|
You should print it out and make notes on it. I'm not joking, there's lots of stuff that (much like the real world) isn't mentioned in the datasheets. And once it's out of early access, you won't have to worry about those undocumented behaviors changing in the next batch of parts!
|
#
¿
Oct 10, 2016 04:10
|
|
|
|
| # ¿ May 13, 2024 23:12 |
|
|
echinopsis posted:
that one's a bit obnoxious you get power efficiency by not doing stuff unless you need to if you already know one condition has failed, you don't need to check the other Jabor fucked around with this message at 08:17 on Oct 10, 2016 |
|
#
¿
Oct 10, 2016 08:09
|
|
|
Sagebrush posted:how do you discover undocumented instructions? are you just supposed to guess based on what you remember from assembly programming class, or does it come up in the story somehow? read your emails (yeah it's the latter)
|
|
#
¿
Oct 10, 2016 08:49
|
|
|
or if you can read chinese you can get all the deets on the logic array chip instead of having to figure it out yourself
|
|
#
¿
Oct 10, 2016 08:52
|
|
|
in it, voted 5
|
|
#
¿
Oct 11, 2016 11:12
|
|
|
for real though, don't beat yourself up comparing your first solution to everyone else's best solution. lots of my first solutions were utter trash, and then I had some insight hours later while doing something totally unrelated that let me optimize the hell out of things
|
|
#
¿
Oct 11, 2016 11:18
|
|
|
Bonfire Lit posted:damnit jabor now I have to figure out how to shave 4 more power off my solution as always, the answer is just to cheat horribly actually, i don't even know if it still works since the update
|
|
#
¿
Oct 11, 2016 13:48
|
|
|
Condiv posted:the immersion is you're an ee making poo poo electronics for chinese manufacturers saving money by not actually meeting the spec, just making something that passes the test cases, is not only permitted but actively encouraged
|
|
#
¿
Oct 12, 2016 09:37
|
|
|
echinopsis posted:big frustration is testing to see if an xbus signal is somethig worth listening to, but then by testing it you lose it, so I always have to copy it to dat and so I need a whole register just for that. I know this is hugely inefficient but I can't see a better solution? yeah it blows the most expedient way is to have a separate MC4k that does the listening, and only forwards it if it's relevant. then your real code can just listen for it using slx. you can look into merging them later if you're gunning for lowest cost.
|
|
#
¿
Oct 13, 2016 10:49
|
|
|
Stereotype posted:This doesn't make sense to me, you can't ever end up with acc==0 in any case with that code, only acc==1 and acc==-999. With the tcp command, if the two arguments are equal then neither branch will be executed, the same way it is when things are first initialized.
|
|
#
¿
Oct 15, 2016 02:51
|
|
|
Sagebrush posted:e: aahhh gently caress you jabor, this always happens. i overexplain and someone beats me with a more cryptic but shorter answer fwiw, i think people appreciate the effort
|
|
#
¿
Oct 15, 2016 02:56
|
|
|
teq and tcp are completely different instructions
|
|
#
¿
Oct 15, 2016 06:15
|
|
|
tgt acc 2 + sub 8 
|
|
#
¿
Oct 16, 2016 09:47
|
|
|
|
| # ¿ May 13, 2024 23:12 |
|
|
big scary monsters posted:not asking for solutions, but do you need ops beyond what you start with to get down to 3 lines on the early problems? it seems like i'd need to do two things on one line at the moment to get there yeah, if you read your in-game emails you'll eventually find out about a few bonus features of the microprocessors you're using some of the later levels would be really tough without them!
|
|
#
¿
Dec 27, 2016 12:04
|
|