syphus |
Posted: 2006-12-18 15:51:27 |
|
|
Expert at chit-chat Posts: 122
Member since: 2006-12-14 19:47:57 |
I did on 3.9 - same thing, as far as I can tell. But _Monk_ in #amigascne told me he has it working in WB3.1 in WinUAE with JIT disabled, cycle-exact disabled, Picasso96 emulated... http://syphus.untergrund.net |
exobuzz |
Posted: 2006-12-18 15:52:55 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
It is possible that there is some memory leak/corruption in the program which only shows up in some cases. When I was having the low memory crashes, it was during the gui_render_tracked() routine. When I disabled that, the crashes stopped. Increasing the amount of free chip ram seemed to stop the crashes for me. |
syphus |
Posted: 2006-12-18 16:16:03 |
|
|
Expert at chit-chat Posts: 122
Member since: 2006-12-14 19:47:57 |
(ps - Now working in Wb3.1 in WinUAE, with FPU/JIT/no cycle-exact, until I switch on Wavemeters and it crashes:P But other than that, fine with a bit of the usual WinUAE lag...) http://syphus.untergrund.net |
syphus |
Posted: 2006-12-19 16:56:11 |
|
|
Expert at chit-chat Posts: 122
Member since: 2006-12-14 19:47:57 |
Update - with 1,8xx,xxx chipmem, I can get Hively running reasonably stably on my hardware. I get some display problems when the polyphony rises above six or so channels, with complex/filterry instruments and after a while the screen starts to flash when autoscrolling. Pattern-scrolling has much less lag than in WinUAE but after some of the aforementioned glitchy autoscrolling and the odd temporary lockup, some of the buttons stop working. Play/stop/InsEdit still work, and keyboard controls, but Load/Save track, Prefs etc don't.
You know I'm no expert, but I suspect having an option to change full patternscroll to simply a cursor scroll might help (or mightn'd it? Maybe turning it off altogether...). http://syphus.untergrund.net |
exobuzz |
Posted: 2006-12-21 03:24:29 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
Hively 68k with the v1.1 changes is available. See the readme in the archive for what is changed from os4 version etc.
http://www.exotica.org.uk/~buzz/ht/ |
exobuzz |
Posted: 2006-12-21 07:33:17 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
I found the bug! (I'm pretty sure). I am unable to crash my latest version. It was a RectFill trashing memory. I'll upload you a test version to check in my ht directory. |
Xeron |
Posted: 2006-12-21 10:05:42 |
|
|
Supreme Being Posts: 448
Member since: 2006-12-13 11:34:00 |
@BuZz
Oh. Cool. Sorry I didn't get around to summarising the changes; I was going to do it today (i'm not working today). I promise i'll do it for future hively versions. AHX forever! |
exobuzz |
Posted: 2006-12-21 13:12:08 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
Well, I just did a diff against v1.0, and read though what had changed :) Wasn't much work. Thanks for the update |
syphus |
Posted: 2006-12-21 14:06:12 |
|
|
Expert at chit-chat Posts: 122
Member since: 2006-12-14 19:47:57 |
Fucking HELL - thank you exobuzz! I'm getting no crashes now, and I'm severely stress-testing it too;) http://syphus.untergrund.net |
spotUP |
Posted: 2006-12-21 19:26:42 |
|
|
Moderator Posts: 197
Member since: 2006-12-15 18:06:32 |
allright! cool news! i guess i'll start working away on that 4 colour skin then! :)
http://www.uprough.net |
spotUP |
Posted: 2007-01-03 18:07:02 |
|
|
Moderator Posts: 197
Member since: 2006-12-15 18:06:32 |
Hmmm a feature for the 68k version, that would make it useable for a lot more 68k users, would be a switch to disable patternscroll. A guy with a 040 reported that hively was unuseable for him, as soon as the patterns started to scroll.
http://www.uprough.net |
Karlos |
Posted: 2007-01-06 19:25:40 |
|
|
Just looking around Posts: 3
Member since: 2007-01-06 13:14:22 |
What sort of performance are we looking at with a basic 68040/25MHz for the 68K version of Hively? Try pie, try. |
syphus |
Posted: 2007-01-06 19:39:03 |
|
|
Expert at chit-chat Posts: 122
Member since: 2006-12-14 19:47:57 |
I don't think anyone's tried that yet, or at least nobody's reported back. As a guide, my 060/50mhz has the *slightest* lag on cursor-movement, but it's hardly noticeable. If I use the heavy skin, performance suffers significantly. http://syphus.untergrund.net |
pulp_pure |
Posted: 2007-05-31 18:57:17 |
|
|
Just looking around Posts: 4
Member since: 2007-05-31 18:37:44 |
Hively 1.3 hangs when I try to load it on my A1200 + Blizzard IV + OS3.9 (I don't have a gfxcard, nor FPU) : I get a guru : "ht-020 Défaillance du logiciel (erreur n°8000000B). Attendez que toute activité du disque cesse.
When I try with snoopdos I can see that it fails when trying to load ht.prefs which doesn't exist at first. So I ran HT from Winuae (it works although I have almost the same workbench as on the a1200) and I copied the ht.prefs to my a1200 but it still hangs... so I tried snoopdos with all its functions enabled and in the long list of its log the only fail I can notice is
ChangeDir WB:W1/Musique/HivelyTracker
#COPY_DIR 1E01A21E
I've found the following info on amiga.org forum : 8000000B is a Line F emulation error that is when you try to execute an FPU command without one being present, or a problem with 68040/060 library.
Did you set your compiler right for the 020 nofpu version ? AGA killed the scene ! |
exobuzz |
Posted: 2007-06-01 14:43:45 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
I think there are problems yes. I'll try and make you another build with some different options - sorry about this. Give me a couple of days (its my birthday tomorrow but im going out in a sec) |
pulp_pure |
Posted: 2007-06-01 21:35:54 |
|
|
Just looking around Posts: 4
Member since: 2007-05-31 18:37:44 |
Firstly, thanks a lot for you reply ! Happy birthday :) Party for four days if you like, I fully respect you for all you have already achieved :)
In fact I've just checked the program because it seems quite cool. Maybe it will be too slow for a 030@50 amiga. Digibooster also lags on the 030. But octamed works fine, sadly it is no longer supported. I know it takes a lot of time to work on such a project. I would be glad to help as much as I can.
To go further with my bugtesting, yoday I've tried with winuae (I chose the 020/no FPU) and I have the same guru as on the A1200. Now I know it's not because of my lack of graphic card or because of some patch (anyway I use as little as possible). AGA killed the scene ! |
exobuzz |
Posted: 2007-06-05 15:18:01 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
I think it might be too slow on an 030. Perhaps you can play 4 track ht tunes at a low mixing rate though. Did the commandline replayers work at all on your setup ? |
exobuzz |
Posted: 2007-06-06 16:53:59 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
right.. the guigfx/render libs i ship with are built for fpu. Please try these:
http://igame.amigahellas.gr/files/guigfx_render_nofpu.lha
Those are ones i found i needed to use on my a1200/030 when running another piece of software. |
pulp_pure |
Posted: 2007-06-10 17:48:43 |
|
|
Just looking around Posts: 4
Member since: 2007-05-31 18:37:44 |
I'm sorry but even with those _nofpu libs, HT still hangs with the same error. The problem is with HT, not the libs. AGA killed the scene ! |
exobuzz |
Posted: 2007-06-11 01:57:20 |
|
|
Moderator Posts: 56
Member since: 2006-12-14 18:21:01 |
I will do more testing. I don't see why ht should use any fpu libraries. as far as im aware linking with libnix will not use the fpu libraries unless asked to, but perhaps I am mistaken. I will test on my own non fpu machine as I said before and get back to you. sorry for the delay,I thought this might have been the problem. Thanks for testing. |