Everything is working great in this build! So I will be taking about a week break from this. Well except single pixel drawing on GPU but more on that later. I included a Hello World program and a screen clear test program running by default.But I would love to see what programs people can come up with. So definitely fork this and make your own programs for it, and link them in the comments below!
Also I am working on adding more instructions for the cpu and the gpu, let me know what might be useful in the comments and definitely check my in progress updates below!
I am running into an issue with per pixel drawing giving a contention error which doesn't make sense as i have tested the chip. unless the rgb screen col and row pins are also outputs i don't get what is going on.
For now though i'm taking a break from that and working on implementing some more of the OPs.
Here is a list of instructions. I'll add descriptions soon. Anything with a * before it is not implemented yet.0 NOP 1 JMP 2 SETA 3 SETB 4 LDA 5 LDB 6 WRA 7 WRB  *8 IA *9 IB *10 DA *11 DB *12 SLA *13 SLB *14 SRA *15 SRB  16 ADD [1 0000]17 SUB [1 0001]18 MUL [1 0010]19 DIV [1 0011] 20 EQL [1 0100]21 NEQL [1 0101]22 GRT [1 0110]23 LSS [1 0111]24 GRQL [1 1000]25 LEQL [1 1001] 26 SSP [1 1010]27 PUSH [1 1011]28 POP [1 1100]29 LTS [1 1101] *30 PROC [1 1110]*31 EPROC [1 1111] 32 SAS [10 0000]
Feb 22, 2020
Apr 18, 2021
Enter Email IDs separated by commas, spaces or enter. Users need to be registered already on the platform. Note that collaboration is not real time as of now. Every save overwites the previous data.
Super impressive build! I step away from the site for a couple of months, just to be seriously one upped!
I may be wrong (and I usually am most of the time), but I've run into contention errors with RAM units before too, especially when there's any feedback or recursion going on between it's input and output (I'm talking about the RGB output from your GPU. Not too sure if this is the case, but might be worth looking into). Probably a bug in the site. Anyway, the way I was able to deal with it was place a flip-flop between the output and whatever I was connecting it to, then couple the clk input to an inverter (you may need to either invert or not, depending on how you set up your clock system).
If I get some time off from school, I going to fork this and play around with it, because it's seriously cool, almost like the Pro-Athena build I saw awhile back.
I saw your comment on my project. I came to check yours, you are a bit further than me! Nicely done
I looked around and it turns out that your contention issue is due to the fact that in the side off the screen your X and Y lines are connected (it's hard to see since the multibit lines are the exact same as the edge of the screen. In order to see it, just delete both the X and the Y lines connecting to the screen and it'll turn red). After deleting that line it appeared to work fine. Hope this helps!
Also EEPROMs in sub circuits can be persistent but must be written to in the sub circuit itself. I'm personally going to use the EEPROM load data function to program my computer (I'll write an assembler using python which will convert it into the data).
I've messed around a bit further, just to see how your implementation works, and have managed to increasing the execution speed to 5120 instructions per second. The forked version is here: https://circuitverse.org/users/4699/projects/16-bit-computer-1c74dd25-0b96-4923-ad3e-7f8e7f5786f4
She is beautiful, what's it name? i say for something like this it needs a name.
made up name: 16-gamma
go to "clock" circuit, make everything a 0 except the first. (n) then you get a error. (stack error)
EDIT: about 13-15 will get the error
Noulmao, the input n is how many times the computer's clock line is pulsed when the circuit clock changes state. This allows the number of cycles the computer executes per second to be controlled, allowing the computer to be sped up or slowed down. Naturally, setting it too high will cause stack errors since it will attempt to update the computer too many times in a single pulse, resulting in Circuitverse's execution stack being filled, causing the error. This is perfectly intended behaviour, and is why the n input exists, since it allows us to limit the number of loops to a safe number. (If you're wondering why I know, this kind of overclocking is my tech).
Wow, how did you make the clock speed adjustable
WOW THIS IS FREAKING AMAZING!
Btw Go to my circuit then come back here And rate my circuit Pls text on May 7