Breakthrough for the ao486 core announced - Cache

FPGAzumSpass
Core Developer
Posts: 382
Joined: Sat May 23, 2020 12:55 pm
Has thanked: 38 times
Been thanked: 407 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by FPGAzumSpass »

D000-DFFF should be fine.

There seem to be some programs using memory in the C800-CFFF area without possibity to turn it off, so i'm not sure how this will be handled finally.
User avatar
tontonkaloun
Posts: 354
Joined: Sun May 24, 2020 7:38 pm
Has thanked: 152 times
Been thanked: 52 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by tontonkaloun »

DEVICE=C:\QEMM\QEMM386.SYS RAM ARAM=B000-B7FF ARAM=D000-DFFF ST:F R:2
User avatar
tontonkaloun
Posts: 354
Joined: Sun May 24, 2020 7:38 pm
Has thanked: 152 times
Been thanked: 52 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by tontonkaloun »

under msdos, there are still sound problems:
Doom no music
Day of the tentacle, some sound effects are missing ...
Until the first cached version, there was no such problem
User avatar
NML32
Posts: 296
Joined: Sun May 24, 2020 6:57 pm
Has thanked: 280 times
Been thanked: 50 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NML32 »

With Cache 6 & 7 Windows 98 locks up during boot. Windows 95 starts fine.
Volfied Locks up with Cache 6 & 7 on earlier builds Volfied would start if L2 cache was disabled.
softtest9
Posts: 158
Joined: Thu May 28, 2020 7:13 pm
Has thanked: 3 times
Been thanked: 21 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by softtest9 »

NML32 wrote: Sun Jul 12, 2020 7:40 pm With Cache 6 & 7 Windows 98 locks up during boot.
I fired up the Windows 98 SE installer earlier today and it complains about the lack of an FPU. So this is likely not a bug, but rather missing functionality in the CPU.
User avatar
SuperBabyHix
Posts: 152
Joined: Sun May 24, 2020 8:26 pm
Has thanked: 44 times
Been thanked: 34 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by SuperBabyHix »

I haven't tried installing Windows 9x on the core, but it might work if you run the installer with the /nm argument (bypasses CPU detection).

A list of install arguments can be found here: http://www.dewassoc.com/support/useful/setup_switch.htm

Edit: I should clarify the installer might work, not necessarily Windows 98 itself.
User avatar
NML32
Posts: 296
Joined: Sun May 24, 2020 6:57 pm
Has thanked: 280 times
Been thanked: 50 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NML32 »

SuperBabyHix wrote: Sun Jul 12, 2020 9:25 pm I haven't tried installing Windows 9x on the core, but it might work if you run the installer with the /nm argument (bypasses CPU detection).

A list of install arguments can be found here: http://www.dewassoc.com/support/useful/setup_switch.htm

Edit: I should clarify the installer might work, not necessarily Windows 98 itself.
I have Windows 95 and 98 installed and running. I was just reporting since Cache6&7 Windows 98 locks up while booting into the OS.
User avatar
SuperBabyHix
Posts: 152
Joined: Sun May 24, 2020 8:26 pm
Has thanked: 44 times
Been thanked: 34 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by SuperBabyHix »

Sorry for the confusion, I was replying to softtest9, who said the installer complained about the lack of an FPU.
User avatar
NML32
Posts: 296
Joined: Sun May 24, 2020 6:57 pm
Has thanked: 280 times
Been thanked: 50 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NML32 »

SuperBabyHix wrote: Sun Jul 12, 2020 10:03 pm Sorry for the confusion, I was replying to softtest9, who said the installer complained about the lack of an FPU.
I think you are correct about installing with a switch. I remember doing something like that when I first installed Win98.
I didn't have any luck installing Win98 SE.
@softtest9 I'm running Win98 4.10.1998
softtest9
Posts: 158
Joined: Thu May 28, 2020 7:13 pm
Has thanked: 3 times
Been thanked: 21 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by softtest9 »

I am aware that there are ways to bypass the hardware check, but I figured it was put there for a reason and haven't messed with it much. I also have a preinstalled HDD image of Windows 98 SE which has always had trouble booting in ao486. I admit that I haven't tried earlier releases of Windows 98 though. Did Win98 4.10.1998 work in ao486 before the cache changes?

Windows 95 does work fine on my end too.
User avatar
NML32
Posts: 296
Joined: Sun May 24, 2020 6:57 pm
Has thanked: 280 times
Been thanked: 50 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NML32 »

softtest9 wrote: Sun Jul 12, 2020 10:55 pm I am aware that there are ways to bypass the hardware check, but I figured it was put there for a reason and haven't messed with it much. I also have a preinstalled HDD image of Windows 98 SE which has always had trouble booting in ao486. I admit that I haven't tried earlier releases of Windows 98 though. Did Win98 4.10.1998 work in ao486 before the cache changes?

Windows 95 does work fine on my end too.
Yes, and it works with one of the earlier cache builds that has the L2 Cache on/off in the OSD, and as far as I can tell works fine with the L2 cache enabled. Win98 fails with the Cache 6 and 7 core builds.
softtest9
Posts: 158
Joined: Thu May 28, 2020 7:13 pm
Has thanked: 3 times
Been thanked: 21 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by softtest9 »

Ah, that does sound like a bug/regression then.
FPGAzumSpass
Core Developer
Posts: 382
Joined: Sat May 23, 2020 12:55 pm
Has thanked: 38 times
Been thanked: 407 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by FPGAzumSpass »

It's maybe related to the more strict check for ROM access.
Before ROM could be written (which is kind of nonsense given the name), now those writes are ignored.

Maybe we need to do a testversion with OSD option for ROM access, so we can check if something is influenced by it.

The problem with these testversions is, that nobody really knows what they are build from.

Maybe BinaryBond can still tell which version is based on which GIT commit?
If not, please in future add the commit hash, so we can find out later when things break.
User avatar
Chris23235
Top Contributor
Posts: 867
Joined: Sun May 24, 2020 8:45 pm
Has thanked: 114 times
Been thanked: 180 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by Chris23235 »

There is some strange behaviour regarding to Joysticks in X-Wing and Tie-Fighter. On the latest official release upon launching the game the calibration screen is coming up.

It still works this way in the first Test-Build with Cache, in all versions from Test2 onwards the calibration screen isn't coming up at all and manually triggering the ingame calibration shows the stick isn't detected at all.

This is not a general Joystick issue, as Doom or Wing Commander have no problems detecting the Joystick in the Test-Builds.
petracus
Posts: 6
Joined: Thu Jun 11, 2020 7:20 am

Re: Breakthrough for the ao486 core announced - Cache

Unread post by petracus »

Duke Nukem works ok.

https://youtu.be/wQSmiWeQdtQ

Doom works with music when launched from w95.
Schyz
Posts: 63
Joined: Sun Jul 12, 2020 4:53 pm
Has thanked: 18 times
Been thanked: 13 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by Schyz »

petracus wrote: Mon Jul 13, 2020 2:18 pm Duke Nukem works ok.

https://youtu.be/wQSmiWeQdtQ

Doom works with music when launched from w95.
I thought it was just the first Duke Nukem, but it's Duke Nukem 3D! that game is not easy to move.
thorr
Top Contributor
Posts: 1142
Joined: Mon Jul 06, 2020 9:37 pm
Has thanked: 565 times
Been thanked: 260 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by thorr »

FPGAzumSpass wrote: Mon Jul 13, 2020 6:03 am It's maybe related to the more strict check for ROM access.
Before ROM could be written (which is kind of nonsense given the name), now those writes are ignored.
This may or may not help at all, but I know that on the Apple II, you can POKE or write to ROM addresses and they act as switches to turn things on and off. It doesn't change the ROM, but it does something when you write to it. I have no idea if that applies here.
User avatar
NML32
Posts: 296
Joined: Sun May 24, 2020 6:57 pm
Has thanked: 280 times
Been thanked: 50 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NML32 »

Win98 boots up with ao486_Cache12_93ec09c.rbf :D
Thank you!!
User avatar
tontonkaloun
Posts: 354
Joined: Sun May 24, 2020 7:38 pm
Has thanked: 152 times
Been thanked: 52 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by tontonkaloun »

Thank you for the job !!!! 😊


always crashes and no sound on some games under Dos, like doom ...
Xevian
Posts: 2
Joined: Sun Jul 12, 2020 12:19 pm

Re: Breakthrough for the ao486 core announced - Cache

Unread post by Xevian »

Looking good :D Have Win98SE running nicely, even have the MT32 Midi working mostly okay withe the Roland Seiral Midi driver!
User avatar
Chris23235
Top Contributor
Posts: 867
Joined: Sun May 24, 2020 8:45 pm
Has thanked: 114 times
Been thanked: 180 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by Chris23235 »

Schyz wrote: Mon Jul 13, 2020 2:59 pm
petracus wrote: Mon Jul 13, 2020 2:18 pm Duke Nukem works ok.

https://youtu.be/wQSmiWeQdtQ

Doom works with music when launched from w95.
I thought it was just the first Duke Nukem, but it's Duke Nukem 3D! that game is not easy to move.
Dark Forces works fine as well and I always found it the most demanding of the 2.5D FPS.
User avatar
NightShadowPT
Posts: 208
Joined: Mon May 25, 2020 9:56 am
Has thanked: 5 times
Been thanked: 9 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NightShadowPT »

FPGAzumSpass just published the latest result of his optimizations of the core... absolutely unbelievable

Ec_hRNGWsAQqtsE.png
Ec_hRNGWsAQqtsE.png (68.39 KiB) Viewed 9416 times

Doom is now almost 4x faster than the current official core. :shock:
PikWik
Posts: 259
Joined: Sat May 30, 2020 7:00 pm
Has thanked: 194 times
Been thanked: 72 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by PikWik »

GV8OTtP.gif
GV8OTtP.gif (5.16 MiB) Viewed 9415 times
FPGAzumSpass
Core Developer
Posts: 382
Joined: Sat May 23, 2020 12:55 pm
Has thanked: 38 times
Been thanked: 407 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by FPGAzumSpass »

I have to thank Sorgelig that we can even have these improvements.

He is currently spending more time then me on this core and while his commits might not be directly visible in those graphs, his changes are very important.

E.g: The last change would not have been possible without prior work and support from him.
User avatar
RealLarry
Top Contributor
Posts: 776
Joined: Mon May 25, 2020 4:04 am
Location: San Junipero/DE/Earth
Has thanked: 93 times
Been thanked: 335 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by RealLarry »

Astonishing, unbelievable and fascinating. Haven't tested this optimized core yet, but followed this thread from beginning and am glad of this co-op'ed work of you. And that's why I'm a patreon supporter.
Contributor of tty2oled, author of tty2tft, tty2rpi and update_tty2xxx
User avatar
NightShadowPT
Posts: 208
Joined: Mon May 25, 2020 9:56 am
Has thanked: 5 times
Been thanked: 9 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NightShadowPT »

FPGAzumSpass wrote: Thu Jul 16, 2020 5:22 am I have to thank Sorgelig that we can even have these improvements.

He is currently spending more time then me on this core and while his commits might not be directly visible in those graphs, his changes are very important.

E.g: The last change would not have been possible without prior work and support from him.
Sorgelig is the unsung hero in many of these cores, but the community is aware of the debt of gratitude we owe him.

Anyway, thank you both for making improvements to this core.
User avatar
NightShadowPT
Posts: 208
Joined: Mon May 25, 2020 9:56 am
Has thanked: 5 times
Been thanked: 9 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by NightShadowPT »

It was already reported here that these "beta" cores that use cache seem to have some problems with Doom having music.

I don't know if it helps somehow, but I have also tried the following and the problem is the same. On the official core I get both sound and music (sound blaster) while on the cache cores I get sound but no music.

- Doom
- Doom II
- Heretic
- Hexen

Whatever bug there is, seems to be transversal to all versions of the engine.

And yes, I am using the exact same versions of Autoexec.bat, Config.sys and the same virtual drive.
User avatar
tontonkaloun
Posts: 354
Joined: Sun May 24, 2020 7:38 pm
Has thanked: 152 times
Been thanked: 52 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by tontonkaloun »

NightShadowPT wrote: Thu Jul 16, 2020 6:28 am It was already reported here that these "beta" cores that use cache seem to have some problems with Doom having music.

I don't know if it helps somehow, but I have also tried the following and the problem is the same. On the official core I get both sound and music (sound blaster) while on the cache cores I get sound but no music.

- Doom
- Doom II
- Heretic
- Hexen

Whatever bug there is, seems to be transversal to all versions of the engine.

And yes, I am using the exact same versions of Autoexec.bat, Config.sys and the same virtual drive.
It's the same thing for me !
No music under dos
music with win95

sound problem with other games like duke3d , Day of the Tentacle ...
User avatar
Jayde Six
Posts: 26
Joined: Mon May 25, 2020 7:05 am
Has thanked: 2 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by Jayde Six »

To Add to the info I've got sound issues with Dark Forces and Warcraft, where the audio starts getting really choppy/distorted after like 10 mins or so.

I checked and Warcraft did it with the release core too so some sound issues (outside of the id games music issue) might not be related to these betas but weren't noticed yet because the games were unplayable before.
User avatar
Chris23235
Top Contributor
Posts: 867
Joined: Sun May 24, 2020 8:45 pm
Has thanked: 114 times
Been thanked: 180 times

Re: Breakthrough for the ao486 core announced - Cache

Unread post by Chris23235 »

Didn't tried Warcraft, but so far I had no issues with Dark Forces, I played an hour or so.
Post Reply