Видео

Of course, I am always adding repository for project in video description. This one is under: github.com/matgla/mspc/tree/mspc_v2 But please keep in mind that it is not well tested yet. Right now I know about two limitations: 1. I forgot about 3.3 power supply from USB, so it can be delivered only through bus 2. I did misclick on SPI flash CS# line for FPGA configuration. Fix was delivered to mspc_v2 branch, but I am waiting for PCB to test it. 3. Most probably more problems, not yet discovered! About parts, you can check BOM from Kicad project. Right now all parts should be available to order, but there is problem with LFE5-UM which for long time was not available at now, now they are quite expensive.

Thanks!

Yes, they are not the best parts to hand solder. But I didn't have a choice :) and I always wanted at least to try solder BGA

me too, it's totally unuseful code :) and it sometimes works because it uses old GCC compiler implementation hacks

Of course! And this card will be a real "gpu" like card. With VGA( and DVI ) signals driven from PIO. And text and graphic modes frame generation on those 2 cortex-m33. I have a proof of concept on two raspberry pico with rp2040 with even some experiments with simple 3d acceleration.

I have plans, but let's the future verify my free time :)

Doom is achievable on the pure MSPC mother board, since it is possible to run doom on RP2040: kilograham.github.io/rp2040-doom/

Hi, it is www.falstad.com/circuit/ I like it for simple circuits, because of its simplicity.

circuitjs by falstad

Yes, most probably more than enough for me :) When I was selecting an FPGA 2 or 3 years ago, I found that the ECP5 family was quite cheap. Initially, I planned to order a 24k/44k LE version, but then I saw that some stores in other EU countries had a reasonable price for an 84k version, so I ordered two. And they have been waiting till now :)

@@matgla95 good decision, they come to perfect use now.

Yes, and I am planning to offload some tasks to RP2350 on both the motherboard and VGA extension, so I should be able to save some logic elements for other tasks

That's rather question for RP2350 datasheet or designers (through raspberry forum), but looking quickly inside documents I don't think so. RP2350 contains encryption module with secure boot, but the flow is to verify image from QSPI flash then load decrypted image inside SRAM. And I would need to deep dive into datasheet, but I believe there is a way to prevent external access through JTAG/SWD and your software should be quite safe (of course except security leaks inside firmware itself)

It's not scrapped :) It's just evolution, complexity of hardware is going to be reduced, but components inter-working that I mentioned will be implemented to communicate between motherboard and extension cards. Previously I planned to use exactly the same implementation for north<->south communication and for north<->extensions. With RP2350 only north<->extensions is left.

thanks!

Standalone chips are not yet available. But you can buy Raspberry Pico 2 which contains RP2350A.

Thanks! I hope it'll be working fine after soldering :D

Yes, exactly! Raspberry Pico 2 is already available to purchase. RP2350 inside contains dual core arm cortex-m33 and dual core risc-v, but you have to select which one you are going to use at startup ( as far as I know )

Thanks! This will be covered in a dedicated episode. I plan to make it together with graphs from the oscilloscope. But it must wait for RP2350 arrival and board soldering and most probably fixing :) this video is just a short overview of the first step!

Thanks!

Thanks!

Of course you're right! I didn't know that, so thanks! So I'm sorry about my mistake inside video.

Thanks!

Yes, that's just overview. I just wanted to cover transition towards RP2350 from my RP2040 board, short overview of design and hopefully in near future short overview of soldering and debugging. Actually design for RP2040 or RP2350 is really well described inside their official hardware design guides. Second thing needed is general knowledge about PCB design and there is a lot of high quality channels with with people much wiser in that domain than me. I personally was learning from Robert Feranec channel.

Thanks! Since I am still waiting for RP2350 availability I am working on video edit for VGA and FPGA extension.

Thanks, keeping fingers crossed for your project !

Thanks, new episodes are under recording. Hopefully I will release them soon!

Thanks!

Thanks!

My goal was to show all steps of the project, but then I realized that it is hard to follow. I hope some day I will find the balance between what to show and what to cut :)

Yeah, that's true, that new videos are much shorter.

Good to know. Latest videos use spectrum filtering to remove voice frequencies inside music. That one unfortunately does not use that yet.

I am glad I helped you. In my experience I realized that almost all concepts are quite simple, even if they look complex.

Writing a filesystem from scratch is also interesting. But for now I have to use something already existing.

That was one of the main goals when I decided to start my RUclips :) Use it whatever you want :)

Yeah, I wanted to show all the steps of the project lifecycle ( except maintenance, since it's too early ) :)

Thanks! I am trying :)

Thanks !

To be honest, me too. C can't be substituted in some environments. But whenever it could be I think it should be. I mostly use C++, but I am looking for tools that can improve or even replace it into safer language transpilers ( like cppfront ), where other compilers like (i.e rust are not fitting). But I really hope that the standardization committee will resolve most of the problems some day.

No Code, No Problems! :)

Yes, but it uses more than that :)

Thanks! I've noticed that, although I am still working on improving the readability of my videos new ones should be at least a little bit better.

Thanks :) currently I am working on adding PIO simulation for the next part. I think I will release a video on how to use it on other projects. Of course it still will be in work in progress mode with contributions warmly welcome :) in which environment have you tried to implement RP2040?

​@@matgla95 to be clear - by implementing I meant Renode implementation. I'm more focused on other MCU series right now ;) looking forward to see the video on how to use Renode. Maybe I'll learn something new from it :D

In thumbnail? At my computer's PCB.

So, let me clarify a little bit. The code is on SPOJ and is executed for 10 test cases. Test data is provided on STDIN in the form: 1 <- test case 1, n = 1 8 <- test case 2, n = 8 ... rest 8 test cases. Your code should output answers on STDOUT separated by whitespace for each of such test cases. According to segmentation fault. I see two problems in the code that you wrote; scanf in the original code reads a decimal number from STDIN in storing it in the k variable; thus, scanf("%d," &k), otherwise k is uninitialized and scanf depending on libC implementation may segfault, it thinks it has one argument according to format string, but it's not placed in argument list and is writing data to some unexpected memory address (most probably null). Additionally, there is '~' before scanf to detect EOF. Otherwise, the program will not end.

I think beauty is subjective and depends on context. Here I used it a little bit ironically, since you can still do things that should not be allowed, at least by default. Hopefully compilers are evolving and, i.e. GCC 14 rejects such code by default which is good in my opinion!

Thanks, for that comment! That software is created mainly for educational purposes. I don't think that I will be able to create OS/Bootloader/(What ever else) that is able to compete with any other not single person project. Also I don't see any real usecases to run it over network. It's closely related to specific hardware. For other cases there are better tools available. But if someone will find a way to use it: good for him! :) I selected GPL for software, because I like that part you have to share with community if you wanna use it. But I wasn't digging deep through all licenses availble, I just selected first one that was fine for me. AGPL also fits, but I don't think I will change all license headers right now. About hardware part, it's an experiment. I am only hobbiyst, so I decided to publish with just MIT. Basically to allow anyone to do whatever they want with that.

@@matgla95 One of the good things about the GNU AGPL is that the network use is redistribution clause doesn't really apply if you do not make the software available over a computer network. In that sense it is pretty much identical to the GPL for non networked software, but if anyone turns it into a service as a software substitute then the network use is redistribution clause kicks in immediately.

Thanks! :)

Sure! Good point and that one is even reported by GCC, when using the -Wall flag :) but the goal of that exercise is to abuse whatever you can in the provided compiler to achieve the smallest possible source code

Thanks!

Originally that code doesn't contain . It was added just to better presentation of execution. The testing from SPOJ is executed for many inputs from stdio and space is necessary to separate outputs for test cases.

Thanks :)

Thanks!

Yup, different languages due to different paradigms enforce to improve the skill set and it seems you have a win-win outcome :)

Yes you're right, I have a problem with recording on Linux ( my camera interface produces crazy delays, most probably Nvidia again broke their driver xd ) which I haven't solved yet, so I temporarily switched to windows.

@@matgla95 That's no excuse to have it running in the background. Honestly, I'd be embarrassed. You do you though. I'm sure your excuses will safe you from the social credit system you help implement by running spyware. Oh, and don't get me started on the smart phone which I'm pretty sure you have too.

No no, I am only using pigeon post :)

@@matgla95 You sound like a smoker telling me that smoked meat takes longer to go bad. I guess the copium is always real. Anyway, you keep making excuses for being the corporations' female canine. Another thing though. Are you actually in Poland? We are always looking for Team Fortress people in Europe. I wrote a fork of the original for the Darkplaces engine with more emphasis on tactics and stealth.

Jerk