3 commits

Author	SHA1	Message	Date
Igor Rončević	46e83a6fbe	Add --perf option to test to collect and diff performance data (#7459) ... ## Description This PR: - Adds `--perf` and `--perf-only` options to `test` to collect gas usages and bytecode sizes, and to run only the tests that actually produce those performance data, respectively. - Adds `[performance]` recipes to `justfile` that fully automate a workflow for collecting and comparing performance data of the compiler output. `--perf` flag instructs `test` to collect performance data from tests of categories `compile`, `run`, and `unit_tests_pass`. Collected gas usages and bytecode sizes are written to files named `<timestamp>-<category>-<build profile>-<branch>.csv`. E.g.: `1020165605-e2e-bytecode-sizes-release-master.csv`. When running tests in parallel (default option), performance data of each individual `test --exact <test_toml_path>` process is written to a piped stdout in JSON format. The main `test` process collects those JSON outputs from all spanned processes. The usual workflow of collecting and comparing performance data between the `master` and feature branch is supported by the following `just` recipes: ```console [performance] perf-e2e filter='' # collect gas usages and bytecode sizes from E2E tests [alias: pe2e] perf-in-lang filter='' # collect gas usages from in-language tests [alias: pil] perf-all filter='' # collect gas usages and bytecode sizes from all tests (E2E and in-language) [alias: pa] perf-diff before after format='md' # generate performance diff between two CSV files [alias: pd] perf-diff-latest format='md' # generate performance diffs between the latest two CSV files per testing category [alias: pdl] perf-snapshot-historical path open='' # collect historic gas usages from a snapshot test that has a `forc test` output [alias: psh] perf-list # list all performance files (*gas-usages-*.* and *bytecode-sizes-*.*) [alias: pl] perf-remove # remove all performance files (*gas-usages-*.* and *bytecode-sizes-*.*) [alias: pr] ``` To obtain the performance comparison it is enough to run: ```console just pa // `just perf-all`. On `master` to get the baseline. just pa // `just perf-all`. On the feature branch to get the improved data. just pdl // `just perf-diff-latest`. To get the diff of the last two collected performance data sets. ``` The `just` recipes should be executed from the root of the Sway repository. Aside from `--perf` writing to predefined files, all other parts of the workflow are `just` recipes and can potentially be tailored to any other workflow. Adding support for collecting gas usages and bytecode sizes from other tests like (snapshot, in-language, SDK harness) will be added in follow up PRs. ## Checklist - [ ] I have linked to any relevant issues. - [x] I have commented my code, particularly in hard-to-understand areas. - [ ] I have updated the documentation where relevant (API docs, the reference, and the Sway book). - [ ] If my change requires substantial documentation changes, I have [requested support from the DevRel team](https://github.com/FuelLabs/devrel-requests/issues/new/choose) - [ ] I have added tests that prove my fix is effective or that my feature works. - [ ] I have added (or requested a maintainer to add) the necessary `Breaking*` or `New Feature` labels where relevant. - [x] I have done my best to ensure that my PR adheres to [the Fuel Labs Code Review Standards](https://github.com/FuelLabs/rfcs/blob/master/text/code-standards/external-contributors.md). - [x] I have requested a review from the relevant team or maintainers.	2025-10-21 18:41:35 +04:00
Daniel Frederico Lins Leite	4e16bedab2	Improve Contract Call by not encoding fuelvm protocol args (#7447) ... ## Description This is an optimization of contract calls on top of https://github.com/FuelLabs/sway/pull/7440. ```csv const_of_contract_call::cost_of_in_array_0,1976,1285 const_of_contract_call::cost_of_in_array_1,3164,1779 const_of_contract_call::cost_of_in_array_16,5096,4753 const_of_contract_call::cost_of_in_array_32,8421,8079 const_of_contract_call::cost_of_in_array_64,60382,14545 const_of_contract_call::cost_of_in_array_8,4724,4382 const_of_contract_call::cost_of_in_b256,2491,2147 const_of_contract_call::cost_of_in_bool,2627,2279 const_of_contract_call::cost_of_in_str_0,3043,2697 const_of_contract_call::cost_of_in_str_1,3345,2998 const_of_contract_call::cost_of_in_str_16,3374,3030 const_of_contract_call::cost_of_in_str_32,3572,3227 const_of_contract_call::cost_of_in_str_8,3768,3421 const_of_contract_call::cost_of_in_tuple_0,3822,3479 const_of_contract_call::cost_of_in_tuple_1,4392,4050 const_of_contract_call::cost_of_in_tuple_2,4761,4418 const_of_contract_call::cost_of_in_tuple_3,5104,4761 const_of_contract_call::cost_of_in_tuple_4,5468,5126 const_of_contract_call::cost_of_in_u16,4493,4141 const_of_contract_call::cost_of_in_u256,4777,4433 const_of_contract_call::cost_of_in_u32,4926,4582 const_of_contract_call::cost_of_in_u64,4934,4589 const_of_contract_call::cost_of_in_u8,4874,4526 const_of_contract_call::in_enum_u64,3186,2843 const_of_contract_call::in_enum_u64_u64,3061,2718 const_of_contract_call::in_enum_u64_u64_u64,3173,2830 const_of_contract_call::in_struct_u64,3635,3293 const_of_contract_call::in_struct_u64_u64,3897,3554 const_of_contract_call::in_struct_u64_u64_u64,4156,3813 const_of_contract_call::isolated_cost_of_in_array_0,956,956 const_of_contract_call::isolated_cost_of_in_array_1,1515,1185 const_of_contract_call::isolated_cost_of_in_array_16,3955,3266 const_of_contract_call::isolated_cost_of_in_array_32,6544,5471 const_of_contract_call::isolated_cost_of_in_array_64,9914,9914 const_of_contract_call::isolated_cost_of_in_array_8,3225,2728 const_of_contract_call::isolated_cost_of_in_b256,1462,1120 const_of_contract_call::isolated_cost_of_in_bool,1358,1047 const_of_contract_call::isolated_cost_of_in_str_0,1451,1140 const_of_contract_call::isolated_cost_of_in_str_1,1547,1237 const_of_contract_call::isolated_cost_of_in_str_16,1552,1241 const_of_contract_call::isolated_cost_of_in_str_32,1553,1242 const_of_contract_call::isolated_cost_of_in_str_8,1551,1241 const_of_contract_call::isolated_cost_of_in_tuple_0,820,820 const_of_contract_call::isolated_cost_of_in_tuple_1,1412,1090 const_of_contract_call::isolated_cost_of_in_tuple_2,1615,1304 const_of_contract_call::isolated_cost_of_in_tuple_3,1766,1455 const_of_contract_call::isolated_cost_of_in_tuple_4,1916,1605 const_of_contract_call::isolated_cost_of_in_u16,1428,1117 const_of_contract_call::isolated_cost_of_in_u256,1431,1120 const_of_contract_call::isolated_cost_of_in_u32,1554,1243 const_of_contract_call::isolated_cost_of_in_u64,1336,1025 const_of_contract_call::isolated_cost_of_in_u8,1352,1041 ``` ## Checklist - [x] I have linked to any relevant issues. - [ ] I have commented my code, particularly in hard-to-understand areas. - [ ] I have updated the documentation where relevant (API docs, the reference, and the Sway book). - [ ] If my change requires substantial documentation changes, I have [requested support from the DevRel team](https://github.com/FuelLabs/devrel-requests/issues/new/choose) - [x] I have added tests that prove my fix is effective or that my feature works. - [ ] I have added (or requested a maintainer to add) the necessary `Breaking*` or `New Feature` labels where relevant. - [x] I have done my best to ensure that my PR adheres to [the Fuel Labs Code Review Standards](https://github.com/FuelLabs/rfcs/blob/master/text/code-standards/external-contributors.md). - [x] I have requested a review from the relevant team or maintainers.	2025-10-20 17:24:58 +00:00
Daniel Frederico Lins Leite	5896b4cac2	Cost of contract calls snapshot tests (#7440) ... ## Description This PR is a prequel for https://github.com/FuelLabs/sway/pull/7419. It will set the baseline, which future PRs need to improve. Apart from that it also: 1 - Fix a problem with `SubtsType` for `TyConstantDecl`; 2 - Introduce a new way to write snapshot tests. Now the test source code can have "blocks" ``` /* START BOOL */ fn in_bool(v: bool) -> bool; /* END BOOL */ ``` These blocks can be manipulated from inside the `snapshot.toml`. In this case, the `toml` is: ``` cmds = [ "forc test --path {root} --release --experimental const_generics", { repeat = "for-each-block", cmds = [ "forc test --path {root} --release --experimental const_generics" ] } ] ``` Which repeats the inner `forc test` for each block, removing all others. That allows me to check the cost of the contract method selection algorithm for each contract method. For example, `cost_of_in_array_64` with all other contract methods costs `60382`, but alone it only costs `58218`, which means that around 2000 was "wasted" in the contract method selection. ## Checklist - [x] I have linked to any relevant issues. - [x] I have commented my code, particularly in hard-to-understand areas. - [ ] I have updated the documentation where relevant (API docs, the reference, and the Sway book). - [ ] If my change requires substantial documentation changes, I have [requested support from the DevRel team](https://github.com/FuelLabs/devrel-requests/issues/new/choose) - [x] I have added tests that prove my fix is effective or that my feature works. - [ ] I have added (or requested a maintainer to add) the necessary `Breaking*` or `New Feature` labels where relevant. - [ ] I have done my best to ensure that my PR adheres to [the Fuel Labs Code Review Standards](https://github.com/FuelLabs/rfcs/blob/master/text/code-standards/external-contributors.md). - [x] I have requested a review from the relevant team or maintainers.	2025-10-16 14:41:51 +00:00