4 ms·

I have posted this before, but I’ve been working on a wrapper for OpenTTD to turn it from a game to a (slightly!) more serious system for research/experimentati

I have posted this before, but I’ve been working on a wrapper for OpenTTD to turn it from a game to a (slightly!) more serious system for research/experimentation, especially using its AI system: https://github.com/michalc/OpenTTDLab https://github.com/michalc/OpenTTDLab

teruakohatu 2y ago
This is really interesting. Submit it as a Show HN.
- michalc 2y ago
  Ah I did! https://news.ycombinator.com/item?id=39254487 https://news.ycombinator.com/item?id=39254487 (Feels a bit soon to repost somehow…)
  - hyperman1 2y ago
    I am looking at https://github.com/michalc/OpenTTDLab/blob/main/examples/02-openttdlab-scaling.ipynb https://github.com/michalc/OpenTTDLab/blob/main/examples/02-... and wondering if this isn't hyperthreading being enabled. If you have 4 real cores, with 2 hyperthreads, the system reports 8 CPUs. But that's fake. Hyperthreads share most of their compute and register resources, so they only work well if a core is spending most of its time waiting. 2 hyperthreads both doing full compute will basically work at half speed each, which is close to what your graph is showing. BTW, you can check on Linux with this command LANG=C lscpu and checking the Thread(s) per core value
    - MoreMoore 2y ago
      I wonder if the plot would look similar if you disabled hyperthreading but kept the same number of workers.
      - michalc 2y ago
        Ah I _think_ I ran this on a MacBook that as far as I know has 8 real cores… But I since found that 4 are “efficiency” cores, so that could be the reason for the poor scaling(?)
        MoreMoore 2y ago
        Seems plausible to me. The efficiency cores seem to be clocked quite a bit lower.
scyzoryk_xyz 2y ago
I’ve been an avid OTTD player for over a decade, and always wondered how someone could do exactly this sort of thing. Some of the addons, like the FIRS industrial economy are quite complex and it can be tricky to assess or estimate optimal routes/investment.
rantouan 2y ago
very good !
indigoabstract 2y ago
I wish you would have given some concrete examples as I was curious what experiments are possible with your code in OpenTTD. There was only one I could find: "exploring the performance of trAIns". I looked at the "trAIns" link and cannot understand what it does, other than being GPL licensed. It's all very cryptic, without any explanations. To me that stands for "if you don't know already, you need not apply". Too bad.
- michalc 2y ago
  Ah thanks for the suggestion! Will see if I can add examples or otherwise make things clearer
  - indigoabstract 2y ago
    Thanks!

teruakohatu 2y ago

This is really interesting. Submit it as a Show HN.

michalc 2y ago
Ah I did! https://news.ycombinator.com/item?id=39254487 https://news.ycombinator.com/item?id=39254487 (Feels a bit soon to repost somehow…)
- hyperman1 2y ago
  I am looking at https://github.com/michalc/OpenTTDLab/blob/main/examples/02-openttdlab-scaling.ipynb https://github.com/michalc/OpenTTDLab/blob/main/examples/02-... and wondering if this isn't hyperthreading being enabled. If you have 4 real cores, with 2 hyperthreads, the system reports 8 CPUs. But that's fake. Hyperthreads share most of their compute and register resources, so they only work well if a core is spending most of its time waiting. 2 hyperthreads both doing full compute will basically work at half speed each, which is close to what your graph is showing. BTW, you can check on Linux with this command LANG=C lscpu and checking the Thread(s) per core value
  - MoreMoore 2y ago
    I wonder if the plot would look similar if you disabled hyperthreading but kept the same number of workers.
    - michalc 2y ago
      Ah I _think_ I ran this on a MacBook that as far as I know has 8 real cores… But I since found that 4 are “efficiency” cores, so that could be the reason for the poor scaling(?)
      - MoreMoore 2y ago
        Seems plausible to me. The efficiency cores seem to be clocked quite a bit lower.

scyzoryk_xyz 2y ago

I’ve been an avid OTTD player for over a decade, and always wondered how someone could do exactly this sort of thing. Some of the addons, like the FIRS industrial economy are quite complex and it can be tricky to assess or estimate optimal routes/investment.

rantouan 2y ago

very good !

indigoabstract 2y ago

I wish you would have given some concrete examples as I was curious what experiments are possible with your code in OpenTTD. There was only one I could find: "exploring the performance of trAIns". I looked at the "trAIns" link and cannot understand what it does, other than being GPL licensed. It's all very cryptic, without any explanations. To me that stands for "if you don't know already, you need not apply". Too bad.

michalc 2y ago
Ah thanks for the suggestion! Will see if I can add examples or otherwise make things clearer
- indigoabstract 2y ago
  Thanks!