• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Softwindows 2 68k

Grex

Member
Alright, I guess there will be no more replies to this thread, as it's been almost a month since the last posting. Anyway, thanks to anyone who participated!

There is only one more thing I can think of: I can look up the credits for Softwindows-1-68k and try to find contact to the developers, using google. Probably they can tell for sure, if there ever was a 68k-Version of SoftWindows-2, and if, where I could get it. However I know from various experiences how painful it can be to search a software-author via google, especially if he/she has a rather common name.

xx(

 

Grex

Member
Hi everyone around! I now have some time again to continue my quest for Softwindows-2-68k (if it exists). As I already said in above post, I want to try to contact one of the developers, but so far I couldn't get a single name! I tried "Get Info" in finder, I read the document that is in the softwindows-drawer, I checked the "about Softwindows"-menu-entry when running softwindows, NOTHING, no credits, not a single name, it all just says "Insignia Solutions".

So, please, does anyone of you know a name of any of the Softwindows 1/2-programmers? Probably there are full credits in the manual-book, which I don't own, because I downloaded my SoftWindows-1 from internet.

:)

 

Grex

Member
Hi everyone!

I just wanted to post a small update. I now managed to contact one of the former developers of SoftWindows. He explained that there has indeed been a 68k-version of SoftWindows 2.0 for Macintosh as well as for Atari. However none of them have ever been released to public. He is not too optimistic about the chances for successfully retrieving that software. But he is (hopefully) trying to find it during the next days.

That's it so far. At least I know now, the rumours were not completely bogus.

 

reukiodo

Well-known member
Any update from the developer?

IIRC, v2 added 486 emulation, which would help substantially with running more software.
 

Concorde1993

Well-known member
SoftWindows sucks anyway. I ran Version 1.0.2 on my IIfx - the 486 OrangePC card is far better in terms of performance and has a far less memory footprint than Insignia’s software.
 

reukiodo

Well-known member
Ah, I am wrong. SoftWindows v2 added 386 emulation, not 486. Even so, 386 opens up a lot more guest software than the 286 from v1.

Macintosh Garden only has the PPC version of SoftWindows v2. The whole point of this specific thread is for the 68k version of SoftWindows v2.
 

NJRoadfan

Well-known member
I don't think SoftWindows had a 68k release. It would have been VERY slow and one of the "killer apps" for PowerPC at the time was PC emulation since the CPU was much more powerful for emulating x86. I recall reading a feature article at the time about what Insignia developers went through at the time to get Windows 3.1 running in 386 Enhanced Mode. They said that it was like triple the number of x86 instructions just to start it from a DOS prompt vs. Standard Mode.

The reality is, most Win16 software didn't really need 386 Enhanced Mode (outside of the obvious Win32s or rarer Win386 runtime from Watcom). Developers either didn't test their code in Standard Mode, or were relying on VxD drivers that were only available in 386Enh. The software itself was mostly strictly written for 286 op codes and that CPU's protected mode of operation.
 

s_pupp

Well-known member
I had a SoftWindows version for 68k, but do not recall which version. It ran, but was so slow on my Performa 475 that it was completely unusable (68LC040 cpu).
 

CC_333

Well-known member
The reality is, most Win16 software didn't really need 386 Enhanced Mode ... The software itself was mostly strictly written for 286 op codes and that CPU's protected mode of operation.
Even so, the 386's implementation of protected mode was more reliable and easier to set up than the 286's was, I believe.

or rarer Win386 runtime from Watcom
Where/what is that?!

c
 

Snial

Well-known member
Even so, the 386's implementation of protected mode was more reliable and easier to set up than the 286's was, I believe.
The i386 supported 4KB paged virtual memory in a flat 32-bit virtual and physical address space and it was possible for the 386 to switch back to real mode from protected mode. This made it possible for its memory management to be largely compatible with contemporary CPUs. The i286 on the other hand was a 16-bit CPU that supported segmented virtual memory of up to 1GB within a 24MB physical address space and it couldn't switch back to real mode by itself.

So, if you're unfamiliar with i286 segmentation, here's a short explanation:

The 8086 was a 16-bit CPU with a 1MB address range and it achieved this using a technique called segmentation. Memory pointers could only address up to 64KB at a time, but whenever you tried to access memory it would add one of 4 different 16-bit segment registers multiplied by 16. Hence a complete address was always: aSegmentRegister*16+aPointer and so the maximum address was 65535*16+65535 = 1114095, which is nearly 1MB+64KB, but that last 64KB would simply wrap around back to the beginning of memory. There were 4 segment registers, one for code (CS), one for data (DS), one for the stack (SS) and an Extra one for copying memory between arbitrary memory regions (ES).

The 68000 by contrast simply supported 32-bit pointers that could in theory address 4GB, but the address signals limited it to 24-bits giving 16MB. Very much simpler!

The 80286 was its true successor and introduced protected mode. The 4 segment registers were kept, but instead of them being multiplied by 16 they indexed a long table in RAM (called the descriptor table) which contained a full 24-bit base pointer and a 16-bit length value. So, now memory addresses were computed as: DescriptorTable[aSegmentRegister].base+aPointer. For 'speed', 6 of the descriptor table entries were cached in the processor too. There were actually at least 3 description tables: the global descriptor table, the local descriptor table and the interrupt descriptor table.

So, this was an awful way to manage virtual memory, because well firstly, the segmentation was now incompatible with the 8086 and secondly, when virtual memory segments of were placed in physical memory, they could all be completely different sizes, so when it came to swapping out segments and swapping in new ones from disk, it could easily involve lots of shuffling around of existing segments or swapping out multiple segments or leaving gaps in physical memory where they didn't align.

In other words Intel turned a poor scheme into a dreadful scheme. It was so bad, that it's really hard to find an Intel 80286 OS that used VM properly. Even Windows 3.1 (in standard mode) didn't manage it, they just treated protected mode as a means of accessing 16MB of physical RAM.

The 80386 fixed this. It supported the segments, but all programmers did was set them all to 0 and then activated page mode. The segmentation just became a waste of transistors, but at least it was simpler to access.

Because 386 paging is pretty similar to 68030 paging, 386 page mode emulation is easier than 286 emulation.
 

CC_333

Well-known member
@Snial It seems I made the understatement of the day, didn't I? Not only is the 386's protected mode better, it's vastly superior!

No wonder nobody bothered to support 286 protected mode. It's a mess!

c
 

dcr

Well-known member
I had a SoftWindows version for 68k, but do not recall which version.
Same here. I had it on my PowerBook 180c. But, I don't know what version of SoftWindows it was.
But, I did manage to install Windows95 in it. Was painfully slow though. I'm thinking I was maybe one of the first Mac users to install Windows95 on a Mac.
 

dcr

Well-known member
Same here. I had it on my PowerBook 180c. But, I don't know what version of SoftWindows it was.
Just checked and it was SoftWindows 1.0.2 on the PowerBook.
I also had SoftWindows 2 but that was for my PowerMac.
 

rplacd

Well-known member
Same here. I had it on my PowerBook 180c. But, I don't know what version of SoftWindows it was.
But, I did manage to install Windows95 in it. Was painfully slow though. I'm thinking I was maybe one of the first Mac users to install Windows95 on a Mac.
Even Windows 3.11 was painfully slow... it was on my Classic II, though. I can imagine that even a SCSI-attached emulation card would do better, if one existed. (If I had the skills to do this, I'd have it expose a virtual hard disk where you might access store emulator files like the virtual hard disk, removable disk images, etc, but have video + input use a sideband...)
 
Top