VOGONS


Would you be interested in an x86-based alternative to the Raspberry Pi, optimized for retro gaming?

Topic actions

  • This topic is locked. You cannot reply or edit posts.

Reply 40 of 201, by Doornkaat

User metadata
Rank l33t
Rank
l33t

Some variant of this thread pops up every now and then:
The OP has an (in principle) good idea for a new production piece of hardware to tackle a (preceived) scarcity on the market or to create a convenient one-size-fits-all solution for certain applications, and would like to ask wether it is worth his time to pursue the idea.
There is generally a decent amount of interest. Wether the idea becomes reality seems to depend on the OP's ability and dedication toward the project though.
This is why I'd say if you want this thing to happen stop asking for permission here and start inquiring with your people who have that expertise about how you can do this and how much of an investment you'd be looking at. It's time to nut up or shut up! 😉
Personally I'm doubtful that currently it is feasibly to produce anything like what you're envisioning but you make it sound like you have people that will be able to give you a much better assessment than I can.

If you want it here's some more feedback on the idea itself:
-Sony and Nintendo went for emulator boxes when they released their mini retro consoles. Both have the engineering capabilities and all data required to recreate their vintage consoles in hardware (plus add some ease of use features) and still decided to go for emulation.
-A great majority of games will run good enough even on current hardware and OS. Sometimes there's a bit of tweaking or a workaround required (and this forum is originally dedicated to finding/creating those) but a lot of stuff just works. Commercial re-releases of old PC games mostly rely on those solutions instead of the user buying dedicated hardware to run a game.
->This is why I believe that to appeal to a broader audience you'd need features that aren't avaliable on modern systems/OS and can not be fatihfully emulated. You need to identify the relevant shortcomings of emulation and make a finished product that avoids those shortcomings.

-Likely you'll also need some sort of software bundle to (optionally?) go with the hardware. Most of us already have the hardware and the games. If you want to appeal to a broader audience it only makes sense to be able to conveniently deliver the OS and at least some of the all time favourites.

-Here on Vogons/Marvin you're inquiring/arguing with a very niche market that is interested in running software on real hardware because of certain nuances and special features. Here on Marvin we are always feature creeping with our builds. We want those extra rendering modes even though we probably aren't really going to use them. If you ask here people will tell you they want nearly all the possible features implemented or at least expansion ports to use our own cards to add specific features. This forum is probably a good place to ask for the make-or-break features but if you only want to release a basic x86-machine with no special features I think most of us are already set.

Reply 41 of 201, by megatron-uk

User metadata
Rank Oldbie
Rank
Oldbie

@WDStudios, you propose a custom, retro gaming SBC - yet you don't understand:

1. Why the need exists for ISA sockets. Hint: Try getting a Dos game to play Adlib music on an AC97 codec, or load a custom GUS patch bank into the same.

2. Why it isn't possible to get an accurate gaming experience with a single system. Have you ever written any code for Dos? Have you thought about how many possibilities there are for a game to generate a timing loop?

3. Why it isn't possible to just mash a whole load of different silicon onto a single die and then just call it a day.

The closest you can get to your unicorn solution is emulation, and for many of us, despite how accomplished it is these days, it simply isn't good enough.

My collection database and technical wiki:
https://www.target-earth.net

Reply 42 of 201, by Oetker

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2021-06-21, 09:15:

Some variant of this thread pops up every now and then:

Yeah exactly, "I'm going to build a new 3dfx card / socket xyz motherboard / etc", while the projects that actually go somewhere are often one or more people working in quiet and then announcing the project once it's in the prototype stage.

Reply 43 of 201, by Socket3

User metadata
Rank Oldbie
Rank
Oldbie
WDStudios wrote on 2021-06-20, 21:33:

Why would you need either of those? Discrete sound cards in general were obsolete by the time either of those came out, with sound generally being integrated onto motherboards, and the SB400 southbridge includes 8-channel AC '97 audio.

If you need to ask this question you don't belong here. Early on board sound cards generally sound bad, lack dos drivers and features some of us are looking for. Also, the majority of user on vogons/marvin are COLLECTORS first, gamers second. Sometimes we'll change the sound card or video card in a retro pc just to experience the hardware.

WDStudios wrote on 2021-06-20, 21:33:

I've never seen a game that required Glide. The closest I've seen is Diablo II, which supports DirectX but ran better with Glide.

Your lack of wisdom is only superseded by your arrogance. Close your mouth and open your mind. There are countless glide games, some of witch are Glide only (if you expect 3d acceleration that is) - Uprising, Tomb Raider 1, Descent 1 and 2, Carmageddon, Pandemonium and so on. And don't give me that crap about "that's only a few games", I'm not interested. What you're looking to play on this SOC could run fine on an OEM pentium 4 + geforce 4 MX that you can get off ebay for like 10$. It's not about what YOU want, it's about what the majority of us would buy - and I'm sure most users here would support the idea of having glide compatibility on such device and the ability to use real vintage hardware with it - witch means ISA, PCI and even AGP slots.

WDStudios wrote on 2021-06-20, 21:33:

That list is garbage. I can tell because the Windows section includes Moo2 and Simcity 2000, both of which I've played on modern hardware (well... hardware from 2014) with no issues, patches, or workarounds whatsoever. The complaint about Simcity 200o is especially amusing because it boils down to "program works exactly as intended and I don't like that".

The list is good - incomplete even. You're being very hard-headed and unpleasant. Are you here to discuss and share ideas or to play the smart-ass and contradict people? You have some knowledge about computers, but not enough to warrant your arrogant wording.

Reply 44 of 201, by WDStudios

User metadata
Rank Member
Rank
Member
Error 0x7CF wrote on 2021-06-21, 07:31:

I would absolutely adore the product you describe, but I do not believe AMD would ever do it. There's just not enough market for it. For all the use cases but retro gaming, there are (arguably better) solutions already. For instance, I could see benefiting from something like this would be old industrial control equipment that still runs DOS, and that is already quite a well-served market by existing hardware, as well as a shrinking market as people try to get away from legacy crust. Things like that don't need the graphics and will be perfectly happy chunking along with a 486-based PC104 SBC with practically no graphics whatsoever for the next unforeseeable number of years.

That is correct; a product like this would primarily find use in applications that demand massively parallel floating-point operations, but where spending $10,000 on a computer with Geforce 3080s in Quad SLI would be severe overkill. Something along the lines of "teaching kids in third-world countries how to use CAD software", for example. Different embedded systems are going to find different niches depending on their capabilities.

Error 0x7CF wrote on 2021-06-21, 07:31:

Suppose AMD makes the chip you desire. Do they sell it on a board with an OS? They can't sell it with Linux since the 9800 is too old for drivers. They can't sell it with Windows

AMD (or VIA, or whoever) wouldn't be put in charge of designing a whole new board. I'd either go with a pre-existing board like the EPIA-P910 and design a complete system around that myself, or I'd commission one of these companies to make the SOC only and then have one of "my people" design the board. For the OS, we'd ideally acquire a Win98 volume license on the secondary market, but barring that, we could sell systems that dual-boot FreeDOS and ReactOS with a CD full of Win98 drivers and a note advising people on how to find their own copies of Win98.

However, I think a lot of people here are getting bogged down in the technical details of how such a thing would get made, particularly if I were to make it. That's really not the point of this thread. I already know people who I can discuss such things with. The point of this thread is to determine the level of market demand for such a product.

Doornkaat wrote on 2021-06-21, 09:15:

-A great majority of games will run good enough even on current hardware and OS. Sometimes there's a bit of tweaking or a workaround required (and this forum is originally dedicated to finding/creating those) but a lot of stuff just works. Commercial re-releases of old PC games mostly rely on those solutions instead of the user buying dedicated hardware to run a game.
->This is why I believe that to appeal to a broader audience you'd need features that aren't avaliable on modern systems/OS and can not be fatihfully emulated. You need to identify the relevant shortcomings of emulation and make a finished product that avoids those shortcomings.

Yeah, it seems like no matter the program, there's always a third-party hack or workaround somewhere to get them to work on new hardware and OSes. For the Win95 ports of C&C and Red Alert, there's the "cnc-ddraw" wrapper (and you have to run the installer in Win95 compatibility mode). If the Earthsiege II installer is giving you an "integer divide by 0" error message, there's a solution at http://www.sierrahelp.com/Patches-Updates/New … llers.html#ESXP. Quake II's OpenGL renderer is too dark on modern GeForce graphics cards? Pull down the console and type "gl_modulate 2". I just recently figured out how to get Blood II's Direct3D rendering mode working correctly. But finding these hacks and workarounds often takes hours, per game... and even then, they're sometimes a bit of a crapshoot. Updated or remastered versions are sometimes available from Steam or GOG, but many people don't like the idea of paying a second time for a game that they already have.

A big part of the draw of a product like this would be that almost everything would install and play correctly from the original disks with no third-party hacks, workarounds, command-line parameters, typing arcane jibberish into consoles etc... there would be some exceptions of course, like the Win95 ports of C&C and Red Alert. I'm pretty sure those were already broken by the time DirectX 9 came out, so getting them to work "out of the box" would mean sacrificing DirectX 9 (and possibly earlier) games, which is far too high a price to pay. But, you know... it would be something.

Doornkaat wrote on 2021-06-21, 09:15:

-Here on Vogons/Marvin you're inquiring/arguing with a very niche market that is interested in running software on real hardware because of certain nuances and special features. Here on Marvin we are always feature creeping with our builds. We want those extra rendering modes even though we probably aren't really going to use them. If you ask here people will tell you they want nearly all the possible features implemented or at least expansion ports to use our own cards to add specific features. This forum is probably a good place to ask for the make-or-break features but if you only want to release a basic x86-machine with no special features I think most of us are already set.

Yes, I'm definitely getting the impression that this isn't a "game not work, wat do?" forum. Or at least that doesn't seem to be the focus.

Doornkaat wrote on 2021-06-21, 09:15:

-Likely you'll also need some sort of software bundle to (optionally?) go with the hardware. Most of us already have the hardware and the games. If you want to appeal to a broader audience it only makes sense to be able to conveniently deliver the OS and at least some of the all time favourites.

Some games have already been released as freeware. Examples that I know of include the first three C&C games and most of the Earthsiege/Starsiege series. Those could be included.

megatron-uk wrote on 2021-06-21, 09:19:
@WDStudios, you propose a custom, retro gaming SBC - yet you don't understand: […]
Show full quote

@WDStudios, you propose a custom, retro gaming SBC - yet you don't understand:

1. Why the need exists for ISA sockets. Hint: Try getting a Dos game to play Adlib music on an AC97 codec, or load a custom GUS patch bank into the same.

2. Why it isn't possible to get an accurate gaming experience with a single system. Have you ever written any code for Dos? Have you thought about how many possibilities there are for a game to generate a timing loop?

3. Why it isn't possible to just mash a whole load of different silicon onto a single die and then just call it a day.

Again, you're getting bogged down in the technical details. I know people in real life who can worry about those things for me. And yes, there will always be compromises, in the design of literally everything that has ever been designed in the history of designing things. The fact that I'd make different compromises than you would, does not indicate a lack of understanding.

Oetker wrote on 2021-06-21, 10:12:

Yeah exactly, "I'm going to build a new 3dfx card / socket xyz motherboard / etc", while the projects that actually go somewhere are often one or more people working in quiet and then announcing the project once it's in the prototype stage.

This wasn't an announcement. It was an inquiry.

Socket3 wrote on 2021-06-21, 12:08:

you don't belong here....

Your lack of wisdom is only superseded by your arrogance. Close your mouth and open your mind...

You're being very hard-headed...

Someone here is being hard-headed and arrogant but it's not me. Reported.

Last edited by WDStudios on 2021-06-21, 19:43. Edited 1 time in total.

Since people like posting system specs:

LGA 2011
Core i7 Sandy Bridge @ 3.6 ghz
4 GB of RAM in quad-channel
Geforce GTX 780
1600 x 1200 monitor
Dual-booting WinXP Integral Edition and Win7 Pro 64-bit
-----
XP compatibility is the hill that I will die on.

Reply 45 of 201, by bZbZbZ

User metadata
Rank Member
Rank
Member
WDStudios wrote on 2021-06-21, 19:30:

However, I think a lot of people here are getting bogged down in the technical details of how such a thing would get made, particularly if I were to make it. That's really not the point of this thread. I already know people who I can discuss such things with. The point of this thread is to determine the level of market demand for such a product.

... I'm definitely getting the impression that this isn't a "game not work, wat do?" forum. Or at least that doesn't seem to be the focus.

I think you're starting to understand that most folks here are not interested in your 'product'. You've spent three pages arguing with everyone who gives you feedback. The only thing you've accomplished is you've revealed your lack of understanding of the technical side, as well as your willful misunderstanding of this user base and how other people's interests differ from yours.

Please go ahead and discuss this with your "people" instead. Your continued posts here aren't changing any minds.

Reply 46 of 201, by WDStudios

User metadata
Rank Member
Rank
Member
bZbZbZ wrote on 2021-06-21, 19:38:

I think you're starting to understand that most folks here are not interested in your 'product'. You've spent three pages arguing with everyone who gives you feedback.

That's because most of the people who provided feedback wanted to argue about the technical details of the implementation rather than whether or not they'd buy such a thing, even though literally zero technical details had been set in stone so there was nothing to argue about. That was their fault, not mine.

Since people like posting system specs:

LGA 2011
Core i7 Sandy Bridge @ 3.6 ghz
4 GB of RAM in quad-channel
Geforce GTX 780
1600 x 1200 monitor
Dual-booting WinXP Integral Edition and Win7 Pro 64-bit
-----
XP compatibility is the hill that I will die on.

Reply 48 of 201, by Doornkaat

User metadata
Rank l33t
Rank
l33t
WDStudios wrote on 2021-06-21, 19:30:

Yes, I'm definitely getting the impression that this isn't a "game not work, wat do?" forum. Or at least that doesn't seem to be the focus.

Marvin isn't but The Guide is. Marvin is aimed at users of actual vintage hardware while The Guide is about getting old software to run on modern systems.

This isn't what my post was about though. What it is about is that if you're convinced of your concept and think you can pull it off you're better off working on the device than telling a group of people (that you honestly do not really seem to understand) why you have a concept they should want and why they're stupid for not wanting it. I think our hobby is different from what you think it is. No problem, really! But if you want feedback from our interest group we're going to tell you what the concept is missing to really appeal to us. If you don't like that feedback you're wasting your time. Better just make the thing happen the way you want it then!
Though what you're describing leaves some (important) boxes unchecked if you can make the device at a reasonable price I'm sure a couple of members are going to buy one. Even if just to toy around with it.

Good luck with your project!

Reply 49 of 201, by Oetker

User metadata
Rank Oldbie
Rank
Oldbie
WDStudios wrote on 2021-06-21, 19:45:
bZbZbZ wrote on 2021-06-21, 19:38:

I think you're starting to understand that most folks here are not interested in your 'product'. You've spent three pages arguing with everyone who gives you feedback.

That's because most of the people who provided feedback wanted to argue about the technical details of the implementation rather than whether or not they'd buy such a thing, even though literally zero technical details had been set in stone so there was nothing to argue about. That was their fault, not mine.

Ok so basically you're only interested in everyone here telling you what a great idea you've got. Maybe start a Kickstarter page, you could even earn money with your idea!

Reply 50 of 201, by WDStudios

User metadata
Rank Member
Rank
Member
Doornkaat wrote on 2021-06-21, 20:13:

Marvin isn't but The Guide is. Marvin is aimed at users of actual vintage hardware while The Guide is about getting old software to run on modern systems.

I was linked directly to Marvin by someone on Facebook who did not indicate that there were sections of the forum other than Marvin. Furthermore, Vogons itself, when you are in one section, does not give any indication that other sections exist. So, at the time when I posted this, I had no idea that there were parts of Vogons other than Marvin. Should I ask a moderator to move this thread somewhere more appropriate?

Doornkaat wrote on 2021-06-21, 20:13:

you're better off working on the device than telling a group of people (that you honestly do not really seem to understand) why you have a concept they should want and why they're stupid for not wanting it.

I never told anyone that.

Oetker wrote on 2021-06-21, 20:17:

Ok so basically you're only interested in everyone here telling you what a great idea you've got.

Or in everyone telling me "no, most retro gamers would rather waste hours hunting down third-party hacks and workarounds for every single game in their collection than in paying $100 for something that just runs everything natively with no problems"

Or in everyone telling me "no, most retro gamers would rather spend thousands of dollars on vintage hardware from eBay than pay $100 for new hardware that plays all of the same games". Which is apparently what Marvin is all about.

Since people like posting system specs:

LGA 2011
Core i7 Sandy Bridge @ 3.6 ghz
4 GB of RAM in quad-channel
Geforce GTX 780
1600 x 1200 monitor
Dual-booting WinXP Integral Edition and Win7 Pro 64-bit
-----
XP compatibility is the hill that I will die on.

Reply 51 of 201, by Dominus

User metadata
Rank DOSBox Moderator
Rank
DOSBox Moderator

To all: please calm down. You are not agreeing but that's not a reason to start calling names etc....
———

Well, it's a forum and forums mostly have different sections. If you don't go to the root you never know...

Anyway, what you envision to be retro computing doesn't seem to allign what we consider retro. Did I see you mention FarCry and how it fares in Virtual Machines?
Retro is for many of us DOS, Windows 3.x, Windows 9x. And that is more complicated to get right as you've been told but don't seem to want to understand and more terrible, don't seem to have any clue on.

And how can you expect us to not be technically examining this idea? Many of us have spent a lot of time with the old tech and also are not out of the loop with new technology. Of course we want to get down to the detail after the initial pitch.

But you have to agree you are not getting further with this here on Vogons (it wouldn't really matter in which section anyway). It's time you have a talk with these persons you allude to. But make sure it's not after the third beer but a serious talk.

Windows 3.1x guide for DOSBox
60 seconds guide to DOSBox
DOSBox SVN snapshot for macOS (10.4-11.x ppc/intel 32/64bit) notarized for gatekeeper

Reply 52 of 201, by Dominus

User metadata
Rank DOSBox Moderator
Rank
DOSBox Moderator

And no, I think we would love that $100 device but we don't think it can be done and you have given us no indication that you could pull this off. In contrary you just ensured us that it can't be done (for $100)

Windows 3.1x guide for DOSBox
60 seconds guide to DOSBox
DOSBox SVN snapshot for macOS (10.4-11.x ppc/intel 32/64bit) notarized for gatekeeper

Reply 53 of 201, by spiroyster

User metadata
Rank Oldbie
Rank
Oldbie
WDStudios wrote on 2021-06-21, 19:30:

That is correct; a product like this would primarily find use in applications that demand massively parallel floating-point operations...

WDStudios wrote on 2021-06-21, 19:30:

For the OS, we'd ideally acquire a Win98 volume license on the secondary market, but barring that, we could sell systems that dual-boot FreeDOS and ReactOS with a CD full of Win98 drivers and a note advising people on how to find their own copies of Win98.

Not going to happen I'm afraid.

Even if you envisige some sort of 'cluster', these OS's are probably the worst choice.

CAD benefits from SMP but not cluster architecture. Perhaps you are thinking more FEA/FEM?

There are already multi-core (18 core) SBC's on the market, or at least there were (parrallela board). And you would be competing against a PI cluster which has an established (all be it relatively small) user base, using hardware which would probably run rings around an 'old architecture', low clock speed, x86 cluster.

But, imo, the main problem with this idea is that something optimised for a massively parrallel architecture, is pretty much the polar opposite of what 'retro' gamers need/want (in most cases).

WDStudios wrote on 2021-06-21, 19:30:

Yeah, it seems like no matter the program, there's always a third-party hack or workaround somewhere to get them to work on new hardware and OSes. For the Win95 ports of C&C and Red Alert, there's the "cnc-ddraw" wrapper (and you have to run the installer in Win95 compatibility mode). If the Earthsiege II installer is giving you an "integer divide by 0" error message, there's a solution at http://www.sierrahelp.com/Patches-Updates/New … llers.html#ESXP. Quake II's OpenGL renderer is too dark on modern GeForce graphics cards? Pull down the console and type "gl_modulate 2". I just recently figured out how to get Blood II's Direct3D rendering mode working correctly. But finding these hacks and workarounds often takes hours, per game... and even then, they're sometimes a bit of a crapshoot. Updated or remastered versions are sometimes available from Steam or GOG, but many people don't like the idea of paying a second time for a game that they already have.

These 'hacks' you mention are software. A new custom SBC (hardware) is probably only going to give you more headaches. You could define your hardware and then write the numerous hacks for each program/game to best fit your hardware, but this is more akin to a console, and this is the problem with any project like this which attempts to provide an out-of-the-box DOS (and especially win9x) gaming experience. Not all games are written to assume the same hardware environment.

WDStudios wrote on 2021-06-21, 19:30:

A big part of the draw of a product like this would be that almost everything would install and play correctly from the original disks with no third-party hacks, workarounds, command-line parameters, typing arcane jibberish into consoles etc... there would be some exceptions of course, like the Win95 ports of C&C and Red Alert. I'm pretty sure those were already broken by the time DirectX 9 came out, so getting them to work "out of the box" would mean sacrificing DirectX 9 (and possibly earlier) games, which is far too high a price to pay. But, you know... it would be something.

Due to the myriad of different configurations avaliable at the time, there is no 'one size fits all'.
Ask 10 people around here what the best configuration/experience for <insert game here> is and you will get 21 answers.

Providing 'configurations' for a desired PC hardware/software setup, on a per game basis, is probably a good start. However without providing the hardware, 'on a chip' for each configuration, this would have to be entirely done in software... which is exactly what VM's and emulators do....and the size of the catalogue of games that would have to be merticulously checked and fixed/'hacked' is quite scary. Ask GOG.

The biggest issue I think you have (imo) is that something designed to crunch numbers, concurrently at that, does not fit a 'retro' gaming mantra. There is a reason "workstation's" exists, it's not just marketing.

Imo, I think you need to do decide what it is your 'problem' is that designing and manufacuring some small SBC will fix. If it's 'out of the box' retro gaming on a cheap device, I would say stay clear.... or at least do some more market research about the actual games you want to run, but give up thinking there is some silver bullet to provide the best, easiest experience. It would most certainly be a more software than hardware 'solution' if it does exist.

However, if you want to design some small, scalabled, cluster-able, x86 SBC, I think there might be a market for it, but it would be very small, and would be purely for academic/curiosity sakes. I say this because the types of programs that require clustering, or multi CPU/GPU cores are written a specific way and are running on that hardware. Anything which may run emabarrisingly parrallel routines will most certinaly get better throughput running less cores at a faster clock rate then a cluster. And you would be competing against the likes of Raspberry PI, who have working, evolved cluster hardware on the market, relatively cheap, and pretty fast... but still not really worth doing any 'real' leg work on. Scaling them up enough to compete with x86 cluster throughput would probably cost more, and ultimately might even end up being less power efficient. idk. Either way though, it's a small market and without proper financial backing (who knows how much, but I suspect not cheap) and time (or big enough team) it wouldn't even get exposed to that small a market, and probably even less chance of it finding a niche providing cheap x86 SBC learning tool.

If you do continue, all I can say is a very sincere 'good luck'.

/2cents

Reply 54 of 201, by Shreddoc

User metadata
Rank Oldbie
Rank
Oldbie

I agree there is no need for us to get emotional, as that is a downward spiral which only sees discussions ended. I think we can all agree this would be a very challenging project, and in some ways, much more than the author envisions.

The author does not want to go too deeply into the technical side, but that is a forlorn wish in this place which is literally a gathering of technical minds who come here to discuss technical things. Realise that, author - discussion at Vogons WILL be technical, whether wished or not. [I also think you need extrapolate that understanding, to realise that your target market for such a device will also tend that way!]

And the crux of the issue here is that the Stated Goal, and the Proposed Way Of Achieving It, are not currently aligned. That is, we enthusiasts know that a Windows 98 system based on ~Socket A era hardware can only cover a limited range of what we, a community of experts in the subject, consider to be "retro PC gaming". We work with this stuff every day in great detail, for decades.

And without a more formal definition of the project e.g. "full compatibility with defined standards xxxx, xxxx, xxxx, xxxx" we are unable to take any evaluation further than that basic truism.

Reply 55 of 201, by kolderman

User metadata
Rank l33t
Rank
l33t

I can't believe we are seriously considering that AMD would develop and send to fab this produce that might sell a few thousand copies (if that, how many customers does serdaco have? I am guessing it will be similar).

It's like dreaming that Nvidia will restore the 3DFX brand and finally release the Voodoo Rampage to succeed the 3xxx series, and game companies will start using the Glide API again. It would be amazing, but keep on dreamin'.

Reply 56 of 201, by WDStudios

User metadata
Rank Member
Rank
Member
Dominus wrote on 2021-06-21, 20:47:

Did I see you mention FarCry and how it fares in Virtual Machines?

Yeah 🤣... that experiment had less to do with "retro gaming" than with my desire to avoid Windows 10 at all costs, which will eventually mean fully migrating to Linux and playing all my Windows games in virtual machines or WINE. Nonetheless, Far Cry does officially support Win98, so it technically falls in the target window even if it's not retro.

Dominus wrote on 2021-06-21, 20:47:

Retro is for many of us DOS, Windows 3.x, Windows 9x.

The target of this project would be Win98 compatibility.

Dominus wrote on 2021-06-21, 20:51:

In contrary you just ensured us that it can't be done (for $100)

I have ensured no such thing. Like I said, literally NO technical details of this project have been set in stone (other than a target of Win98 compatibility).

So, you're not going to do anything about Socket3 breaking the forum rules?

spiroyster wrote on 2021-06-21, 22:04:
Even if you envisige some sort of 'cluster'... […]
Show full quote

Even if you envisige some sort of 'cluster'...

... CAD benefits from SMP but not cluster architecture....

...you would be competing against a PI cluster...

...if you want to design some small, scalabled, cluster-able ... the types of programs that require clustering ... get better throughput running less cores at a faster clock rate then a cluster ... Raspberry PI, who have working, evolved cluster hardware ... Scaling them up enough to compete with x86 cluster...

Why are you obsessed with clusters? I never said or implied anything about clusters :\

spiroyster wrote on 2021-06-21, 22:04:

You could define your hardware and then write the numerous hacks for each program/game to best fit your hardware

Oh hell no. The whole point of this is to mitigate or eliminate the need for such hacks.

spiroyster wrote on 2021-06-21, 22:04:

Not all games are written to assume the same hardware environment.

Nor are they all written to assume the same software environment, as I already pointed out with the Win95 ports of C&C and Red Alert. There's no perfect solution, but there are certainly better possible solutions than what's currently available.

spiroyster wrote on 2021-06-21, 22:04:

Due to the myriad of different configurations avaliable at the time, there is no 'one size fits all'.
Ask 10 people around here what the best configuration/experience for <insert game here> is and you will get 21 answers.

Again, I never said anything about a "one size fits all" solution or "best configuration". The goal is out-of-the-box compatibility with the largest possible number of DOs and Win9x games, at a reasonable price.

Shreddoc wrote on 2021-06-21, 22:31:

And the crux of the issue here is that the Stated Goal, and the Proposed Way Of Achieving It, are not currently aligned. That is, we enthusiasts know that a Windows 98 system based on ~Socket A era hardware can only cover a limited range of what we, a community of experts in the subject, consider to be "retro PC gaming".

Limited, but not significantly so. I've already acknowledged that there will be the occasional games, like the Win95 ports of C&C and Red Alert, that were just programmed stupidly or based on certain failed predictions of what future hardware and software would support. And there will always be the occasional person who insists that "it's not retro gaming unless it supports Glide and Adlib", but that person does not speak for everyone. I think it would be a mistake to assume that self-proclaimed "experts" at Vogons are representative of the average person who just wonders why their copy of Doom 3 won't run on their Windows 10 computer.

kolderman wrote on 2021-06-21, 23:40:

I can't believe we are seriously considering that AMD would develop and send to fab this produce that might sell a few thousand copies

A few thousand to the retro gaming market. As I've already said, the product would need to be cheap enough and general-purpose enough to sell outside the retro gaming market. That was what happened with the R-Pi: it was intended for educational purposes but ended up being cheap enough and general-purpose enough to sell far beyond that market.

Since people like posting system specs:

LGA 2011
Core i7 Sandy Bridge @ 3.6 ghz
4 GB of RAM in quad-channel
Geforce GTX 780
1600 x 1200 monitor
Dual-booting WinXP Integral Edition and Win7 Pro 64-bit
-----
XP compatibility is the hill that I will die on.

Reply 57 of 201, by bZbZbZ

User metadata
Rank Member
Rank
Member

Why are you still here, bulk replying everyone, defiantly disputing that feedback that YOU asked for?

If you think the self-proclaimed experts at Vogons don't represent the customer for your idea, why do you seem so desperate to make us agree with you?

You could be using your exemplary powers of persuasion on these 'people you know' and go convince AMD (or VIA, or whoever) to build your $100 (or not $100) SBC that finally lets the average person play DOOM 3 on Windows 98.

Reply 58 of 201, by mR_Slug

User metadata
Rank Member
Rank
Member

I'm a bit confused by this long thread. so forgive me if i have missed a few things.

The target is win 98 compatibility. So we are excluding those very old cpu-timed games that only run on an XT. And we are excluding games from the XP era that don't support 98.

So we are left with (broadly) 2 types of games. Those well-behaved games that run on 98. And those poorly-behaved games that run on 98. Well and Big Rigs:-) So is this device only going to support the well-behaved games? If so what are the advantages over the existing solutions?

Or is this going to go for the poorly-behaved games too?

The Retro Web | EISA .cfg Archive | Chip set Encyclopedia

Reply 59 of 201, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

You know what would probably be easiest though not cheap... Use a SBC such as one of these Ryzen based boards with onboard decent graphics:
https://www.electromaker.io/blog/article/best-x86-sbc

Then use whatever wrappers, emulators, etc. needed for general use.

For ISA slots, you can use the not-so-cheap USB2ISA boards and their version of DOSBox or a modified newer version of DOSBox to allow sound passthrough.

For PCI and AGP, you would have to have your own custom SoC or chipset or reuse old parts in your design if they are available.

PCI should be doable through USB but again you would have to make your own custom solution.

AGP should be doable through Thunderbolt but again that would be a custom solution.

Yamaha modified setupds and drivers
Yamaha XG repository
YMF7x4 Guide
Aopen AW744L II SB-LINK