PlayStation MiSTer Core Repository Created

[Xbytez]

Robert Peip (@AzumFpg) has created the public repository for the PlayStation core he is working on!

https://github.com/RobertPeip/PlayStation_MiSTer

[MostroW]

Looks like a good start, quite curious on how this core will progress in the time to come!

[pgimeno]

Oh my that's so exciting! Wish he published his software emulator as well.

[Sectorseven]

So no analog out through the IO board, it will be interesting to see how users react to that.

[hiddenbyleaves]

So no analog out through the IO board, it will be interesting to see how users react to that.

I am desperately trying not to react to it at the moment.

[Duffygag]

So no analog out through the IO board, it will be interesting to see how users react to that.

Using Direct Video probably

[FPGAzumSpass]

Maybe there will be the possibility to disable the scaler reading part of the DDR3 and output analog instead.

But i don't focus on such things now.

[FoxbatStargazer]

So did Bob get you a CRT yet?

[Blitzwing]

Is this a new ground up core or does it build upon the work by Laxer et al?

[Blitzwing]

Maybe there will be the possibility to disable the scaler reading part of the DDR3 and output analog instead.

But i don't focus on such things now.

Does your workflow of building a software emulator first and using that as the basis for you FPGA core help with debugging and getting the logic right, rather than starting from scratch in Quartus?

[gamesfan]

Is this a new ground up core or does it build upon the work by Laxer et al?

New ground up.

[Tackemon]

Is this a new ground up core or does it build upon the work by Laxer et al?

From:
https://www.retrorgb.com/misterfpga-psx ... spass.html

None of the open sourced material from Laxer3A or ultraembedded has currently been utilized for the development of his PSX FPGA project.

FPGAzumSpass stated it was easier for him to start from the ground up, rather than use the work previously done on PSX FPGA development.

[hiddenbyleaves]

Maybe there will be the possibility to disable the scaler reading part of the DDR3 and output analog instead.

But i don't focus on such things now.

My excitement rating had dropped to 99% but has now returned to 100%.

[darksakul]

So no analog out through the IO board, it will be interesting to see how users react to that.

I am more than fine with this. If you ask me HDMI is the way.
But I realize that sucks for those CRT fans.

[lroby74]

So no analog out through the IO board, it will be interesting to see how users react to that.

No CRT no party!

[Duffygag]

It seems people in this thread don't know what Direct Video is

[lroby74]

It seems people in this thread don't know what Direct Video is

I am using Direct Video on 2° Mister (connected to 2 14" Philips CRT monitors) but on first one i need to not use it

[aberu]

Oh my that's so exciting! Wish he published his software emulator as well.

He usually does eventually. The software emulator is often something that runs at like 1/3rd speed is solely for helping the development process anyways, so it's not much of something to play games on.

[FPGAzumSpass]

For fullspeed gaming my emulator would require a ~12Ghz CPU, so it's worthless for that.
Most people don't understand why and just say it's bad design, so I rather not publish it.
Also I received complains about the (missing) UI. Well, for my basic tests i don't need any great UI.

The more I design the emulator to work like hardware, the slower it will be and this time I maybe did too much.
On the other hand, I could port the CPU nearly 1:1, except the cache, which helped a lot.

Hopefully the other components will also transfer over so nice.

[Duffygag]

For fullspeed gaming my emulator would require a ~12Ghz CPU, so it's worthless for that.
Most people don't understand why and just say it's bad design, so I rather not publish it.
Also I received complains about the (missing) UI. Well, for my basic tests i don't need any great UI.

The more I design the emulator to work like hardware, the slower it will be and this time I maybe did too much.
On the other hand, I could port the CPU nearly 1:1, except the cache, which helped a lot.

Hopefully the other components will also transfer over so nice.

Keep up the great work mate, unfortunately the world is filled with entitled 40 years old idiots, with no hardware/software understanding, who never had the chance to grow up and develop into decent human beings, just ignore them - The community is beyond lucky to have ppl like yourself.

[Neocaron]

For fullspeed gaming my emulator would require a ~12Ghz CPU, so it's worthless for that.
Most people don't understand why and just say it's bad design, so I rather not publish it.
Also I received complains about the (missing) UI. Well, for my basic tests i don't need any great UI.

The more I design the emulator to work like hardware, the slower it will be and this time I maybe did too much.
On the other hand, I could port the CPU nearly 1:1, except the cache, which helped a lot.

Hopefully the other components will also transfer over so nice.

My knowledge in this is very limited, but if I understand your process, you basically replicate every FPGA gate at a software level on a regular CPU which allows you to troubleshoot things and make improvement directly, but with very very slow performance. However once you reach a great level of accuracy you then translate this very slow but thorough software emulator to the FPGA chip using the same backbone. Am I close enough? Do you think there is a chance for perspective-correct texturing using floating-point precision?

Anyhow, very happy you are the one to be tacking this! Your emulators are absolutely amazing, though I still don't understand why the multiplayer cores are not already part of the official update, since they are game changers for each console you did it for.

[FPGAzumSpass]

The multiplayer cores are not in the update script, because the script as is cannot fetch from a branch. Hopefully it's added some day.


The emulator is not built on gate level, my description in FPGA also is not on gate level.
I prefer doing behavioral design as much as I can.
But it does handle every component/chip with the original resolution.

E.g. common emulators only do 1-2 cpu stages, because everything else is just for reaching high CPU clock speed in hardware.
My emulator does all 5 stages, which obviously costs 2-3x performance, but in FPGA I have to do 5 stages anyway to reach the required clock speed.

Also I trigger EVERY component/chip for every clock cycle like it would happen on real hardware.
Most emulators instead work with events, waiting until some action is done and only trigger the chip once.

[Neocaron]

Very interesting read! Thanks And yes I hope the update script gets updated as well, I feel some people are probably missing out on them, or your lastest improvements, because they are not included.

[Blitzwing]

It seems people in this thread don't know what Direct Video is

Oh we do, it's just some use both IO analogue out and the HDMI

Whilst I could plug in and unplug the adapter and the switch ini's I'd rather stick with how I use the MiSTer now, that said if it can't be done it can't be, just having a PS1 core would be an amazing achievement.

[Blitzwing]

For fullspeed gaming my emulator would require a ~12Ghz CPU, so it's worthless for that.
Most people don't understand why and just say it's bad design, so I rather not publish it.
Also I received complains about the (missing) UI. Well, for my basic tests i don't need any great UI.

The more I design the emulator to work like hardware, the slower it will be and this time I maybe did too much.
On the other hand, I could port the CPU nearly 1:1, except the cache, which helped a lot.

Hopefully the other components will also transfer over so nice.

I guess the software CPU being designed to be ported to FPGA means most of the logic executes in parallel? Great for an FPGA not so much for a CPU which works sequentially.

Don't understand how that is bad design, it just isn't meant for CPU's

Edit, your other post answered that question nevermind.

[pgimeno]

I imagine that the software emulator is a "reality check" on how the hardware works and behaves, and allows tuning or fixing things, testing ideas etc. faster and with more debugging tools than doing it directly in hardware form. That sounds like good approach.

Watching the Croc demo video suggested that the speed was quite good, so I guess I got a wrong impression. Nevertheless, I think there's value in publishing the emulator, in that users may check behaviours in certain games to see if they match the real machine the way they expect, and if not, report them even before the hardware is implemented, allowing for correction of those aspects and planning the hardware implementation in advance.

But yeah, I guess that's an "ideal world" view and that the reality is that people will complain if the software emulator doesn't perform as they expect, so it may not be a good idea in practice after all.

[FPGAzumSpass]

The software emulator doesn't support all games yet and still has bugs and no sound at all, which causes even more bugs.

For Lynx and Wonderswan I "completed" the emulators before doing the hardware, as they were not as big.
But this time it would have taken too long, so i'll be doing it as I did with GBA: implementing the hardware until some games run, then go back to all the missing things.

It's a longer process, but i don't risk forgetting how things work when I finally got 99% of the games working perfect in the emulator some day next year.
Also i need this to keep myself motivated. I cannot work for 1 year on such a project with no (playable) outcome.


As for the croc demo: the playback was at 2x speed, so about 80% of original speed, I didn't want to torture all viewers with the slow games i have to see

Edit: I'm really looking forward to test on FPGA, because it will be 100% speed. It will feel like running in fastforward.

[Moondandy]

Getting exciting. Has it been decided what the .rbf name will be for this core?

[MottZilla]

So no analog out through the IO board, it will be interesting to see how users react to that.

No CRT no party!

I agree. But this is just the very beginning. It's understandable to be a low priority at the start. Lots of people certainly hope it'll be added eventually. I certainly do as I prefer analog RGB via the I/O board.

[Xbytez]

Another PlayStation 1 core development update: