like 1
  5. 4 hours ago, Robert said:

    When I tried to download the new 32 and 64 bit 7z files with an older browser, it wanted to download .man files instead. Works ok on a newer browser.

    It's a misconfiguration at the server level, mime types configuration, with wget you get this :

    --2022-03-04 14:09:08-- http://raine.1emulation.com/archive/raine32-0.93.4.7z
    Résolution de raine.1emulation.com (raine.1emulation.com)… 216.245.218.214
    Connexion à raine.1emulation.com (raine.1emulation.com)|216.245.218.214|:80… connecté.
    requête HTTP transmise, en attente de la réponse… 200 OK
    Taille : 2902845 (2,8M) [application/x-troff-man]

     

    (the text is in french, sorry, but the mime type appears between () at the end : application/x-troff-man, it's a man page indeed, but it's not my fault).

    On my side, it's a simple <a href link, produced by some basic html file, I don't touch the mime type configuration...

    And, from there for example : https://github.com/mime-types/ruby-mime-types/issues/32
    the mime type should be application/x-7z-compressed

    it's minor though, even the most basic browser would kinow how to download this, whatever the mime type is.

    • Thanks 1
  6. Yeah sorry I should have fixed that the 1st time. The reason why there is the option for neocd/neogeo is that it's a single driver for roughly 100 games, so the search for these speed hacks is generic there, and sometimes it can fail to find something good. You are right it's usually not really necessary anymore nowadays but it's in the spirit of raine, optimize as much as possible, just because doing it is great ! ;)

    It would have been nice to have the option for all the drivers, for those who prefer to play with a choppy animation when the game is overloaded with sprites, but it's tedious to do it now, so maybe one day, but not for now...

    Except for neogeo where the driver sometimes fails to find the right spot, these hacks are made to end the emulation of the main cpu when it just waits for the vbl, instead of just wasting cycles looping for a vbl which won't come because it requires an irq. It saves some cycles, but on modern hardware, it's around 1 or 2% usually (on an optimized build, more on a debug build in C), but it's the right thing to do (mame prefers to emulate these wait cycles, because it's more accurate).

  7. 5 minutes ago, Neville said:

     

    How do you do that?

    neocd/neogeo options / allow speed hacks : no

    And you'll probably need to reload the game after that because the speed hack installs itself very fast !
    They'll be disabled by default for this game in next version.

  8. 2 hours ago, mer-curious said:

    Hey Tux! Thanks for this super fast fix! I'm glad it wasn't that hard to fix in the end! 🙏

    By the way, I've stumbled upon this site which collects shaders for various emulators:

    https://emulation.gametechwiki.com/index.php/Shaders_and_Filters

    In the Emu-France shaders pack for Raine/BSnes there are some which aren't bundled with Raine. I was thinking if it would be a good idea to link to this page/pack in Raine's downloads page? Maybe this could help new users know about them and try them.

    However, there are some shaders in this pack which didn't work here. The 2xSal-convert gave me a black screen and the two NTSC ones returned a syntax error. But except these there are some interesting ones such as XBRZx4, scanlines-3x and 5xBR-v3.7c+CRT.OpenGL(hunterk).

    But I still dream of Raine supporting the new GLSL shader format one day, which currently have much more shader options than we find in this pack, for example, for CRT/scanlines, see the numerous possibilities here:
    https://github.com/libretro/glsl-shaders/tree/master/crt/shaders

    I've already played with some of them in Retroarch and some work really well with these old arcade games.

    Anyway, just some suggestions to consider for future releases.

    Thank you so much again for your time and work. 👍

    Yeah yeah the pack was made hastily, I doubt any direct3d shader would work since raine uses opengl, although maybe ?  (seems unlikely though !).
    So yeah some shaders are broken, they were clearly not even tested, just assumed to be compatible.

    For the glsl shaders, yeah good idea, but I am no shader expert, I already had some trouble to add the old shaders support, so... !
    But I agree, it would be a good idea ! (actually the name is bad, all shaders for opengl use what is called glsl).

    • Like 1
  9. 2 hours ago, Neville said:

    I'm seeing some glitches in "Neo Turf Masters", Windows x86 build, using the Scanline-3x.shader. Don't know if this is normal.

    8ZJ7Qyd.png

    They appear in the hole overview screen only.

    make sure you disable speed hacks for this game, I thought they were disabled automatically but apparently not.
    I just tested the neocd version with this shader, apparently it's ok...

  10. 1 hour ago, mer-curious said:

    Hey Tux!

    Just a quick feedback about the sound associations issues which I reported in the other thread: everything is working correctly now. The only issue I noticed is if a track start playing in the game, and then before it finishes you decide to enable the loop effect for it, when you return to the game the loop function will not take effect immediately, that is, the track will end and stop. I don't know if it's a limitation of the function, so I'm just reporting in case you didn't know about it.

    But I found a work-around for this: if you enable the loop function and click on "play track" in "Manage associations", the loop will work for that very turn of the game.

    Anyway, it would be just a tiny refinement since I cannot think of many use cases for this. I only stumbled upon it doing some later adjustments in my KOF2001 associations.

    Thank you so much for this very fast new release!

    Yeah it's very tiny indeed, I wouldn't have thought about using the loop in play track to change the setting when returning to game...

    Well considering that before this version returning to game would simply stop the current track which played, it's already a huge improvement that it does not stop anymore !

    Actually the loop trick you find is the "exploit" of a bug, I forgot to restore the loop status when returning to game, so here it just takes the last loop status... ! Ok well I'll restore it from the track setting next time and not from the last value it had.

    • Like 1
  11. 37 minutes ago, pmc2 said:

    for me, the background color has disappeared on the 32-bit version (there is no more blue), that'all ^^

     

    ... and if you had read the announcement with attention you would have noticed the advice to go to options / colors / revert to

    to restore the colors to their preferred theme (blue by default). Before you ask, the effect is not immediate, you need to close the dialog and open a new one to notice a change, so after choosing which colors to revert to the easiest path is to just quit and reload the emu.

  12. This binary is only about the sound associations, particularly those related to kof2001, so if you are not interested in this, you can skip this one because it changes the colors encoding so if you load it with a previous config file you'll need to use the "revert to" command from the colors menu in options. It also includes a new dll for libmpg123, which is included in the archive for this version, it will go to the dlls package soon for the next versions.

    You can see the details of what happened (and eventually get the audio tracks from mer-curious !) from there :

    http://raine.1emulation.com/download/latest.html

    • Like 2
  13. Well it does a lot of rom bank switches while on this screen, a few / frame, which is a little weird, I wonder what it needs to fetch while on a selection screen, but anyway the bank switches are just emulated by a pointer which changes, it's extremely fast, and you can see that with the profiler %, they remain low. I thought maybe some kind of weird input which is checked, but it's neogeo, meaning all the games use the same inputs normally (cartridges !). So for now no idea...

    For the fix for sdl_sound, I just reverted to the decoder used in the 1.2 version, which worked well, so it means you wouldn't have this bug in the 1.2 version. I think it's quite unlikely to see that fixed, the author seems to have mostly abandoned the project, plus the bug is not in sdl_sound itself, it's in the decoder he chose, which is too experimental. It was convenient though, very small and very fast, so can be merged without needing an external lib, libmpg123 which I use now is huge in comparison !

  14. And finally found a workaround for your weird colors in game in windows after displaying the sound association dialog :

    it can't be fixed, because it's totally crazy ! Actually the color you see is the color used by the lines of the cross to indicate a loop is enabled. Here you couldn't even see the lines since their alpha was 0, but the color indexes still counted ! So the only way I could find to work around it is to make sure the last line drawn here is white, this way there is no color applied to the game bitmap and everything is fine !

    I am glad it's windows specific, because it doesn't make any sense !

    So you'll notice now there is 1 green line and 1 white line for the cross... temporary workaround until someone finds something better !
    The fix is in windows only of course, I keep the normal cross otherwise.

    I think that's the end of the fixes for your problems, I had a quick look at the slowdown in the game menu, and no idea for now !

    • Like 1
  15. The invisible cross in the loop checkboxes was simply a bad handling of its colors, it's because the colors changed their encoding in sdl2, I tried to keep some compatibility with the old ones, but it's the last problem here, I switch to the new colors, so next time you'll run raine, you'll have to use the "revert to" option in the colors menu (from options).

    The sound which loops badly is specific to the mp3 format, at least for me, I don't have it in wav or ogg format. This has been reported there : https://github.com/icculus/SDL_sound/issues/27

    The problem is that the author doesn't seem to react much. So for now the best solution is to use ogg or wav instead of mp3, it's not specific to the vbr contrary to what I thought 1st, it cuts the end even with a plain 128k mp3 ! It's possible the problem doesn't happen in the sdl1.2 version, they changed their sound decoders in sdl2_sound, and might have chosen wrongly the mp3 decoder here. It's fixable since it's open source, but it's not an easy task !

    Just checked : it also works correctly with flac, it's really specific to mp3 for me.

    Ok problem fixed by taking back the old mp3 decoder we had for sdl-1.2, and which used mpg123 (it uses minimp3 here). It works, but it requires to link with libmpg123, which means 1 more dll for windows, yeah !!!

    • Like 1
  16. 8 hours ago, mer-curious said:

    Ok, thanks for the info! It's probably a graphical glitch in the GUI then, right?

    Sadly no, it's an "update glitch" you could say, with sdl1.2 I could update just tiny regions of what is displayed, now I can't anymore, the whole thing must be redrawn, and it creates problems in this kind of case.

    8 hours ago, mer-curious said:

    Interesting. Here I can play this track normally until the end in all the players I've tested with it, namely foobar2000, Media Player Classic, VLC, Windows Media Player and Chrome. I used foobar2000 to convert the wav tracks to mp3 using the LAME encoder, which is a very popular mp3 encoder. Also, Raine 0.91.18 plays fine this file without cutting the last second. I've tried to convert the file with LAME encoder again using a constant bit rate profile but it still kept skipping the last part of the track. Then I tried a generic mp3 encoder in VLC and the cut not only persisted but got worse in Raine. Finally, I tried this new converted file in Raine 0.91.18 and there was no cut there.

    So there seems to be some kind of incompatibility with Raine SDL2 in terms of playing (some) mp3 files integrally?

    Anyway, converting the tracks using the Ogg Vorbis format works around this issue for the moment with the latest Raine versions.

    My bad, I used audacity to load that, actually the track was playing fine, it's just you don't often hear tracks ending like that, so I confirm there is indeed a problem with the loop setting while using mp3 ! Maybe some other formats. Only way to fix it for now : convert to wav (using audacity for example !), and edit your games.cfg to use the wav version instead of the mp3, it will work flawlessly.

    Notice that I was stuck at an audacity v2.x, the latest one 3.1.3 has a nice loop button with a tutorial on how to make nice loops, but for this track it seems to loop perfectly without updating it (in wav !). I'll try to fix the mp3 problem later, it's odd we never noticed that earlier though, it's still sdl_sound, I don't think they changed the way they played mp3, so it's probably been there for quite some time...

    8 hours ago, mer-curious said:

    I'm happy you were able to reproduce it somehow. I could try to switch to Linux eventually, but I suppose it's also good to have this stability parity between different operational systems.

    Yeah no worry, I'll have a look, I don't know yet what it is though.

    8 hours ago, mer-curious said:

    Interesting. I don't remember that exactly now. But anyway, I guess fixing this slowdown could benefit other NG games too perhaps?

    Yeah yeah but it's not easy usually !

    8 hours ago, mer-curious said:

     

    Thank you so much again for your attention and time spent with all this testing. 🙏

     

    • Like 1
  17. 7 hours ago, mer-curious said:

    So, I'm also experiencing a weird slowdown in the order select screen in KOF2001. The cursor movement becomes very slow when I try to move left or right. I also notice that the animations in the screen as well as the CPU order selection are also affected by this slowdown.

    For comparison, I went and tried this game with FBNeo and couldn't reproduce this slowdown there. I've made a video of the issue showing in Raine and then comparing it to how FBNeo performs with this game:

    https://drive.google.com/file/d/1SaypCppv08YnY5WZ5XRjsWO1XB5McJpR/view?usp=sharing

    I don't if it's possible to see the slowdown, but it's very easy to feel it when trying to move the cursor with a controller or keyboard. When the CPU order is finished, the speed returns to normal.

    It's not an issue that hinders the game-play though, but maybe it could also affect other moments of the game which I'm yet to realize if I start playing more KOF2001.

    Anyway, I'm reporting it in case it was an unknown issue.

    Thank you so much again for your time and attention.

    I seem to remember you found something very similar in another kof game, and we never found the solution, and there was no over slow-down found.
    I might try to take a look, but not now... it's the kind of thing where I'll search for a long time, and it's probably not worth it.

    • Like 1
  18. 7 hours ago, mer-curious said:

    Hey Tux! I was trying to create sound associations for the KOF2001 game and then stumbled upon some issues which I think could be fixed or improved. I'll number them below to help us with the discussion on each of them:

    1. The "play track" function is broken in the new SDL2 versions. When we select "play track" both the associated track and the sound command are played simultaneously.

    I don't have this, only the associated track is played for me. There is a problem about the loop setting though, it plays in loop if there was a sample which was playing in loop when going to this dialog, which is not very reasonable. Tested on the intro song 21h of course, with a short wav sample looping, I hear only the wav file when doing "play track".

    7 hours ago, mer-curious said:

    2. The loop check-box is no longer functional in the SDL2 versions, neither by pressing the space bar or by clicking in the box with the mouse cursor.

    Yeah you can't see if the box is checked or not, but it seems to accept a change of setting (the setting is correctly saved in the config file but you don't see it on screen).

    Yeah I know where it's coming from, it's annoying.

    7 hours ago, mer-curious said:

    3. It would be useful that when we had the loop check-box marked for a track it also looped in the Raine player so we could preview the looping effect before starting the game.

    Yeah well for now associate something which loops to the 1st song, 21h, and when the song starts ingame, go to play anything in this dialog, it should loop !

    Yeah I know, it's ridiculous... !

    7 hours ago, mer-curious said:

    4. I don't know what the "Test this" option is for.

    Already replied. Its main purpose was to find the sound effects commands actually, starting a song the usual way, then testing what a given command does to it.

    7 hours ago, mer-curious said:

    5. Since the loop feature is broken in the SDL2 versions, I used an SDL1.2 version to create the sound associations for KOF2001. Then when I tried it with the latest Raine version I noticed there is an issue with the loop feature and the track played in the KOF2001 character select screen (file SND01_00041.mp3 from my pack). The loop is applied before the end of the track (maybe 1 or 2 seconds to the end). This produces a noticeable seam effect when the track loops, which doesn't happen in the SDL1.2 version I tested. I suppose this is happening because the track is too short (just 8 seconds)? Anyway, it would sound better if it could loop perfectly in the most recent versions as well as we had in the SDL1.2 ones.

    I don't have that, I used a 1.79s wav, and it loops perfectly...

    After seeing at the end of the post that you posted directly your files : encoding problem apparently, I used 2 players to play the file, mpg123 and mpv and both stop before the end, so if it loops it creates this effect, normal. Maybe switching to cbr (constant bitrate) can work around this, or change the encoder. Actually I don't have any player which can play this file to the end, they all stop before it.

    7 hours ago, mer-curious said:

    6. There is a bug in a part of the game in which we leave a stage straight to a story demo. Two tracks are played at the same time, which wasn't supposed to happen. The only track that should play in this part is file SND01_00048.mp3 (sound command 37h). Here's a save state right before you enter this cutscene:

    https://drive.google.com/file/d/1FjLgfoq8zvRf1AEWQbAUrfzq3KjBT2_L/view?usp=sharing

    Ok, bad news for you, you are by far the one who uses this function the most, and since you didn't test kof2001 thoroughly and neither did I, there remained bugs !
    For the little story, like all the last games, its z80 rom is based on garou, but it's an evolution of it, and the 2nd array of offsets was wrong here, I was a little rusty for this stuff so it took me a little while to find something which works everywhere (assuming we don't forget another game). It seems to work now.

    7 hours ago, mer-curious said:

    7. I don't know if this is related to the "Sound associations" settings, but if you click on "Manage associations" and then return to the game the colors will be pinkish, take a look:

    https://drive.google.com/file/d/1Db0Den6qU0ttR_Qt2CJEy2lywwXinadh/view?usp=sharing

    I don't have this, but I don't even have the neogeo music playing for this game, probably a neogeo bios setting. I tried ingame, but it doesn't change colors...

    Might be windows specific ?

    Yep, windows specific, I can reproduce it in wine... ! Oh well... switch to linux ? ;-)

    7 hours ago, mer-curious said:

    I guess these are all the issues I've found so far with the sound commands feature. I'm sharing my games.cfg file together with my KOF2001 soundtrack pack to help you reproduce the problems:

    https://drive.google.com/file/d/15V40Yph1yVsw1crovgIDJ9oKIfdSV84O/view?usp=sharing

    Thank you so much in advance for your help.

     

    PS: In the following post I will report a slowdown issue which I'm experiencing in KOF2001.

     

    • Like 1
  19. Test this is just to test a specific command, displayed in the field just before. Actually it's equivalent to entering a number in the field and just press return, it's tested as a sound command. Useful mainly when trying to understand how the song codes work for a game, it's normally useless for the games you are using, except when you already know a sound effect number and want to play it directly without browsing the others in the top menu.

    For the rest, I'll look later, I'm having a little break for now... !

    • Like 1
  20. Yeah, annoyed by a nasty thread bug in 0.93.2 for those who installed without a previous config file, so here is a fix.
    Not a few changes except that, 0.93.2 was only 2 days ago, but :

     - curl should now know how to download a parent set when needed (not totally sure it will work all the time, but it should).
     - A fix for the sound dialog which got some strange strings inside when you tried to change the driver

     - And all the rest are memory leak fixes, most of them harmless, but the file selector leaked quite a lot of memory and I hadn't noticed, since it's not used very much it was not a very big deal, but still it's better like that. these fixes are thanks to https://sourceforge.net/projects/memwatch/

    Notice that this annoying bug could be worked around by simply quitting raine once using the quit command so that it writes its config file, so it was not so disastrous, but here is the fix anyway... !

    http://raine.1emulation.com/download/latest.html

    • Like 2
  21. 38 minutes ago, mer-curious said:

    Hey Tux! I suppose the fix you have provided to FaustXXX's bug report should also deal with Neville's freeze as well, right? In case it doesn't, I just found out that you don't have to go full-screen nor play with the Video settings to trigger the issue. Just try to load a game after populating the game list and this should make the program freeze.

    A video of it happening: https://drive.google.com/file/d/1YC16cm52TfcY-hG4Rck7TbHGzmbOuPEt/view?usp=sharing

    I suppose it should definitely have to do with the lack of a config file in the config folder because this is what we all had in common in our setups. So in the end it wasn't a CPU overclock or video drivers issue.

    Anyway, thank you again for your super fast work!

    Yep the cause was indeed the lack of config file. It's something new with sdl2, there is a slight modification in the sound thread, it starts faster now, so here it asks for some data before the sound chip had time to initialize and it creates a crash. For some reason it happened only if there was no config.

    And since I almost never remove my config file, it took me quite some time to notice, even when I reboot in windows to test a new version it keeps its config file.

    The fix is just moving a call a few lines down, very easy to fix, but quite hard to find indeed.

    Anyway those who kept their config files never experienced this, and in linux the config file is in a separate directory, which makes it even less likely to run a new version without any config file if raine was used at least once before...

    • Like 1
  22. 1 hour ago, faustXXX said:

    Last version (x64) still has same problem.

    Video (if i could publish link, remove space)

    www.t2e.pl / www / files / 2022-02-23 / Raine v0.93.2 2022-02-23 20-28-34-t2e.pl.mp4

    I don't know why you didn't post the link directly, there were spaces everywhere in your link, anyway...

    You finally found something !

    1) loading gauntlets 1st fails because it needs the parent set which is gauntlet and it hasn't got it.

    2) after loading gauntlet you get a crash because the default sound init is bad for some reason, if you want to fix it on your side, either open the sound options dialog once, or just quit so that the config file is saved, the crash won't come back !

    And yeah it should know about loading directly the parent set from internet archive, but anyway...

    I'll see about fixing this sound issue, even if it happens only when you run without any config file !

    edit : fixed both, it now knows how to load gauntlets even if the rom directory is empty, and it doesn't crash on the 1st sound init like it did here. But I won't make another binary today, sorry, too much coding with all the allocations fixes today, I'll see that tomorrow, thanks for the report anyway !

    • Like 1
  23. 1 hour ago, Neville said:

    As I said, no overclocking: https://www.pccomponentes.com/procesador-amd-ryzen-7-2700x-37-ghz

    But I may have been abusing my CPU lately, re-compressing video while I go through my usual stuff, including fiddling with Raine.

    That's the advantage of non overclocked cpus : you can "abuse" them continually, there will never be any crash ! There are serious stability tests for an overclocked cpu, like running for a whole night memtest86 in loop, and it should not get a single error after a whole night (or even better, a whole day).

    Anyway...

  24. 53 minutes ago, Neville said:

    No, my CPU is not overclocked. The figure could be wrong, though, it's the one Windows gave me.

    Well I actually have the same cpu on my main machine. From cpupower frequency-info :

    hardware limits: 1.55 GHz - 3.20 GHz
      available frequency steps:  3.20 GHz, 2.80 GHz, 1.55 GHz

    It's a good cpu, easy to overclock, but which can also become very easily unstable, I had a quite a few problems when I first got it.

    So 3.7 GHz is very far from its normal frequency, even with some small boost, and freezes as you describe are quite common for an overclocked cpu over its limit.
    With its 8 cores + 8 threads, it makes 16 virtual cpus, it's a beast to compile things, even at its vanilla settings, and it's much better to have something always stable.

    Anyway you say it isn't well maybe, but I can't reproduce this, so until someone can reproduce that reliably, I'll just forget this...

     

×
×
  • Create New...