|
I'm really enjoying it so far. Got stuck on Combiner for a moment so I went back and had another look at Multiply, the only Section 2 puzzle I moved on from without finishing. Got it completed with 1 module more than the listed ideal.
|
# ? Sep 7, 2017 02:04 |
|
|
# ? May 8, 2024 03:46 |
|
NickPancakes posted:I'm really enjoying it so far. Got stuck on Combiner for a moment so I went back and had another look at Multiply, the only Section 2 puzzle I moved on from without finishing. Got it completed with 1 module more than the listed ideal. Nice! I'm probably going to add achievements for more of the side levels at some point, Multiply among them - it's enough of an obvious challenge that it'd be good to acknowledge. News:
|
# ? Sep 8, 2017 05:26 |
I've been doing other project lately, but I have made sure to mess around with the new versions when they popped up in Steam. I seriously find it hard to believe it comes out in only a few days! I also got my brother who's into electronics to tell me "There's no such thing as regfiles. You must be talking about an OS's registry." And they say video games don't teach you anything! Speaking of updates, were there any historical figures or electronic concepts you decided against putting in due to some reason or another?
|
|
# ? Sep 15, 2017 11:25 |
|
Doom Goon posted:I've been doing other project lately, but I have made sure to mess around with the new versions when they popped up in Steam. I seriously find it hard to believe it comes out in only a few days! I also got my brother who's into electronics to tell me "There's no such thing as regfiles. You must be talking about an OS's registry." And they say video games don't teach you anything! Heh, excellent. There's no particular historical figures who I can think of - anyone who I was interested in I mixed in somehow, though it might've just been a personality trait or quirk for one of the characters. I was pickier about electronic concepts. An early decision was to exclude binary/two's complement logic from the game - I had plenty of interesting concepts without them, and I was personally sick of games that taught binary. Some of the implementation details of real processors are abstracted away as a result of this - for example, the way that instructions are packed into a single word of memory, or how real input/output selectors (multiplexers) select on individual bits. I don't think it's any great loss. Some other ideas would have been interesting, but didn't fit in for scale reasons. Branch prediction would be the logical next concept to introduce after the game's current ending, but would require extremely large programs - difficult to read and annoying long to run through. The memory hierarchy and caching would also be interesting, but would again slow testing down enormously. Power consumption and heat dissipation are the big concerns of modern processor design, but Silicon Zeroes doesn't cover them at all. Probably they would involve rearranging the modules to avoid dumping too much heat in one place while still trying to not use up too much space overall, maybe also positioning heat sinks...? I worry that adding those kinds of geometric/positioning incentives would make the game less fun by making 'optimized' machines completely visually incomprehensible, but it's possible that a mechanic like this could show up in some kind of expansion or sequel, if the game does well enough to earn one. Likewise, interrupts would be a fun twist to the mechanics, but would require quite a lot of UI work to fit in. Someday, perhaps!
|
# ? Sep 16, 2017 17:08 |
Speaking of memory and cacheing, does that mean no ? e: You really can't say that this: wouldn't be fun to do
|
|
# ? Sep 16, 2017 18:26 |
|
Watermelon Daiquiri posted:Speaking of memory and cacheing, does that mean no It would be fun, in principle! Like I said, though, the trouble is the timescales: cache misses are going to mean order-of-magnitude slowdowns, which means an order-of-magnitude increase in sim runtime. Also, responding to that diagram with "it would be fun" is probably the second-nerdiest thing I've ever done, right after designing this entire game.
|
# ? Sep 17, 2017 00:09 |
I haven't played the game, but what about pausing the machine/ff-ing to the point when the cache would fill? I don't know how feasible this would be or if it'd even fit in with the game as is, but I'm thinking about it as a bonus condition-- you can try and use the cache to 'speed up' your answer, but watch out for the misses and collisions! It could drastically hurt your score, even if it worked out in the end or something. Hell, just implement an entire drat mips machine lol
|
|
# ? Sep 17, 2017 13:45 |
|
Watermelon Daiquiri posted:I haven't played the game, but what about pausing the machine/ff-ing to the point when the cache would fill? I don't know how feasible this would be or if it'd even fit in with the game as is, but I'm thinking about it as a bonus condition-- you can try and use the cache to 'speed up' your answer, but watch out for the misses and collisions! It could drastically hurt your score, even if it worked out in the end or something. Hell, just implement an entire drat mips machine lol I agree, pausing/fast-forwarding is what you'd have to do! The problem is, doing that is nontrivial. It'll be a lot easier to have this discussion once you've actually played the game - let's shelve it for, uh, the next 25 hours or so Also, I just realized I didn't link to the talk I gave on Wednesday, so here it is: https://www.youtube.com/watch?v=p82iThhtj9c
|
# ? Sep 17, 2017 16:22 |
Awesome post, although my eyes glazed over once I got to the branch prediction link, lol. That said, I did just figure out how two's complement works so net win! Oh, I don't know if this is considered a "bug" per se, but I did notice if you copy the Op Select puzzle into the Palette and paste it into Arith CPU puzzle it uses the two-input instead of the three-input you use on that puzzle. Not that it matters because it can't match against instructions that aren't used!
|
|
# ? Sep 18, 2017 04:14 |
|
Doom Goon posted:Awesome post, although my eyes glazed over once I got to the branch prediction link, lol. That said, I did just figure out how two's complement works so net win! Haha, yeah, that's Wikipedia technical writing for you. There are people who can make dense, technical topics accessible... but most of them aren't writing for wikipedia! The op selector thing is more the absence of a feature than anything. Op Selectors only 'upgrade' themselves in puzzles that add opcodes to the ones they're already using; puzzles that just have different opcodes they don't touch, since swapping out what ports do what would be do weird and unexpected things to their connections to other modules. I have a note to rethink this behavior at some point (you're not the first person it's bothered), but it's fiddly and small, so I haven't gotten to it yet.
|
# ? Sep 18, 2017 16:37 |
|
Oh yeah. Also, It's out!
|
# ? Sep 18, 2017 17:05 |
|
Just beat Arith CPU, what do I win?
|
# ? Sep 18, 2017 19:11 |
|
Elswyyr posted:Just beat Arith CPU, what do I win? The greatest prize of all: more puzzles!
|
# ? Sep 18, 2017 20:43 |
|
(Spoilers for Multiply) trying to do everything with LUTs in TIS-100 has broken me
|
# ? Sep 18, 2017 21:08 |
|
overeager overeater posted:(Spoilers for Multiply) trying to do everything with LUTs in TIS-100 has broken me This... is art.
|
# ? Sep 18, 2017 22:40 |
|
I didn't realize this was a PF game. Also, getting a quote from Zach of Zachtronics fame is so...right. I look forward to getting my rear end kicked by yet another game, but I'll have a good time doing it!
|
# ? Sep 18, 2017 23:54 |
Speaking of that: https://twitter.com/zachtronics/status/909840355784237056
|
|
# ? Sep 19, 2017 07:39 |
|
Threes is making my head hurt. I'm guessing this was your intent The concept introduced in board 3 was a nice little twist, but it does kind of feel like I'm still doing setup for the "real" puzzles to come. How many boards are there in total?
|
# ? Sep 19, 2017 08:55 |
|
Canine Blues Arooo posted:I didn't realize this was a PF game. Also, getting a quote from Zach of Zachtronics fame is so...right. TheOneAndOnlyT posted:Threes is making my head hurt. I'm guessing this was your intent There are five boards total, counting Asides. The last one is by far the hardest, as you might expect. Threes is a puzzle that's very easy for some people and very hard for most. One of those puzzles where something just has to 'click' for you to get it!
|
# ? Sep 19, 2017 17:46 |
|
Fuuuuuuuuck I just figured out the solution for Threes and now I have to sit at work and hopefully remember it for six more hours until I can get home and actually test it. On the upside, the fact that I'm sitting at work thinking about Silicon Zeroes puzzles means that you've got a great game on your hands
|
# ? Sep 19, 2017 19:28 |
|
very cool game, enjoying it a lot. threes and multiply are my favorite puzzles so far - i'm pretty stumped on how to get threes down to 8 modules from 10 and i don't even want to talk about my solution to multiply. one thing i found odd was that my solution for the final puzzle on board 3 is identical to my solution for the final puzzle on board 2, just with one extra part (a register taking the outputs of the instruction decoder before processing it). were they intended to be so similar? makes me feel like i missed something.
|
# ? Sep 19, 2017 20:33 |
|
Saphire_flames posted:very cool game, enjoying it a lot. threes and multiply are my favorite puzzles so far - i'm pretty stumped on how to get threes down to 8 modules from 10 and i don't even want to talk about my solution to multiply. one thing i found odd was that my solution for the final puzzle on board 3 is identical to my solution for the final puzzle on board 2, just with one extra part (a register taking the outputs of the instruction decoder before processing it). were they intended to be so similar? makes me feel like i missed something. Haven't solved it myself, but I think the main concept the puzzle is meant to get across is pipelining: by having each stage of execution separated by latches, the clock speed increases, since the stages operate concurrently instead of waiting for the whole CPU to stabilize before the next instruction.
|
# ? Sep 19, 2017 21:58 |
|
I can't believe the two inputs of the subtractor are called the "minuend" and the "subtrahend."
|
# ? Sep 21, 2017 04:20 |
|
nrook posted:I can't believe the two inputs of the subtractor are called the "minuend" and the "subtrahend." It's very strange to me that I've gotten exactly no complaints about that. Aside from you, and I know you in real life, so you don't count. Did you notice the names of the inputs to the Adder?
|
# ? Sep 21, 2017 05:34 |
well you have dividend--- it makes sense! (still haven't gotten around to getting the game, but at my next paycheck ill pick it up)
|
|
# ? Sep 21, 2017 06:54 |
|
I just wanna say I love this game. Been a big fan of all these engineering games since the beginning, and out of all of them this one has by far the best campaign progression I've seen, the iteration of small concepts that eventually comprise a larger system and the introduction and then recontextualization of concepts like briefly exploring the parallel architecture to demonstrate address collision before it needs to be used in the more complex pipelined cpu, and the couple of levels where the manual clock timing mechanic is introduced and then making it automatic once the player has proven they understand how that poo poo works, its all kinda brilliant. Silicon Zeroes is secretly the best educational game of 2017, educational in a way these kind of games usually aren't, and maybe more coherently presented than most textbooks on the subject. I usually like a bit more room for creative solutions in these games than I feel like SZ offers, I tended to go back and redo things in different ways in for instance Shenzen and Manufactoria. 61/68 deep in this game and not feeling that urge here, most of them are at or below ideal anyway. I'm not sure that's much of a criticism though, because I'm pretty sure more flexibility would hurt the educational arc. I'll pretty happily put this game away with 1/5th the time clocked as on Shenzen feeling positively about the experience. Anyway, dang, nice work on this one, good game.
|
# ? Sep 22, 2017 06:58 |
|
homeless snail posted:I just wanna say I love this game. Been a big fan of all these engineering games since the beginning, and out of all of them this one has by far the best campaign progression I've seen, the iteration of small concepts that eventually comprise a larger system and the introduction and then recontextualization of concepts like briefly exploring the parallel architecture to demonstrate address collision before it needs to be used in the more complex pipelined cpu, and the couple of levels where the manual clock timing mechanic is introduced and then making it automatic once the player has proven they understand how that poo poo works, its all kinda brilliant. Silicon Zeroes is secretly the best educational game of 2017, educational in a way these kind of games usually aren't, and maybe more coherently presented than most textbooks on the subject. Thank you! The way you wrote it in that first paragraph perfectly sums up what I was hoping to achieve with the game - I'm really glad that it came across in play! Agreed that most of the puzzles don't have as much room for wiggling as other genre entries. Still, there are people doing some very clever stuff with some of the puzzles - I'm going to ship an update with Shenzhen-style leaderboards in a few days, and hopefully that'll encourage some people to go back and compete a bit more!
|
# ? Sep 22, 2017 20:39 |
|
I must be missing something extremely obvious, but I'm not sure how Parallels is possible without a Register. I want to pass - to my Writers in the first clock period while my other components are pulling up instructions, but there aren't any modules available that can output or store it for just one period. Am I being really dumb?
|
# ? Sep 23, 2017 19:34 |
do you have flip-flops? e: I've also tried the demo, and while I'm slightly disappointed that its mostly tutorial and nothing that requires more than a minutes thought, It promises something amazing. What's the highest difficulty rating (assuming thats what the stars are?) Watermelon Daiquiri fucked around with this message at 20:58 on Sep 23, 2017 |
|
# ? Sep 23, 2017 20:43 |
|
Watermelon Daiquiri posted:do you have flip-flops?
|
# ? Sep 23, 2017 20:56 |
sorry, flip flop is another name for a latch e: just to clarify, what is it you are trying to store? Does '-' mean subtraction?
|
|
# ? Sep 23, 2017 20:58 |
|
Watermelon Daiquiri posted:sorry, flip flop is another name for a latch
|
# ? Sep 23, 2017 21:04 |
|
TheOneAndOnlyT posted:I must be missing something extremely obvious, but I'm not sure how Parallels is possible without a Register. I want to pass - to my Writers in the first clock period while my other components are pulling up instructions, but there aren't any modules available that can output or store it for just one period. Am I being really dumb? The absence of a register is extremely intentional, to prevent you from doing exactly what you're trying to do. For what you can do: think about the name of the level and the way that the tick limit is phrased. Watermelon Daiquiri posted:e: I've also tried the demo, and while I'm slightly disappointed that its mostly tutorial and nothing that requires more than a minutes thought, It promises something amazing. What's the highest difficulty rating (assuming thats what the stars are?) Four stars. There's some fun ones waiting for you!
|
# ? Sep 24, 2017 00:03 |
|
PleasingFungus posted:The absence of a register is extremely intentional, to prevent you from doing exactly what you're trying to do.
|
# ? Sep 24, 2017 00:55 |
Oh, the null. Is the null supposed to represent a hi-z-state?
|
|
# ? Sep 24, 2017 16:41 |
|
Watermelon Daiquiri posted:Oh, the null. Is the null supposed to represent a hi-z-state? I'm going to give that a tentative 'yes', assuming that I read the wikipedia page correctly. It's equivalent to not having the port wired to anything.
|
# ? Sep 24, 2017 17:33 |
|
I picked up this game and have played it so far much like I played TIS-100 and Shenzhen - it's a ton of fun and then suddenly, I am completely stuck! This game is super-cool, though it's made me realize to what extent the whole timing/delay thing messes with my head. EDIT: my brain just kind of completely shuts off when I get to the third board. Something about managing clock ticks and timings to hold onto information between objects is really confusing to me... "Swap" is the first puzzle that just has me totally stymied. I understand that I only get one Reader, so I need to build a counter-type thing to give it 0 for a while and then 1 for a while, but I keep ending up with it returning the output from memory for such a short amount of time that the inputs on the next objects don't get a chance to stabilize. DOUBLE EDIT: Once you get to board 3, it seems like you just plain can't have a clock speed lower than 13 while there's a Writer on the board because the input values initialize to "?" instead of "-". Am I understanding that correctly? Ignoranus fucked around with this message at 18:45 on Sep 25, 2017 |
# ? Sep 25, 2017 18:24 |
From what i played in the demo, the '?' is a race condition where an input is an impossibility, such as depending on an output further along in the circuit. Are there any shift registers in this?
|
|
# ? Sep 25, 2017 18:57 |
|
I bought your game. It's great. Thanks! Also cool to see new things being added quickly, like the online and friend list score comparisons that just showed up.
|
# ? Sep 25, 2017 20:14 |
|
|
# ? May 8, 2024 03:46 |
|
Ignoranus posted:I picked up this game and have played it so far much like I played TIS-100 and Shenzhen - it's a ton of fun and then suddenly, I am completely stuck!
|
# ? Sep 25, 2017 20:55 |