| --- Log opened Wed Nov 26 00:00:44 2014 | ||
| olofk | Wow... http://motherboard.vice.com/read/gods-lonely-programmer | 11:52 |
|---|---|---|
| stekern | I'll install that on my wife's computer | 14:07 |
| Limb | Read that yesterday on HN. Impressive work for one person | 15:11 |
| poke53282 | stekern: You might want to give her just a homebrew computer then. Take a lookt at: http://www.mycpu.eu/ | 17:09 |
| poke53282 | I met this guy once on a exhibition. | 17:10 |
| poke53282 | The CPU is similar to a MOS6502 (including a mul and div instruction) and was built fully with standard electronic parts. | 17:12 |
| poke53282 | The magic here lies in the eeprom. What he does is mostly table lookups. For an 8 Bit Data Bus, this is possible with a few MB. | 17:14 |
| stekern | yeah, well, the reason I intended to install it on my wife's computer was more because of ideological reasons | 17:15 |
| stekern | this is way cooler if you're going to have a homebrew cpu from standard parts anyway: http://www.nablaman.com/relay/ | 17:16 |
| Limb | That machine would drive people mad after hearing those relays for a while | 17:18 |
| poke53282 | :) But this might be a little bit too slow for me. Some pages of the mycpu project are even hosted on the CPU. | 17:18 |
| poke53282 | If the sound is homogeneous, I might not care. | 17:19 |
| poke53282 | Like a fan. | 17:20 |
| Limb | I wonder how fast the clock speed is | 17:20 |
| stekern | 1 Hz iirc | 17:20 |
| poke53282 | There is a video | 17:21 |
| Limb | Yea I was watching it hehe | 17:21 |
| poke53282 | I would say 10Hz | 17:21 |
| poke53282 | But might be indeed only 1 Hz. | 17:22 |
| poke53282 | A little bit bulky for my home. | 17:25 |
| poke53282 | The MyCPU project uses even microcode. I was very fascinated about the design, because this was the first computer which I understood on the logic level. | 17:28 |
| poke53282 | and just in case someone hasn't seen this yet: http://www.visual6502.org/JSSim/index.html | 17:40 |
| jakobkn | what kind of associativity does the or1200 tlb support? | 20:02 |
| stekern | jakobkn: 1-way | 20:21 |
| jakobkn | ok, thanks | 20:37 |
| jakobkn | what's the reasoning for not going fully associative? | 20:42 |
| wallento1 | the comparators I suppose | 20:47 |
| jakobkn | that means FPGA resource limitations? | 21:00 |
| jakobkn | My textbooks say 1-way associativity is horrible for performance. How is it working out in the openrisc core? | 21:02 |
| -!- knz_ is now known as knz | 22:04 | |
| olofk | Ahhh!!! Git, I hate you! | 22:08 |
| stekern | jakobkn: yes the trade-off is more complicated hardware | 22:13 |
| stekern | jakobkn: if you want an example of an implementation of a fully associative tlb, here's one: https://github.com/skristiansson/eco32f/blob/master/rtl/verilog/eco32f_mmu.v | 22:19 |
| olofk | Anyone knows what to do here? https://github.com/olofk/fusesoc/issues/61 | 22:28 |
| olofk | I don't know why the stack protector is yelling at the VPI stuff | 22:29 |
| olofk | And why does it do that at link time? Isn't that just a switch to enable some compile-time stuff? | 22:29 |
| --- Log closed Thu Nov 27 00:00:46 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!