mirror of
https://github.com/Bubberstation/Bubberstation.git
synced 2026-08-22 20:48:56 +01:00
ok-behavior-trees
6
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
|
0716e3fff4 |
Atmos refactor & speedup by utilizing BYOND 516 vector functions (#96448)
## About The Pull Request ### Summary This PR changes internal structure of `/datum/gas_mixture`: `gases[gas_id][MOLES]` refactored into `moles[gas_id]`, `gases[gas_id][ARCHIVE]` into `moles_archive[gas_id]` and `gases[gas_id][GAS_META]` into `gas_meta` static variable. This allows us to use BYOND 516 vector functions for calculating total moles and heat capacity. Also it simplifies some parts of the code, allowing us to get rid of macros `ADD_GAS` and `ASSERT_GAS`. According to the profiler `/turf/open/process_cell` time is reduced by ~20%. ### Details `gas_mixture.gases` was a nested 2d-list with MOLES, ARCHIVE and GAS_META for each gas_id. For example, to get gas moles you had to do `gases[gas_id][MOLES]`. I've changed this structure to be as follows: `moles[gas_id]` - moles for the gas, `moles_archive[gas_id]` - archived version of moles, `gas_meta[KEY][gas_id]` - static var with meta information for the gas. Since I removed key GAS_META from the gases, `gas_meta` was moved to the static variable and the order of keys in the array was changed from `gas_meta[gas_id][META_KEY]' to 'gas_meta[META_KEY][gas_id]`. This was done to allow using it in vector calculations (for example heat capacity or fusion power). Static variable access is very fast and it is considered as accessing a global in the bytecode. Byond 516 introduced new vector functions: `values_sum`, `values_dot` and others. These functions are very fast, but operate only on associative lists. This allows us to change the way we calculate total_moles and heat_capacity - very hot and heavily used functions. `total_moles()` became just `values_sum(moles)`, and `heat_capacity` is just a dot product: `values_dot(moles, gas_meta[META_GAS_SPECIFIC_HEAT])`. As a side bonus, since `moles` is just an associative list, you don't really need old macros `ADD_GAS` and `ASSERT_GAS` - all they did was to make a copy of a list[3] with default value [0, 0, gas_meta] for specific gas. Now when you're adding gas you can just use `moles[gas_id] += amount` and when you query amount of gas you can just query the key (for example `moles[/datum/gas/oxygen]`) if the key does not exist, it returns null and works as 0 for all arithmetic and logic operations. For example, old code would be `if (!air.gases[/datum/gas/oxygen] || air.gases[/datum/gas/oxygen][MOLES] < 1)` and now it is `if (moles[/datum/gas/oxygen] < 1)`. This simplifies some parts of the code and also speeds things up. For the performance comparison I used Tracy profiler. I've done many different tests, and they all show slightly different numbers, but overall speedup for `process_cell` is about 20%. (-20% to average time per call from ). My testing setup was as follows: Load Icebox, drop 30/60/90 radius bomb in the middle of the bridge, set code to blue, wait 10 minutes until the round ends. Also I fixed random seed in the master controller and in the planetary gas randomization so generated maps are the same between tests. Althought it's not very realistic, it generates a lot of samples for the `process_cell` (around ~3.5M per 10 minutes). Another test I did was a plasmafire in an 8x8 space, on runtime station, it showed (-24% time on process_cell). Another test was a emagged holodeck burn test, it showed (-13% time) As for other functions of gas_mixture: `total_moles`: -50%(2x speedup), `heat_capacity`: -65%(3x speedup), `share`: -30%, `react`: -20%. Timings of all those functions is in microseconds range and they are very hot (call count is in the same order as process_cell) <details><summary>Some pictures from profiler</summary> <img width="569" height="642" alt="process_cell" src="https://github.com/user-attachments/assets/76fa0c27-719d-485d-9bfc-859fef788999" /> <img width="572" height="315" alt="image" src="https://github.com/user-attachments/assets/f68497e9-4db8-4a9c-b43f-ad04e6dc5cac" /> <img width="569" height="317" alt="image" src="https://github.com/user-attachments/assets/d7249e7b-f344-47a6-8b39-1bab0521182d" /> <img width="541" height="316" alt="image" src="https://github.com/user-attachments/assets/1bef71fd-533d-40fa-a85e-7a803ad322f7" /> <img width="519" height="409" alt="image" src="https://github.com/user-attachments/assets/51165289-174e-403d-a09c-787dd9af136a" /> <img width="523" height="318" alt="image" src="https://github.com/user-attachments/assets/fe4b1db0-17d8-47f9-8fd0-e2ecef2ee66a" /> </details> <details><summary>Setting up a profiler</summary> If you wanna to reproduce my results here is a list of steps 1. download: https://github.com/goonstation/byond-tracy-writer (this one has offsets for my version 1677) 2. build the dll, drop in the tgstation/ folder 3. download rtracy https://github.com/Dimach/rtracy 4. download Tracy profiler (0.13.1) https://github.com/wolfpld/tracy 5. uncomment `#define USE_BYOND_TRACY` in `_compile_options.dm` 6. build tgstation 7. open dream daemon, run the desired test, after round end dream daemon closes 8. navigate to tgstation/data/profiler, find the `123412341234.utracy` file 9. run `rtracty 123412341234.utracy` 10. open tracy-profiler.exe, press Connect, save the profiler data 11. repeat steps 5-10 with another branch, save another profiler data 12. open tracy-profiler, open first data, press compare, open second data </details> ## Why It's Good For The Game ## Changelog 🆑 refactor: Atmos refactor & speedup by utilizing BYOND 516 vector functions /🆑 --------- Co-authored-by: san7890 <the@san7890.com> |
||
|
san7890andGitHub
|
265ceb3eb2 |
Fixes Flakey Ashwalker Lung Unit Test Failures on gateway_test (#94821)
## About The Pull Request
Closes #94794
Both of these unit tests were using the outdated way of changing the gas
mix on a turf, probably because they were more than a few years old
apiece. The modern way is using the `/datum/gas_mixture` and
`parse_gas_string()` to either retrieve the gas mix from cache or
generate it on-demand. Neither of these tests were doing that, and they
worked well enough until they got mutated in #94771
(
|
||
|
FlufflesTheDogandGitHub
|
01c7ef7de1 |
Stops gateway CI from generating lavaland and space ruins (#94771)
## About The Pull Request Lightens up the Gateway CI, as it's a very memory-intensive test that creeps ever closer to the 4gb limit. ## Why It's Good For The Game Functional and consistent CI is good ## Changelog N/A |
||
|
GhomandGitHub
|
778ed9f1ab |
The death or internal/external organ pathing (ft. fixed fox ears and recoloring bodypart overlays with dye sprays) (#87434)
## About The Pull Request This PR kills the abstract internal and external typepaths for organs, now replaced by an EXTERNAL_ORGAN flag to distinguish the two kinds. This PR also fixes fox ears (from #87162, no tail is added) and mushpeople's caps (they should be red, the screenshot is a tad outdated). And yes, you can now use a hair dye spray to recolor body parts like most tails, podpeople hair, mushpeople caps and cat ears. The process can be reversed by using the spray again. ## Why It's Good For The Game Time-Green put some effort during the last few months to untie functions and mechanics from external/internal organ pathing. Now, all that this pathing is good for are a few typechecks, easily replaceable with bitflags. Also podpeople and mushpeople need a way to recolor their "hair". This kind of applies to fish tails from the fish infusion, which colors can't be selected right now. The rest is just there if you ever want to recolor your lizard tail for some reason. Proof of testing btw (screenshot taken before mushpeople cap fix, right side has dyed body parts, moth can't be dyed, they're already fabolous):  ## Changelog 🆑 code: Removed internal/external pathing from organs in favor of a bit flag. Hopefully this shouldn't break anything about organs. fix: Fixed invisible fox ears. fix: Fixed mushpeople caps not being colored red by default. add: You can now dye most tails, podpeople hair, mushpeople caps etc. with a hair dye spray. /🆑 |
||
|
|
2e5bfe5be6 |
Refactors and optimizes breath code (Saves 12% of carbon/Life()) (#74230)
## About The Pull Request ### How things work As things currently stand, when a mob breaths several things happen (simplified to focus on the stupid) We assert the existance of all possible breathable gases, and pull partial pressures for them Then we walk through all possible interactions lungs could have with these gases, one by one, and see if they're happening or not As we go we are forced to cleanup potential alerts caused by the previous breath, even if those effects never actually happen At the end we clear out all the unused gas ids, and handle the temperature of the breath. ### What sucks There's I'd say 3 different types of gas reactions. - You can "need" a gas to survive. o2, n2 and plasma all fall into this category - A gas can do something to you while it's in your system. This applies to most gas types - Variation on the previous, some gases do cleanup when they're not in your system, or when there isn't much of them in the first place The main headache here is that second one, constantly cleaning up potential side effects sucks, and fixing it would require a lot of dummy variables There's other suckage too. Needing to constantly check for a gas type even if it isn't there is stupid, and leads to wasted time It's also really annoying to do subtypes in this system. There is what amounts to a hook proc you can override, but you can't override the reaction to a gas type. It also just like, sucks to add new gases. one mega proc smells real stupid. ### Improvements In the interest of speed: - I'd like to build a system that doesn't require manually checking for gas - Reacting to gas "disappearing" should be promoted by the system, instead of being hacky. - I would like to avoid needing to assert the existence of all possible gases, as this is slow on both the assert and the garbage collect. In the interest of dev ergonomics: - It should be easy to define a new gas reaction - It should be easy for subtypes to implement their own gas reactions. The current method of vars on the lung is all tangled up and not really undoable as of now, but I'd like to not require it - It should be possible to fully override how a gas is handled ### What I've Done Lungs have 3 lists of proc paths stored on them Each list handles a different way the lung might want to interact with a gas. There's a list for always processing on a gas (we use this for stuff that's breathed), a list for handling a gas in our breath, and a list for reacting to a gas previously being in our breath, but not any more. Lungs fill out these lists using a helper proc during Initialize() Then, when it comes time to breath, we loop over the gas in the breath and react to it. We also keep track of the previous list of partial pressures, which we calculate for free here, and use that to figure out when to call the loss reactions. This proc pattern allows for overrides, easy reactions to removals, lower indentation code and early returns, and better organization of signal handlers It's also significantly faster. Ballpark 4x faster ### Misc Removes support for breathing co2, and dying from n2 poisoning. They were both unused, and I think it's cringe to clutter these procs even further Added "do we even have oxyloss" checks to most cases of passive breathing. This is a significant save, since redundant adjustoxy's are decently expensive at the volume of calls we have here. Fixes a bug with breathing out if no gas is passed in, assigning a var to another var doesn't perform a copy Rewrote breathe_gas_volume() slightly to insert gas into an immutable mix stored on the lung, rather then one passed in This avoids passing of a gas_mixture around just to fill a hole. I may change my mind on this, since it would be nice to have support for temperature changing from a hot/cold breath. Not gonna be done off bodytemp tho lord no. Uses merge() instead of a hard coded version to move the gas ids over. This is slightly slower with lower gas counts but supports more things in future and is also just easier to read. ## Why It's Good For The Game Faster, easier to work with and read (imo) Profiles: [breath_results_old.txt](https://github.com/tgstation/tgstation/files/11068247/breath_results_old.txt) [breath_results_pre_master.txt](https://github.com/tgstation/tgstation/files/11068248/breath_results_new.txt) [breath_results_new.txt](https://github.com/tgstation/tgstation/files/11068349/breath_results_new.txt) (These profiles were initially missing #73026. Merging this brings the savings from 16% to 12%. Life is pain) --------- Co-authored-by: san7890 <the@san7890.com> |
||
|
Dani GloreandGitHub
|
f9fe79a307 |
Organ Unit Tests & Bugfixes (#73026)
## About The Pull Request This PR adds a new unit test for all organs, a new unit test for lungs, and includes improvements for the existing breath and organ_set_bonus tests. Using the tests, I was able to root out bugs in the organs. This PR includes an advanced refactor of several developer-facing functions. This PR certainly represents a "quality pass" for organs which will make them easier to develop from now on. ### Synopsis of changes: 1. Fixed many fundamental bugs in organ code, especially in `Insert()`/`Remove()` and their overrides. 2. Added two new procs to `/obj/item/organ` named `on_insert` and `on_remove`, each being called after `Insert()`/`Remove()`. 3. Added `organ_effects` lazylist to `/obj/item/organ`. Converted `organ_traits` to lazylist. 2x less empty lists per organ. 4. Adding `SHOULD_CALL_PARENT(TRUE)` to `Insert()`/`Remove()` was very beneficial to stability and overall code health. 5. Created unit test `organ_sanity` for all usable organs in the game. Tests insertion and removal. 6. Created unit test `lungs_sanity` for `/obj/item/organ/internal/lungs`. 7. Improved `breath_sanity` unit tests with additional tests and conditions. 8. Improved `organ_set_bonus_sanity` unit tests with better documentation and maintainable code. --- ### Granular bug/fix list: - A lot of organs are overriding `Insert()` to apply unique side-effects, but aren't checking the return value of the parent proc which causes the activation of side-effects even if the insertion technically fails. I noticed the use-case of applying "unique side-effects" is repeated across a lot of organs in the game, and by overriding `Insert()` the potential for bugs is very high; I solved this problem with inversion-of-control by adding two new procs to `/obj/item/organ` named `on_insert` and `on_remove`, each being called after `Insert()` and `Remove()` succeed. - Many organs, such as abductor "glands", cursed heart, demon heart, alien hive-node, alien plasma-vessel, etc, were not returning their parent's `Insert()` proc return value at all, and as a result those organs `Insert()`s were always returning `null`. I have been mopping those bugs up in my last few PRs, and now the unit test reveals it all. Functions such as those in surgery expect a truthy value to be returned from `Insert()` to represent insertion success, and otherwise it force-moves the organ out of the mob. - Fixed abductor "glands" which had a hard-del bug due to their `Remove()` not calling the parent proc. - Fixed cybernetic arm implants which had a hard-del bug due to `Remove()` not resetting their `hand` variable to `null`. - Fixed lungs gas exchange implementation, which was allowing exhaled gases to feedback into the inhaled gases, which caused Humans to inhale much more gas than intended and not exhale expected gases. ### Overview of the `organ_sanity` unit test: - The new `organ_sanity` unit test gathers all "usable" organs in the game and tests to see if their `Insert()` and `Remove()` functions behave as we expect them to. - Some organs, such as the Nightmare Brain, cause the mob's species to change which subsequently swaps out all of their organs; the unit test accounts for these organs via the typecache `species_changing_organs`. - Some organs are not usable in-game and can't be unit tested, so the unit test accounts for them via the typecache `test_organ_blacklist`. ### Overview of the `lungs_sanity` unit test: - This unit test focuses on `/obj/item/organ/internal/lungs` including Plasmaman and Ashwalker lungs. The test focuses on testing the lungs' `check_breath()` proc. - The tests are composed of calling `check_breath` with different gas mixes to test breathing and suffocation. - Includes gas exchange test for inhaled/exhaled gases, such as O2 to CO2. ### Improvements to the `breath_sanity` unit tests: - Added additional tests for suffocation with empty internals, pure Nitrogen internals, and a gas-less turf. - Includes slightly more reliable tests for internals tanks. ## Why It's Good For The Game **Organs and Lungs were mostly untested. Too many refactors have been submitted without the addition of unit tests to prove the code works at all.** Time to stop. _Time to get some help_. Due to how bad the code health is in organs, any time we've tried to work with them some sort of bug caused them to blow up in our faces. I am trying to fix some of that by establishing some standard testing for organs. These tests have revealed and allowed me to fix lot of basic developer errors/oversights, as well as a few severe bugs.  ## Changelog 🆑 A.C.M.O. fix: Fixed lungs gas exchange implementation, so you always inhale and exhale the correct gases. fix: Fixed a large quantity of hard-deletes which were being caused by organs and cybernetic organs. fix: Fixed many organs which were applying side-effects regardless of whether or not the insertion failed. code: Added unit tests for Organs. code: Added unit tests for Lungs. code: Improved unit tests for breathing. code: Improved unit tests for DNA Infuser organs. /🆑 |
