Repository navigation
dev call 20261012
Jitsi: Recording is on.
Ryan: Okay. So even if a few people are here, this will get recorded and transcribed. So we're at least performing for the future audience. Thanks for showing up, though, guys and girls. Welcome, Susan. All right. News of the week. I may go represent us at the Rust Bitcoin Summit in Nashville next week. All the cool Rust projects show up. And I guess we're trying to be one, so I'll probably show up. We use the library rust-bitcoin for its sort of fundamental types. That's sort of like the bottom layer library that Mujina uses. So we're actually users of... Some of the code that they'll talk about at the summit. But I imagine it's also about libraries like rust-bitcoin, BDK, LDK. But I'm sure we'll find a place to fit into as an application layer slash user of libraries. I'm also going to run a Mining Illustrated at the Workspace in Chicago in November. So I don't remember the exact date but some point in November you'll probably see that on Twitter they'll announce it when it becomes the next cool thing happening at the Chicago Bitcoin Collective I don't know if you guys see tweets about that occasionally but it's it's very cool highly recommend that we do something in Chicago someday And then there's a hash rate heating workshop in New York in November, the beginning of the week that I'm going to do the thing in Chicago. Tyler, you want to say a word about that? We joined their meeting and
Tyler: Yeah, sure. It's a cool idea because Andy, we all know, is an awesome hacker in the Bitcoin mining space and Heatpunk space. He's working with Harrison, who you might know from HRF. He works at HRF, but it sounds like he has organized a hacker space similar to PlebLab in New York City called Freedom Labs. Not affiliated with HRF, just something he's passionate about with a group of guys out there, which is rad to see, and they're kind of kick-starting with a event about some of the awesome mining systems that PizzAndy and Aadhi have put together, the 3D printer and the filament dryer, and formulating that into more of an educational event to invite hackers maybe outside the Bitcoin ecosystem, 3D printing type folks, but also the Bitcoiners in the area, followed by a get together at PubKey after the event, but they reached out to Ryan and I about potentially coming out or at least being involved in the program structure to tell that narrative of, okay, how do we get from the industrial Bitcoin mining to this crazy 3D printing device from Aadhi and PizzAndy and the open mining stack being a big part of that narrative. And I think they're going to film some content out of it. It could be good for marketing, getting awareness on our projects. So we'll see if we're able to make it out there, but at the very least we want to support that event.
Ryan: Yeah. Maybe at a minimum I could fly out there with our demo. But I don't know yet. It's subject to our travel budget perhaps, which is minimal at the moment. But it's nice of them to think of us and try to include us in this. I mean, we'd love to do more in New York at some point. Sounds like Bitstorian, for example, has a lot of access to or does a lot of stuff for students at the colleges. And we'd love to participate in something like that.
Ryan: All right. On to notable PRs for the last two weeks. So there's been some action finally on the SV2 PR from Jayr. I have posted my long and long-awaited review of this. So that is now out there and pending. I will poke Jayr separately just to make sure he knows, but he probably got GitHub notifications about this too. But I guess the summary is I think we're very close to being able to, with a few adjustments, to pull this into Mujina and then be able to call SV2. We'll probably call it something like a beta feature. When we first do it. Just because who knows if it has some corner cases that breaks or something.
Tyler: Don't forget, Ryan, that we need to totally abuse and take advantage of marketing everything.
Ryan: You're right.
Tyler: Let me know when I can smash a tweet about SV2 support.
Ryan: Okay. Will do. It looks like I got a contribution to the SV2 repositories merged today. So that's cool. Since we joined the working group, we might as well contribute a little bit, right? I had to fix something. So that's that. As a part of this SV2 support, I made a number of changes in PR number 125. And I've asked Jayr to review this. Anyone's welcome to review this though. So like there were a number of things in that SV2 PR that changed some stuff in like common Mujina code. And for some of it, rather than try to explain what I wanted instead, it was just easier for me to just sort of write this on my own and make a preparatory commit that fixed some of the things in the common code of Mujina that was lacking for SV2 support. And so there's this PR. People are welcome to review, Jayr especially. And then his PR can get rebased on this. And then the combination will be what we can merge to call it SV2 actually in our main line. Also, people are welcome, even if you don't want to review this code, people are welcome to test this. So, like, if you pull Jayr's branch and just run it with a Bitaxe or even a CPU miner perhaps. Although with that low difficulty, I'm not sure what pool you can usefully use for testing. But with a Bitaxe, certainly you can test against Braiins and CKPool and DMND pool and all those SRI stuff. All those are valid ways to test SV2. So I welcome testing for this also.
Ryan: All right, there are a number of other PRs that are smaller things that I just noted here because there was activity the last two weeks. If Jeremiah or Jolade were here, I'd ask them to talk about this. But maybe I'll skip it because Susan is here. And Susan, would you like to say something about this next one? I'll admit I haven't looked closely at these yet, so I don't really have any strong opinions until I take some time to look at them. But you want to talk about this one you posted?
Susan: Yeah, so the PR was a follow-up from the other one that I had done. I followed the dependency reduction discussion. And from there, bitvec was listed as one of the dependencies that could be replaced with custom code. So I just followed up on the discussion and implemented the approach that was suggested on the discussion.
Ryan: Cool. Thank you. This is a good background task here Susan's talking about to go through and remove some dependencies by replacing them with our own custom code when it makes sense to do that. So thank you for this. I will be sure to take a look soon. I also appreciate that other people are reviewing these, that reviewing other people's PRs is a great way to contribute. Awesome. Thank you, Susan. We also have a similar contribution from Wandia. Am I saying your name? Right? You're here? Do you want to tell us what you did here?
Wandia: Right, so hi everyone, my name is Wandia. And I just started looking at Mujina last week, the two weeks ago. Right, so as I was setting up everything, I realized that when I was running just checks, it was failing because I had an older Python version, so it was just like cleaning up the docs because I didn't see it anywhere and I was thinking maybe for a beginner that will be helpful.
Ryan: Yeah, absolutely.
Wandia: Yeah.
Ryan: Okay, thank you. I will take a look at this one. I'm just thinking on the fly. It also occurs to me, and maybe I'm thinking of a different repository, like the Bitaxe firmware one, but I had a just recipe that was like, check that all the prerequisites were installed, at least. Maybe if we don't have one of those in Mujina, we should add them. Or maybe this needs to be added to it, in addition to a documentation thing. If we had a just setup or something like that. I don't remember if I'm thinking of a different repository. OK, cool. Thank you for testing this and for circling back and contributing to the documentation when you found something that was confusing or new.
Ryan: There is a PR. I've talked about this last couple of weeks for adding speed control to the fan on the Bitaxe when we're using it with its RHAP firmware. I think it's pretty complete, but I would love for more people to test it because nothing would be worse than merging something that tried to cause a benefit by reducing fan noise and power consumption but didn't work right and ended up smoking overheating Bitaxes. So the more testing we get on this, the better before we can merge it. Also because I've seen different behavior across Bitaxes from different manufacturers when it comes to how good the heatsink is or how well even the fan controller hardware works. So I will wait to hear a little more about some testing before I merge number 33 there. Number 112, we talked about this last week. I had posted a PR that updated the documentation to explain the new repository for the Bitaxe pass-through firmware that we no longer share it with the all those other Bitaxe boards, and there's kind of a standalone repo. So I went ahead and merged that. It was just an op. So that new users, when they come looking for instructions on how to run on Bitaxe, pull firmware from the right place. There's a related PR here over on the old repository that this sort of belongs to Skot and Johnny, I guess. But I had put a PR over there just to... So we updated our own docs, but I wanted to make sure those other docs didn't point people in the wrong direction too, that those other docs pointed to this new repository somehow, if people came following old instructions, looking for bitaxe-raw firmware over there.
Ryan: Let's see, new issues in the last two weeks. Or activity on those issues. Let me click through this one, because I don't entirely remember. OK, so this is a red team finding that I converted over to an issue where we just didn't handle a lot of cases of malformed stuff come from the pools. And I think the reason I noted this here is because someone has offered to work on part of it. And I think there's a PR pending for this too. So I will reply to this and say, yes, of course. PRs are always welcome. Looks like she also found an issue. Yeah, I've noticed this too. And I've, or I've seen other people mention it that, Hot plugging still has some issues. Sometimes if you disconnect one board, Mujina will be confused about which board the cable got pulled on and will, in software, disconnect the wrong board. So this really only affects you if you're using multiple hash boards and unplugging them while it's running. But that is definitely something we want to be working. I will probably kick the tires on this again and convert this discussion into an issue, because it really is an issue, I think. Wow, thank you for testing this and filing about it.
Ryan: And then to follow up on last week, I guess we talked a lot about BZM2 support. So there's been activity on that front. Ronald submitted those PRs, revisions to those PRs a couple more times. So that's good. I talked to Ronald about how maybe the 256 Foundation could get an RDS, and he offered to send one if we would cover the shipping. So I can do that and follow up with him on that because that will help me evaluate and merge this stuff better if I can actually run it on the hardware. If Ronald were here, I would invite him to explain his changes, but that's okay. Also, I don't know if you saw, if you get notifications on the repo, but Johnny posted this yesterday. Johnny9 posts new fresh links and his intentions per our conversation last week about getting the new Bitaxe Bonanza to run. With the same drivers in Mujina. So I'm really excited about that. I was pleased with our discussion last week. We seem to have concluded, oh, there is a lot of life left in these Intel chips. Not only did Ronald talk about there being a lot of hardware out there that he's aware of, a lot of these RDS units and other things like that. But also the Bitaxe Bonanza is about to come out and we definitely want that to be supported in Mujina because it's a different chip type. So it looks like Johnny's re-energized to make that happen. And again, if he were here, I'd invite him to talk about this. Johnny, when you read this, your checklist of things seems perfect here and will help in any way we can. And hopefully I can get my hands on a Bitaxe Bonanza soon also so that I can test the stuff as we merge it. So probably the hardest part of all this will be. Not hardest, but biggest part of this will be like consolidating the work Johnny's doing and the work that Ronald's already done into one driver for the chip at least. So that'll be fun. I think we can, that can't really begin in earnest though, until this next round of Bitaxe Bonanza boards lands. So that Johnny has one so that maybe I have one, Skot has one. Maybe we can even get one out to Ronald if he's also doing the testing on this chip. Anyway, we got to find a way to reconcile these two branches a little bit at some point. But that'll be cool because then that Bitaxe Bonanza can run just like we run Bitaxe now, the Bitaxe Gamma right now as a development board. So that's nice to see. And that pretty much covers the activity in the last couple of weeks.
Ryan: My big outstanding things here still are the next big milestones are installers for S19 and some merging of board support for that into mainline. So stay tuned for that. Once that lands, we can make a big splash about it, Tyler, and we can call for testing because that's something anyone with an S19 can test. You're muted. Sorry, you're muted, Tyler.
Tyler: Which S19 models, remind me, are we targeting to support by the time an installer comes out? Just curious.
Ryan: J Pro and K Pro. Although it may be very, very close to work on other similar things. I just don't have those similar things here to test with.
Tyler: Okay. Do we have a vanilla S19, Dylan, we could send, Ryan?
Dylan: We just have J Pros. Mike might have some.
Eco: I have an A Pro. A as in alpha.
Ryan: What the heck is that? I've never heard of that.
Eco: You want it? I'll send it to you.
Ryan: I'll have to look it up first. I don't even know what's inside of it.
Tyler: If we send you a board, would you prefer a board or a rig?
Ryan: Probably just a board. Okay. And then we can talk about the rig differences. So, like, I guess it's also important to note this is for the Amlogic control boards. And not the other types of control boards that might be floating around. I'm aware of two other kinds, a Xilinx one and a BeagleBone-based one. I think those are both rarer.
Tyler: We were at the Compass Distribution Center, which is in Denver, a couple weeks ago, and I saw just pallets of machines that were marked for scrap, and I'm sure there's some working boards in there. So if you made a little wish list, I'm happy to just zip over there and ask them.
Dylan: We have Xilinx and BeagleBone on hand.
Ryan: Okay. I have a BeagleBone here. I don't have a Xilinx here. Yeah. Well, I mean, we'll probably just release what we got first. And then quickly we can expand to... It'll only be small adjustments from there to new hash board models with the same control board probably. And then only a little bit more work to have new images for those other control boards. So the way it'll look is each control board will probably have its own image. But that image will support any miner using that control board. Does that make sense on how that matrix would look? Yeah. So two dimensions for us to work on there. Rig support, I guess, and varying control board support. Yeah, I'll put together a list. I'll try to put together a list of like what models even exist and what we know about how common they are. And then we can start like checking off, like who has what, where, and what we still need to get from people like Compass. Or I think Mike Hamilton had a bunch of stuff he offered to send us. Schnitzel or Eco seems to have a trailer full of stuff. Never heard of the A Pro, so that'll be fun.
Eco: I just sent you on Signal a picture of the control board, if it's any help. But yeah, I've got a bunch of that stuff just sitting in a storage unit. Wow. Collecting dust. Cool. If it can be put to good use with you, I'm happy to... Make that happen that looks like a Xilinx control board might be I'm not sure if I'm zooming in correctly what is the total hash rate of an A Pro it's 110 is the nominal hash rate It's been a fantastic miner. Never had any problems with it. What firmware have you run on it? Just the stock firmware that it shipped with.
Ryan: Cool.
Eco: Yeah.
Ryan: I have no idea. It looks so clean.
Eco: I wonder how popular these are. Well, it looks clean in that picture because I took that picture right when I unboxed it. So it was completely like brand unused, completely unused and brand new when that picture was taken. Cool. And it's spent its life ever since then inside a Upstream Data black box outdoors. So probably doesn't look quite as clean in there any longer.
Ryan: Neat. Oh, sorry. I guess if I were on the ball, I would have shared this picture somehow, but that would probably take me forever to do that in an opsec safe way. It looks just like a J Pro from the, from the top, except I think it has a Xilinx board in it.
Ryan: Okay. Well, open floor. Does anybody have anything they want to talk about?
Dylan: No, you say it. Whatever. Are there instructions for using RHAP-D like there were for bitaxe-raw or running, you know, Gammas or whatever?
Ryan: I think so. Instructions in what way? Like, what do you want to do with it? Like, for installing?
Dylan: For installing and just running it.
Ryan: Let me click through. I think this is the doc you want. How do I get to the non? Just take me to the file. Something like this, like if you start in Mujina. There's this board documentation for all the boards we support, which tells you what firmware to put on them. I think if this doesn't say in enough detail, then if you click through and follow that firmware link, it'll tell you how to build and install this. Yeah. And I simplified it significantly. Again this is a somewhat environment sensitive thing so if you try this I'll be curious to hear how it goes for you because the instructions used to be harder I simplified them but it might be you might run into something where like on your machine the version of such and such is different or sometimes this stuff relies on upstream servers where versions change and stuff like that too so yeah be sure to tell me how it goes if you try this Okay, well, I don't really have much more to share at the moment.
Tyler: What would be most helpful for you, Ryan, for people to put any resources or attention on?
Ryan: Well, I guess at this point, testing on that SV2 branch, or if you have a Bitaxe, testing of that fan control behavior. Those are the two things that are sort of ready for testing that are new. Soon, when we get some of this additional stuff merged, there'll be a lot more stuff to test. S19 support not only comes with a lot of changes for supporting those boards, but all the API and configuration stuff is going to land very shortly there too. And so all that new API stuff will be ready for testing and persistent configuration files stuff. Home Assistant integration. It'll be ready. Can't wait till this runs on the hash tub.
Tyler: Yeah. Can't wait to open up that tank of immersion fluid to flash it in firmware.
Ryan: Better you than me. OK, well. Anyone else?
Susan: I have a question.
Ryan: Yes.
Susan: Sometime back, Ronald mentioned that he had some BZM2 hardware lying around, and I wanted to ask if it's possible for us to get one shipped to Kenya.
Ryan: That's a great idea.
Susan: Yeah.
Ryan: Are you in our Telegram, 256 Telegram chat?
Susan: Yes, I am.
Ryan: I know that Ronald's in there. You should mention it to him in there and we'll all give it a thumbs up. Are you involved or going to be involved in the open source lab that Gridless is trying to open up there? Are you aware of that?
Susan: Yeah, I'm aware of it. I've heard of them speaking about it, so they haven't mentioned or told anyone to participate in it.
Ryan: Yeah, I think they're still getting started, but that might be a great place for us to make sure that one or two of these units get set up permanently. Since they take a lot of setup and power to run.
Tyler: I just tagged him in the Telegram group so that we can all kind of upvote it and see what he chimes in with.
Ryan: Yeah, perfect. All right, good idea. Good requests. I'm sure at some point we can make sure that some Bitaxe Bonanzas make their way over there too.
Susan: I also are yeah no go ahead in regards to the PR for the failing just checks I and Wandia were able to do like an audit on why it was failing and it's actually linked to a PR that you recently made where you specified The Python version for the just checks. So for anyone who is using an older version. The test will fail. I've linked it, yeah, yeah. Oh mention yeah, got it.
Ryan: OK. Thank you. If you want to add a comment to this PR or something, if there's more to say about that, that'd be great.
Susan: Okay.
Ryan: Eco, what's the plan or thinking about 256 dev kits at the moment? Eco or Tyler?
Eco: Or are we still planning? I mean, still planning. I don't think we have anything concrete yet. I think it's a good idea. And I'd like to support it.
Ryan: As long as we were thinking about hardware and distributing hardware, that just came to mind.
Eco: I guess for me, the biggest question mark for me right now or the, I guess, variable is that I need to get my pick-and-place machine up and running. As it turns out, I'm not going to be able to do it at the warehouse that it's in right now because my sister-in-law actually needs that space to work. So she wants my stuff out of there. So I need to find a home to set that stuff up so I can start producing hardware that would include like LibreBoards so we're getting really close to we're gonna go with a third party vendor to produce the v3 of the LibreBoard so that we can do our validation testing. So we're probably going to have that in our hands possibly by the end of October, definitely by the beginning of November. So we're just about ready to pull the trigger and buy a handful of prototypes from a third party manufacturer so that we can validate v3. Once that's done, then they're ready to start getting produced. But you know, I would love for those to be part of a dev kit. I think it's a critical component and needs to be part of it. I'd love to produce them myself. I just, I need the, I need the workspace. So I know there had been some discussion about possibly using the space, to make a new home for the pick-and-place machine, but that's kind of, you know, it takes time to work through all those details. So, you know, we're still figuring that out, but I think that probably looks like the best prospect. So all that is just a long winded answer to say, I don't know when I'm going to be able to start producing hardware, but I would love to be producing it as soon as possible to help support something like a dev kit initiative.
Ryan: Cool. All right. Well, that's all the ideas I can think of to stretch this out. Anyone else? All right. Well, I think we're good for the next two weeks then. Thank you for joining again, everybody. And we'll keep our nose to the grindstone, which have a lot of news in the next coming weeks. Once we have installers people can use on S19s, I think that'll really open the floodgates as far as users and probably bug reports and all the above. So that'll be fun. This is the calm before the storm.
Dylan: Exciting times. Thanks, Ryan.
Ryan: All right. Thanks, guys. Thanks, everybody.
Dev-call summaries and transcripts live in the Dev Calls discussion category. Transcript pages here are linked from their discussion threads.