redbags

Rat Race is my earliest memory of gambling. Must be 40+ yrs ago in a seaside arcade in Haylin Island (not sure I spelt that correctly). Will download and take a look.

Maybe but not sure how accurate that is. I don't know anything about the systems that came before this but perhaps they didn't use HASP i.e. rio cabinets and it's leftover code. As far as I can tell barboot only deals with installing the HASP software - it doesn't actually communicate with the HASP.

The RAM disk is strange and I can only assume it was intended to host ephemeral data hence they carved a bit out of RAM. Never seen one used on a production system before. The PE bar7.tmp sounds suspiciously like the real barboot so I think you are on the right track. You might also notice they have a tendency to name exe stuff .tmp i.e. devcon, ftpserver etc. The developer.txt file is simply a check for the presence of the file - there is nothing of significance in it from what I could tell. So just go ahead and add the file and barboot will display text informing you it's running in developer mode. Barboot also has some check for the presence of pressing of left alt key - again - not especially clear what it's intended for. There is another check for c:\settings\settings.ini which is quite important to be present so make sure you copy it if moving to another system. Ge.exe is some sort of packed version of the real executive which is called executive.exe. If you don't have the ram drive configured then just create a small V:\ drive and the real executive will be copied there and launched. If all goes well you will get the executive running and then be greeted by an alarm telling you that the cabinet dongle is missing. And then you're stuck. Re: elvis20. I believe you have the game but it's in an encrypted form. I think it gets unencrypted via the licence on the HASP which is why when you poke it - it all looks like junk. Re: hasp - I did read there are some which allow freeplay. Maybe expired or something. Otherwise I think you will need to get one. Trying to work around it without having one is going to be a serious undertaking. I would say I've been looking at this system on/off for around one year and I probably haven't scratched the surface. It's very complex with a lot of moving parts. Be prepared to burn a lot of hours/days/weeks looking at raw computer data

I went into the tank on this one. Based on where I've gotten so far I still maintain that there is NO physical dependency on the ATI graphics card - it's all configuration. Here is how graphics work in the exec:- Exec --> SDL --> OpenGL (windows) --> OpenGL (device driver) In the above - the exec is using certain functions which are not exposed directly through the Windows XP OpenGL library as it's basically a very old/small version of openGL (v1.1 I believe). Instead, the Windows OpenGL library is forwarding those to the device driver through a function pointer. The functions appear to be OpenGL 1.2 or higher. What I believe this means is as long as the installed video card implements OpenGL 1.2 or higher than you're good to go. The caveat here is that it's possible further along the line some hardware specific dependency comes into play but this seems unlikely as the video card manufacturers generally played quite well with aligning with the API specs. The fact that the T8 uses another graphics card from another manufacturer supports this idea as well i.e. they'll be doing spec driven code.

I don't believe it's been reverse engineered but that's probably because nobody has looked too closely at it. Because the system/games are unpacked at rest it's not reallly much of a barrier to anything.

It's Windows XP embedded. The application software is all just standard windows x86 32bit. Hardware wise - I know the graphics card is an Radeon X1300. In terms of all the other hardware - not exactly sure as I no longer have the physical unit. Opinions tend to vary but I'm not convinced there is physical dependencies on the underlying PC hardware except for the fact that Windows XP embedded needs to be able to boot + install the necessary drivers. PC hardware aside there are two key peripheral dependencies when considering the PC environment. One is the MPU - this is a proprietory piece of hardware connected via USB which manages all the cabinet hardware i.e. note acceptors, hoppers, led lamps etc. The other is the Sentinel HASP - this essentially serves to provide the encryption/protection of some of the software + individual games. If your goal here was to get the system up and running on something other than the OEM hardware - MPU/HASP and all that other stuff aside - I'm pretty sure that is possible. If your goal was can I run these games via some sort of emulation or on hardware that eliminates the need for the proprietory hardware - what I can say is that you will need very strong reverse engineering skills and the appetite to take on a huge task. There is probably somewhere in the region of 50Mb+ of code before you even get anywhere near a game. For anyone not versed in reverse engineering - 50Mb is massive. 10s of thousands of lines of code and some very complex code I might add.

Never say never but this just seems very unlikely to me. For starters the driver used to support the graphics is stock AMD. And modifying the BIOS of a commercial graphics hard is both a.) technically challenging and b.) something almost certainly off limits from AMDs perspective. If you look at the behaviour outlined by slots magic - sticking in any old graphics card doesn't break it - the output just comes out incorrectly. And I'm sure that's a gfx card configuration vs something more technical. The exec in T7 land is drawing to a single canvas and managing offsets etc to know where to draw to recreate the 'multiple' display effects. But the graphics card and driver has the job of sending half of that canvas to one monitor over one port and the other half to the other monitor over the other port. Things change in T8 because of the addition of an additional screen but in T7 that's how it works. I suspect that if you were to connect a single screen which supports the resolution of 2546 x 1200 (typing off hand here but it's something like that) to any graphics card it would work as expected. And final thought - keeping the graphics card uniform makes a lot of sense when you look at at the setup of these systems and how locked down/controlled they are. Fewer hardware types = less management overhead. So I think it's less about custom graphics vs a tightly controlled system where introducing new hardware becomes a headache. Now someone can come along and say.....I did all this via the OS on a T7 and it still didn't work to prove me wrong. But installing and configuring a graphics without somehow interacting with the OS - it's hard to know what you actually did.

I suppose my wider point is that I'm pretty sure they haven't done anything as low level like creating custom silicone, drivers or firmware and such such like on the gfx. I think that's a bit of a myth / something mired in mystery. I suspect any replacement card you got working either shared the same driver or that default driver setting came with the required settings. Out of interest - did you ever tinker with the graphics setup in the actual operating system?

Too bad none of us have a T7 to test it on. But I suspect the issue is one of graphics card setup + configuration vs underlying code issue. The API at that time did not support multiple displays as draw canvases so it was left to the developer to treat them as two screens but managed via a single desktop resolution. Apparently horizontal span mode needs to be configured at the driver level vs extended mode or some variation on that. And looking at the code there are some fairly strong indicators it is drawing to a single resolution and doing a bunch of operations on that to make sure things are drawn to the correct viewport. So if someone out there has a broken T7 graphics card and is willing to tinker then I think it's possible to get any card working so long as it supports the relevant API and required resolution. I'm only going on registry settings but it's something like 2560x1024.

Well that sounds more like a configuration thing vs graphics card not working. I don't suppose you tried but if you bombed out of the exec and reverted to the OS i guess they were working. I was looking at the driver install earlier so will go and take a look again. There are a lot of settings

Out of curiosity - when you say suitable - what was the criteria and where did you land? I was looking at the T7 spec and I can see it's a AMD Radeon X1300. What happens if/when you stick another non-X1300 gfx card in? Black screens?

I double checked and the exec uses OpenGL for the API and another set of libraries called SDL to abstract out some of the graphical operations. I'll dig deeper and see if I can find out where the hard dependency on this is coming from.

From memory the exec is based on a fairly old Open GL spec - I'll take a look later and confirm. Generally speaking it's less about the emulation and more about the API spec. I think what is probably unique about those graphics cards (for those times) is they probably had larger memory to deal with the resolution across two screens and multiple output ports. I'm reasonably sure a modern virtual graphics card could do the lifting of a 20+ yr old graphics card. It's also possible to use GPU pass through these days if not (I've seen this for LLM workloads but not gaming but assume it's the same).

I suspect you need to go after the exec as opposed to the individual games. I've never seen the individual game code because it's encrypted at rest and I got rid of my working system before poking around in that. But I think the exec is still doing quite a bit of the heavy lifting. HASP is the bigger blocker though. I do think getting T7/T8 games running on any PC will happen at some point once someone with the right level of skills and determination puts their mind to it. It's the skills part holding me back

I'm pretty sure this is not the case. It might have been true in previous versions of the software but there is nothing in the code which ties specific PC hardware to the software. Obviously you still have HASP and the MPU hardware/functions to deal with but that software (OS + App) will run on something other than the specific hardware.

I wonder if he ever managed to get one of these and if so how much paid? I'm in the market for one but hard to come buy. Only option seems to be buying a whole unit

Welcome aboard Andrew. I would personally be fascinated if you ever had any stories to share on the T7/random technology. It seems like such a leap in technology compared to the fruit machine stuff that I'd be so curious what it was like going through that transition. Hope you enjoy your stay here. Regards

I never did get around to finishing this. Turns out that doing the reel calcs is quite a bit harder than I imagined. Once I'm in front of the laptop I'll share what I have and someone else can look if it's of interest. Something I never really understood was why game developers used multiple reelsets? Was it just a case that it's easier to use multiple sets to home in on the required rtp and profile vs plugging away at a single set of strips? @stevedude2 But since I've seen how the designers can use excel to do the maths I imagine that exercise isn't overly taxing or as much effort as I once thought.

hey mate. It did die a little bit (xmas + kids) but just fired up the laptop to switch my brain back into mode. I wonder how much I've forgotten. Let's see.

So I ran into difficulties getting to the same number as you for the four of a kind tens. But. As I started typing out the message the light came on and I got it sorted. Exactly the same as my previous post. I think I'm going to do the calcs in code vs excel as the formulas are getting a bit messy. Let's see how I get on.

@stevedude2 I'm getting very close to your numbers but I'm missing something. See below: - 5OAK ME YOU DIFF TEN 718800 718536 264 JACK 395952 395760 192 QUEEN 248784 248544 240 KING 103632 103392 240 ACE 53313 53097 216 JP 80142 80142 0 WILD 48 48 0 If we use the 5OAK TENs as an example I am doing the following: - =(tencountreel1+wildcountreel1)*(tencountreel2+wildcountreel2)*(tencountreel3+wildcountreel3)*(tencountreel4+wildcountreel4)*(tencountreel5+wildcountreel5)-wild5oakhitstotal so...... =(15+3)*(14+2)*(14+2)*(10+2)*(11+2)-48 = 718800 I've tried a few different theories to try and get to your number but they are not working. Are you able to spot anything obvious? Cheers. Update - Oh I see it now. The difference is the same as your previous post. The W,W,W,W,T is not a winning line as the W,W,W,W pays more than W,W,W,W,T so is the winning line.

Could I possibly look at your formula for 4OAK wilds? I'm on cell 2 and failing already haha

Cheers mate. I'm getting very similar numbers to yours but I'm slightly off. I'm going to start at 5oak wilds and work back. I might shout if I can't crack it. Hands full at moment so hoping to take a good look later on this evening. Will also answer other questions as best I can.

I was in the process of going to ask you another question but if you want to take a stab at it then I for one would be very interested and will park it for now. I've zipped up a bunch of files which might be of interest. Let me know if you need anything further. data.zip

@stevedude2cheers buddy. I used wilds and only three reels to try and simplify the sums for that specific example and because I have all of the three of a kind wins dumped out to a csv so had something to cross reference against. Armed with your knowledge above I will take do the calcs for all sets tomorrow.