Cross-binding benchmark — parity + timing (1 thread)

Single-thread numbers — the cleanest cross-language comparison. Most external libraries don’t honour OMP_NUM_THREADS at the algorithm level (sklearn, pls::plsr, plsregress, ikpls all run the PLS algo serially even when BLAS is multi-threaded), so a multi-thread comparison would compare pls4all’s OpenMP path against everyone else’s single-threaded loop. That’s a real, useful number — see the thread sweep page — but this main page sticks to 1 thread for honest apples-to-apples timing.

Generated: 2026-05-30 10:15:30 UTC CSV: tmpnxk259td.csv (664 cells)

Host

CPU	12th Gen Intel(R) Core(TM) i9-12900K
Cores	12 physical / 24 logical
Frequency	3187 MHz nominal
L3 cache	30720 KB
RAM	62.7 GB
OS / kernel	Linux 6.6.87.2-microsoft-standard-WSL2 (x86_64) · kernel 6.6.87.2-microsoft-standard-WSL2
Python	3.11.15

How to read a cell

Each cell shows <reported_ms> <gate>. C++ and external library cells use Gate 2, the reference parity against the method’s canonical oracle. Internal binding cells use Gate 1, the binding parity against the native C++ baseline:

• ✓ exact/pass — gate-specific tolerance passed

• ⚠ drift — finite mismatch, below 10× the method’s gate-2 tolerance or below the binding drift band

• ✗ divergent — outside the gate’s tolerance band

• ⏳ cell hit the 24 h wall-clock guard

• — see Why a cell is empty below

Bold = fastest cell on the row, counted only among cells whose relevant gate is ✓. For internal bindings that is Gate 1; for C++ and external libraries it is Gate 2. Drift / divergent / empty cells never carry the bold.

Timing uses the adaptive protocol: run #1 is the warmstart; if run #1 exceeds 5 min it is reported and the cell stops. Otherwise run #2 is the first scored sample and the warmstart is excluded. A 2-run cell reports run #2 alone; a 3-run cell reports the mean of runs #2-#3; 10/20/40-run cells report the median of runs after the warmstart. All backends in a single cell read the same orchestrator-generated CSV dataset. See methodology.md for the full details.

Why a cell is empty (—)

An empty cell means the backend did not produce a timing for that (algorithm, size) combination. Four distinct reasons, all reported as — in the matrix:

1. External library doesn’t implement the algorithm. Example: sklearn has no native CPPLS or sparse SIMPLS; plsregress (Octave) only does plain PLS; ikpls is plain PLS only; ropls only does OPLS; mixOmics covers PLS / sparse PLS / PLS-DA. Filling this column requires the external maintainer to add the algorithm — out of our control.

2. pls4all wrapper missing for that algorithm. Example: pls4all.sklearn (tier 2) doesn’t yet ship a class for continuum_regression or kernel_pls; pls4all.R.formula doesn’t have a formula wrapper for every algorithm yet. This is the path we still need to cover on our side — visible TODO.

3. Bench script missing the dispatch case. A few cells can fail because a binding bench script doesn’t yet route a specific algorithm to its underlying call. Pure tooling gap, no library impact.

4. Run too slow / OOM / crashed. The cell hit the 24 h guard (⏳) or raised a runtime error.

Each — represents one of these four reasons. The CSV (results/full_matrix.csv) carries a reason column that tells you exactly which one for any given cell.

Column names: anything starting with pls4all. is one of this project’s bindings; the others use their real package name (sklearn, ikpls, pls, ropls, mixOmics, plsregress). Every algorithm table keeps every column — — cells are intentional and show where coverage is still missing, either on our side (an algorithm not yet wrapped in a tier-2 class) or on the external side (e.g. sklearn doesn’t implement CPPLS / OPLS, plsregress only does plain PLS, ikpls only does plain PLS). Full per-column description: see Backend definitions at the bottom of this page.

aom_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	4.20 ms ✓	4.30 ms ✓	4.16 ms ✓	4.15 ms ✓	7.84 ms ✓	9.43 ms ✓	9.87 ms ✓	17.6 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	41.9 ms ✓	—	—	—	—	—	—

aom_preprocess — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.85 ms ✓	2.52 ms ✓	1.76 ms ✓	1.75 ms ✓	6.07 ms ✓	7.96 ms ✓	7.33 ms ✓	6.56 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	2.13 ms ✓	—	—	—	—	—	—

approximate_press — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	28.4 ms ✓	28.0 ms ✓	28.1 ms ✓	28.1 ms ✓	24.7 ms ✓	25.5 ms ✓	25.7 ms ✓	25.3 ms ✓	—	69.5 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

bagging_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	9.82 ms ✓	10.2 ms ✓	9.30 ms ✓	1.69 ms ✗	4.62 ms ✗	6.41 ms ✗	4.93 ms ✗	5.11 ms ✗	11.1 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

bipls_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	3.83 ms ✓	4.36 ms ✓	3.60 ms ✓	3.69 ms ✓	8.26 ms ✓	11.7 ms ✓	8.83 ms ✓	9.90 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	365.6 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

boosting_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.23 ms ✓	1.23 ms ✓	1.52 ms ✓	2.07 ms ✗	5.05 ms ✗	4.95 ms ✗	6.27 ms ✗	6.92 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	18.5 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—

bve_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	510.0 ms ✓	443.3 ms ✓	563.8 ms ✓	42.4 ms ✗	58.7 ms ✗	66.5 ms ✗	141.0 ms ✗	152.1 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	46.9 ms ✓	—	—	—	—	—

cars_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	780.7 ms ✓	752.8 ms ✓	772.6 ms ✓	—	13.3 ms ✗	10.2 ms ✗	13.3 ms ✗	16.4 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	188.9 ms ✓	—	—	—	—

continuum_regression — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	2.73 ms ✓	2.76 ms ✓	2.63 ms ✓	2.77 ms ✓	5.55 ms ✓	6.02 ms ✓	6.62 ms ✓	6.62 ms ✓	—	—	—	—	—	—	—	—	—	2.27 ms ✓	31.6 ms —	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

cppls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.88 ms ✓	1.79 ms ✓	1.87 ms ✓	1.97 ms ✓	4.57 ms ✓	5.65 ms ✓	10.6 ms ✗	6.07 ms ✗	—	9.91 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

di_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	2.58 ms ✓	2.57 ms ✓	2.74 ms ✓	2.82 ms ✓	9.69 ms ✓	12.4 ms ✓	12.4 ms ✓	12.8 ms ✓	—	—	—	—	—	—	—	3.78 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

ds — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.64 ms ✓	1.73 ms ✓	2.09 ms ✓	2.27 ms ✓	10.2 ms ✓	14.0 ms ✓	12.6 ms ✓	11.4 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	31.0 ms ✓	—	—	—	—	—	—	—	—	—	—

ecr — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	2.16 ms ✓	2.15 ms ✓	2.24 ms ✓	2.37 ms ✓	5.88 ms ✓	6.96 ms ✓	8.61 ms ✓	7.41 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	62.4 ms ✓	—

emcuve_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.3 s ✓	1.3 s ✓	1.3 s ✓	1.98 ms ✓	4.51 ms ✓	5.44 ms ✓	5.57 ms ✓	5.25 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	922.4 ms ✓	—	—	—	—	—

fused_sparse_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.93 ms —	1.86 ms ✓	1.95 ms ✓	2.13 ms ✓	6.13 ms ✓	7.70 ms ✓	7.80 ms ✓	7.55 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

ga_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	607.6 ms ✓	625.8 ms ✓	756.5 ms ✓	3.94 ms ✗	7.38 ms ✗	8.59 ms ✗	9.76 ms ✗	16.0 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	218.7 ms ✓	—	—	—	—	—

gpr_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
120×25	5.91 ms ✓	6.75 ms ✓	6.17 ms ✓	2.75 ms ✓	20.9 ms ✓	12.6 ms ✓	14.2 ms ✓	8.10 ms ✓	2.25 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

group_sparse_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	8.36 ms ✓	4.69 ms ✓	7.86 ms ✓	3.79 ms ✗	9.70 ms ✗	11.0 ms ✗	11.0 ms ✗	10.2 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	2.98 ms ✓	—	—	—	—	—	—	—	33.7 ms —	—	—	—	—	—	—	—

interval_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.5 s ✓	2.0 s ✓	1.5 s ✓	3.75 ms ✗	10.9 ms ✗	11.7 ms ✗	10.3 ms ✗	13.9 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	632.3 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

ipw_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	360.4 ms ✓	358.1 ms ✓	366.9 ms ✓	1.81 ms ✗	4.86 ms ✗	5.51 ms ✗	5.86 ms ✗	5.15 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	22.0 ms ✓	—	—	—	—	—

irf_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
120×30	—	—	—	2.02 ms ✓	2.92 ms ✓	3.44 ms ✓	3.91 ms ✓	3.37 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

iriv_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
80×25	268.0 ms ✓	269.9 ms ✓	270.2 ms ✓	25.3 ms ✗	29.4 ms ✗	30.0 ms ✗	30.5 ms ✗	30.8 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	274.8 ms ✓

kernel_pls_rbf — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	3.07 ms ✓	3.02 ms ✓	3.23 ms ✓	3.23 ms ✓	6.34 ms ✓	7.05 ms ✓	7.21 ms ✓	6.83 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	28.6 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

lw_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	25.6 ms ✓	13.3 ms ✓	37.2 ms ✓	9.19 ms ✗	20.3 ms ✗	23.2 ms ✗	23.2 ms ✗	24.6 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	23.9 ms ✓	—	—	—	—	—	—

mb_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×60	2.37 ms ✓	2.73 ms ✓	2.85 ms ✓	2.59 ms ✓	7.82 ms ✓	10.9 ms ✓	10.2 ms ✓	11.5 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	2.99 ms ✓	—	—	—	—	—	—

mir_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.19 ms —	1.21 ms ✓	1.27 ms ✓	1.47 ms ✓	3.98 ms ✓	4.72 ms ✓	5.04 ms ✓	5.11 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

missing_aware_nipals — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.27 ms ✓	1.24 ms ✓	1.25 ms ✓	1.39 ms ✓	4.25 ms ✓	4.88 ms ✓	4.85 ms ✓	5.41 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	13.6 ms ✓	—	—	—	—	—	—	—	—	—

n_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×48	19.5 ms ✓	19.3 ms ✓	19.0 ms ✓	2.03 ms ✗	4.51 ms ✗	5.67 ms ✗	5.68 ms ✗	5.37 ms ✗	—	—	—	—	—	—	—	—	—	—	—	19.3 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

o2pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.21 ms ✓	1.20 ms ✓	1.19 ms ✓	1.37 ms ✓	2.69 ms ✓	3.44 ms ✓	3.58 ms ✓	3.42 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	7.21 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

on_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	6.20 ms ✓	7.71 ms ✓	3.02 ms ✓	2.69 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	12.2 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

one_se_rule — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	2.10 ms ✓	2.07 ms ✓	2.08 ms ✓	1.13 ms ✗	3.25 ms ✗	3.87 ms ✗	4.01 ms ✗	3.96 ms ✗	—	11.6 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

opls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.87 ms ✓	1.93 ms ✓	1.88 ms ✓	2.12 ms ✓	4.35 ms ✓	5.13 ms ✓	5.40 ms ✓	5.23 ms ✓	—	—	—	—	15.4 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pcr — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	2.92 ms ✓	2.84 ms ✓	2.95 ms ✓	2.68 ms ✓	5.15 ms ✓	6.41 ms ✓	11.8 ms ✓	6.21 ms ✓	2.71 ms ✓	7.99 ms —	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pds — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.37 ms ✓	1.55 ms ✓	1.44 ms ✓	1.86 ms ✓	10.6 ms ✓	10.1 ms ✓	14.2 ms ✓	13.7 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	43.1 ms ✓	—	—	—	—	—	—	—	—	—	—

pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.70 ms ✓	1.73 ms ✓	1.69 ms ✓	1.97 ms ✓	4.79 ms ✓	5.26 ms ✓	9.99 ms ✓	5.92 ms ✓	2.16 ms ✓	8.01 ms —	1.92 ms —	9.72 ms —	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_cox — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.79 ms ✓	1.84 ms ✓	1.75 ms ✓	2.13 ms ✗	4.52 ms ✗	5.52 ms ✗	5.43 ms ✗	5.51 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	2.03 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_diagnostic_dmodx — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.25 ms ✓	1.30 ms ✓	1.25 ms ✓	1.41 ms ✓	4.16 ms ✓	4.79 ms ✓	4.47 ms ✓	5.71 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	19.8 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_diagnostic_q — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.22 ms ✓	1.19 ms ✓	1.19 ms ✓	1.30 ms ✓	3.10 ms ✓	3.57 ms ✓	3.47 ms ✓	3.96 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	13.8 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_diagnostic_t2 — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.19 ms ✓	1.22 ms ✓	1.22 ms ✓	1.31 ms ✓	2.67 ms ✓	3.38 ms ✓	4.11 ms ✓	3.70 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	14.2 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_glm — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	3.57 ms ✓	6.25 ms ✓	7.99 ms ✓	2.74 ms ✗	7.97 ms ✗	22.7 ms ✗	18.4 ms ✗	25.8 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	240.6 ms ✓	—	—	—	—	—	—	—	—	—	—	—

pls_lda — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	3.14 ms ✓	7.18 ms ✓	7.48 ms ✓	—	—	—	—	—	7.32 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_logistic — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	3.14 ms ✓	4.04 ms ✓	3.34 ms ✓	—	—	—	—	—	3.44 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_monitoring — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.23 ms ✓	1.23 ms ✓	1.18 ms ✓	1.32 ms ✓	2.90 ms ✓	3.54 ms ✓	3.74 ms ✓	3.81 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	14.2 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pls_qda — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	2.65 ms ✓	2.09 ms ✓	2.12 ms ✓	2.11 ms ✗	4.06 ms ✗	4.67 ms ✗	4.99 ms ✗	5.46 ms ✗	2.36 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

pop_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	6.24 ms ✓	5.68 ms ✓	7.23 ms ✓	6.51 ms ✓	12.0 ms ✓	12.5 ms ✓	13.2 ms ✓	24.7 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	47.8 ms ✓	—	—	—	—	—	—

pso_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
80×25	219.8 ms ✓	384.2 ms ✓	219.3 ms ✓	4.05 ms ✗	6.11 ms ✗	6.29 ms ✗	7.12 ms ✗	7.33 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	163.8 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—

random_frog_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	—	—	—	2.19 ms ✓	4.98 ms ✓	5.67 ms ✓	5.94 ms ✓	5.96 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

random_subspace_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	13.5 ms ✓	25.2 ms ✓	13.2 ms ✓	1.54 ms ✗	5.11 ms ✗	5.82 ms ✗	5.56 ms ✗	6.22 ms ✗	13.0 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

randomization_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	181.1 ms ✓	182.3 ms ✓	182.5 ms ✓	2.67 ms ✓	6.34 ms ✓	5.94 ms ✓	6.80 ms ✓	6.29 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	54.5 ms ✓	—	—

recursive_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	2.92 ms ✓	2.82 ms ✓	2.79 ms ✓	3.10 ms ✓	5.74 ms ✓	6.55 ms ✓	7.36 ms ✓	6.99 ms ✓	48.1 ms ✓	141.6 ms —	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

rep_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	605.7 ms ✓	592.6 ms ✓	588.4 ms ✓	3.35 ms ✗	5.16 ms ✗	6.34 ms ✗	6.07 ms ✗	6.53 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	241.7 ms ✓	—	—	—	—	—

ridge_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.75 ms ✓	2.14 ms ✓	1.83 ms ✓	1.75 ms ✓	4.62 ms ✓	6.20 ms ✓	5.59 ms ✓	5.56 ms ✓	2.19 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

robust_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	2.07 ms ✓	1.95 ms ✓	1.95 ms ✓	2.15 ms ✓	4.38 ms ✓	5.55 ms ✓	6.62 ms ✓	5.77 ms ✓	—	—	—	—	—	—	—	—	17.9 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

rosa — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	7.07 ms ✓	2.04 ms ✓	2.41 ms ✓	2.56 ms ✓	5.46 ms ✓	6.38 ms ✓	6.68 ms ✓	6.80 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	2.9 s —	—	1.43 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

scars_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	7.65 ms ✓	6.52 ms ✓	10.2 ms ✓	4.14 ms ✗	4.85 ms ✗	5.55 ms ✗	5.72 ms ✗	5.92 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	3.61 ms ✓	—	—	—

shaving_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.0 s ✓	411.6 ms ✓	440.9 ms ✓	2.32 ms ✗	5.59 ms ✗	7.94 ms ✗	6.29 ms ✗	6.31 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	38.2 ms ✓	—	—	—	—	—

sipls_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	2.52 ms —	3.21 ms ✓	3.18 ms ✓	2.82 ms ✓	7.86 ms ✓	8.15 ms ✓	8.12 ms ✓	8.80 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

so_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×30	1.62 ms ✓	1.62 ms ✓	1.67 ms ✓	1.77 ms ✓	3.26 ms ✓	3.99 ms ✓	4.11 ms ✓	4.00 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	981.7 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

spa_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	499.6 ms ✓	536.6 ms ✓	515.0 ms ✓	1.97 ms ✗	4.72 ms ✗	6.23 ms ✗	5.75 ms ✗	5.86 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	104.0 ms ✓	—	—	—	—	—

sparse_pls_da — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	7.43 ms ✓	11.1 ms ✓	4.50 ms ✓	2.45 ms ✗	7.20 ms ✓	8.68 ms ✓	8.53 ms ✓	8.43 ms ✓	—	—	—	—	—	—	82.9 ms —	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	3.58 ms ✓	—	—	—	—	—	—	—	—

sparse_simpls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.89 ms ✓	1.82 ms ✓	1.93 ms ✓	1.88 ms ✓	4.60 ms ✓	5.64 ms ✓	5.36 ms ✓	6.30 ms ✓	—	—	—	—	—	3.55 ms ✓	13.4 ms —	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

st_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	403.7 ms ✓	409.2 ms ✓	418.3 ms ✓	1.82 ms ✗	6.05 ms ✗	5.27 ms ✗	5.46 ms ✗	5.45 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	17.8 ms ✓	—	—	—	—	—

stability_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	690.4 ms ✓	820.9 ms ✓	1.7 s ✓	3.70 ms ✗	15.9 ms ✗	6.82 ms ✗	8.38 ms ✗	8.71 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	492.8 ms ✓	—	—	—	—	—

t2_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.67 ms —	1.67 ms ✓	1.69 ms ✓	1.83 ms ✓	4.21 ms ✓	6.34 ms ✓	5.80 ms ✓	6.75 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	37.0 ms ✓	—	—	—	—	—

uve_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.5 s ✓	1.5 s ✓	1.7 s ✓	3.28 ms ✗	10.3 ms ✗	7.89 ms ✗	6.77 ms ✗	6.20 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	222.0 ms ✓	—	—	—	—	—

variable_select_coef — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.56 ms ✓	1.56 ms ✓	1.61 ms ✓	2.87 ms ✓	15.0 ms ✓	9.19 ms ✓	17.2 ms ✓	18.1 ms ✓	—	50.9 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

variable_select_sr — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	542.8 ms ✓	1.5 s ✓	1.2 s ✓	8.28 ms ✗	24.1 ms ✗	11.2 ms ✗	31.6 ms ✗	22.3 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	18.9 ms ✓	—	—	—	—	—

variable_select_vip — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	2.55 ms ✓	2.70 ms ✓	2.47 ms ✓	1.97 ms ✓	6.19 ms ✓	6.51 ms ✓	5.86 ms ✓	6.61 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	16.2 ms ✓	—	—	—	—	—

vip_spa_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
80×40	—	—	—	1.34 ms ✓	2.19 ms ✓	2.80 ms ✓	2.83 ms ✓	2.87 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

vissa_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
80×25	—	—	—	10.8 ms ✓	11.2 ms ✓	11.1 ms ✓	21.4 ms ✓	49.1 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

weighted_pls — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×50	1.93 ms ✓	1.82 ms ✓	1.77 ms ✓	1.92 ms ✓	4.57 ms ✓	5.88 ms ✓	5.23 ms ✓	5.95 ms ✓	2.12 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

wvc_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	1.49 ms ✓	1.47 ms ✓	1.48 ms ✓	1.63 ms ✓	4.23 ms ✓	4.97 ms ✓	5.72 ms ✓	5.30 ms ✓	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	14.4 ms ✓	—	—	—	—	—

wvc_threshold_select — 1 thread

samples × features	pls4all.cpp.blas+omp	pls4all.registry	pls4all.python	pls4all.sklearn	pls4all.R	pls4all.R.formula	pls4all.R.pls_compat	pls4all.R.mdatools_compat	ref.python_scikit_learn	ref.r_pls	ref.python_ikpls	ref.r_mixomics	ref.r_ropls	ref.python_chun_keles_spls	ref.r_spls	ref.python_diplslib	ref.r_chemometrics	ref.python_stone_brooks_1990_py	ref.r_jico	ref.python_tensorly	ref.r_kernlab_pls	ref.r_omicspls	ref.r_mdatools	ref.r_multiblock	ref.python_onpls	ref.python_numpy	ref.python_auswahl	ref.python_pyswarms	ref.r_mboost	ref.r_plsrglm	ref.r_base	ref.r_softimpute	ref.python_chun_keles_splsda	ref.r_sgpls	ref.python_nirs4all	ref.r_plsvarsel	ref.r_enpls	ref.python_scars_numpy_port	ref.r_pls_stats	libPLS	ref.python_iriv_numpy_port
200×40	336.1 ms ✓	344.6 ms ✓	344.7 ms ✓	1.67 ms ✗	4.07 ms ✗	4.66 ms ✗	5.51 ms ✗	4.59 ms ✗	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	13.0 ms ✓	—	—	—	—	—

Backend definitions

Each column in the per-algorithm tables above is one of the entries below. Columns are named owner.language[.variant]: pls4all.* are this project’s own bindings; everything else is an external library shipped by its own maintainers.

Column	Language	Tier	What it actually runs
pls4all.cpp.blas+omp	C++	pls4all + BLAS + OpenMP	libn4m built with both PLS4ALL_WITH_BLAS=ON and PLS4ALL_WITH_OPENMP=ON — the recommended production config.
pls4all.registry	Python	pls4all canonical	benchmarks.parity_timing.registry.MethodSpec.pls4all_fn — the canonical per-method pls4all entry point
pls4all.python	Python	pls4all raw	pls4all._methods.<algo>_fit(ctx, cfg, X, y, …) — direct FFI binding
pls4all.sklearn	Python	pls4all idiomatic	pls4all.sklearn.<Class> — sklearn-style BaseEstimator with .fit() / .predict()
pls4all.R	R	pls4all raw	pls4all_method(algo, X, y, ...) — unified dispatcher (33 fits + 24 selectors + 4 diagnostics)
pls4all.R.formula	R	pls4all idiomatic	pls(y ~ ., data), cppls(...), sparse_pls(...), … — base R formula+S3 wrappers
pls4all.R.pls_compat	R	pls4all pls-compatible	plsr() / pcr() for PLS/PCR/CPPLS, formula-built dispatcher path for the rest of the method matrix
pls4all.R.mdatools_compat	R	pls4all mdatools-compatible	pls(x, y, ...) for PLS/PCR/CPPLS, matrix dispatcher path for the rest of the method matrix
ref.python_scikit_learn	external	reference	registry-declared external reference library
ref.r_pls	external	reference	registry-declared external reference library
ref.python_ikpls	external	reference	registry-declared external reference library
ref.r_mixomics	external	reference	registry-declared external reference library
ref.r_ropls	external	reference	registry-declared external reference library
ref.python_chun_keles_spls	external	reference	registry-declared external reference library
ref.r_spls	external	reference	registry-declared external reference library
ref.python_diplslib	external	reference	registry-declared external reference library
ref.r_chemometrics	external	reference	registry-declared external reference library
ref.python_stone_brooks_1990_py	external	reference	registry-declared external reference library
ref.r_jico	external	reference	registry-declared external reference library
ref.python_tensorly	external	reference	registry-declared external reference library
ref.r_kernlab_pls	external	reference	registry-declared external reference library
ref.r_omicspls	external	reference	registry-declared external reference library
ref.r_mdatools	external	reference	registry-declared external reference library
ref.r_multiblock	external	reference	registry-declared external reference library
ref.python_onpls	external	reference	registry-declared external reference library
ref.python_numpy	external	reference	registry-declared external reference library
ref.python_auswahl	external	reference	registry-declared external reference library
ref.python_pyswarms	external	reference	registry-declared external reference library
ref.r_mboost	external	reference	registry-declared external reference library
ref.r_plsrglm	external	reference	registry-declared external reference library
ref.r_base	external	reference	registry-declared external reference library
ref.r_softimpute	external	reference	registry-declared external reference library
ref.python_chun_keles_splsda	external	reference	registry-declared external reference library
ref.r_sgpls	external	reference	registry-declared external reference library
ref.python_nirs4all	external	reference	registry-declared external reference library
ref.r_plsvarsel	external	reference	registry-declared external reference library
ref.r_enpls	external	reference	registry-declared external reference library
ref.python_scars_numpy_port	external	reference	registry-declared external reference library
ref.r_pls_stats	external	reference	registry-declared external reference library
libPLS	external	reference	registry-declared external reference library
ref.python_iriv_numpy_port	external	reference	registry-declared external reference library

Versions per backend

Column	Versions
pls4all.cpp.blas+omp	language=C++ (via ctypes bridge) (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; libopenblas 0.3.33 blas=1; libomp ? openmp=1; libn4m=0.98.0+abi.1.9.0; numpy=2.4.6; registry_method=pls; timing_schema=adaptive-v1
pls4all.registry	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libopenblas 0.3.33 blas=1; libomp ? openmp=1; pls4all=0.98.0; numpy=2.4.6; registry_method=pls; timing_schema=adaptive-v1
pls4all.python	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libopenblas 0.3.33 blas=1; libomp ? openmp=1; pls4all=0.98.0; numpy=2.4.6; registry_method=pls; timing_schema=adaptive-v1
pls4all.sklearn	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libopenblas 0.3.33 blas=1; libomp ? openmp=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; pls4all=0.98.0; numpy=2.4.6; sklearn_class=PLSRegression; timing_schema=adaptive-v1
pls4all.R	language=R 4.3.3; pls4all=0.98.0; registry_method=pls; blas=linked-BLAS; timing_schema=adaptive-v1
pls4all.R.formula	language=R 4.3.3; pls4all=0.98.0; registry_method=pls; formula_facade=True; blas=linked-BLAS; timing_schema=adaptive-v1
pls4all.R.pls_compat	language=R 4.3.3; pls4all=0.98.0; registry_method=pls; facade=pls_compat; formula_facade=True; blas=linked-BLAS; timing_schema=adaptive-v1
pls4all.R.mdatools_compat	language=R 4.3.3; pls4all=0.98.0; registry_method=pls; facade=mdatools_compat; matrix_facade=True; blas=linked-BLAS; timing_schema=adaptive-v1
ref.python_scikit_learn	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_scikit_learn; reference_library=scikit-learn; reference_version=1.8.0; timing_schema=adaptive-v1
ref.r_pls	language=R 4.3.3; pls=2.8.5; timing_schema=adaptive-v1
ref.python_ikpls	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_ikpls; reference_library=ikpls; reference_version=4.0.1.post1; timing_schema=adaptive-v1
ref.r_mixomics	language=R 4.3.3; mixOmics=6.26.0; timing_schema=adaptive-v1
ref.r_ropls	language=R 4.3.3; ropls=1.34.0; timing_schema=adaptive-v1
ref.python_chun_keles_spls	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_chun_keles_spls; reference_library=chun_keles_spls; reference_version=1.0; timing_schema=adaptive-v1
ref.r_spls	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_spls; reference_library=spls; reference_version=2.3.2; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.python_diplslib	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_diplslib; reference_library=diPLSlib; reference_version=2.5.0; timing_schema=adaptive-v1
ref.r_chemometrics	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_chemometrics; reference_library=chemometrics; reference_version=0.7.x; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.python_stone_brooks_1990_py	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=python_stone_brooks_1990_py; reference_library=stone-brooks-1990-py; reference_version=1.0; timing_schema=adaptive-v1
ref.r_jico	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_jico; reference_library=JICO; reference_version=0.1; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.python_tensorly	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_tensorly; reference_library=tensorly; reference_version=0.9.0; timing_schema=adaptive-v1
ref.r_kernlab_pls	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_kernlab_pls; reference_library=kernlab+pls; reference_version=0.9.33+2.8.5; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_omicspls	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_omicspls; reference_library=OmicsPLS; reference_version=2.1.0; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_mdatools	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_mdatools; reference_library=mdatools; reference_version=0.15.0; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_multiblock	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_multiblock; reference_library=multiblock; reference_version=0.8.10; timing_schema=adaptive-v1
ref.python_onpls	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=python_onpls; reference_library=OnPLS; reference_version=github tomlof/OnPLS; timing_schema=adaptive-v1
ref.python_numpy	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=python_numpy; reference_library=numpy; reference_version=2.4.6; timing_schema=adaptive-v1
ref.python_auswahl	—
ref.python_pyswarms	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_pyswarms; reference_library=pyswarms; reference_version=1.3.0; timing_schema=adaptive-v1
ref.r_mboost	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_mboost; reference_library=mboost; reference_version=2.9-11; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_plsrglm	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_plsrglm; reference_library=plsRglm; reference_version=1.5.1; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_base	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_base; reference_library=base; reference_version=R 4.3.3; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_softimpute	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_softimpute; reference_library=softImpute; reference_version=1.4-1; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.python_chun_keles_splsda	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_chun_keles_splsda; reference_library=chun_keles_splsda; reference_version=1.0; timing_schema=adaptive-v1
ref.r_sgpls	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_sgpls; reference_library=sgPLS; reference_version=1.8.1; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.python_nirs4all	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_nirs4all; reference_library=nirs4all; reference_version=in-tree; timing_schema=adaptive-v1
ref.r_plsvarsel	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_plsvarsel; reference_library=plsVarSel; reference_version=0.10.0; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.r_enpls	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_enpls; reference_library=enpls; reference_version=6.1; timing_driver=rscript-batch; timing_schema=adaptive-v1
ref.python_scars_numpy_port	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_scars_numpy_port; reference_library=scars_numpy_port; reference_version=1.0.0; timing_schema=adaptive-v1
ref.r_pls_stats	language=R (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=r_pls_stats; reference_library=pls+stats; reference_version=R 4.3.3; timing_driver=rscript-batch; timing_schema=adaptive-v1
libPLS	language=matlab (host Linux x86_64); blas=libscipy_openblas 0.3.31.188.0 blas=1; numpy=2.4.6; reference_id=matlab_libpls; reference_library=libPLS; reference_version=1.95; timing_schema=adaptive-v1
ref.python_iriv_numpy_port	language=Python 3.11.15; blas=libscipy_openblas 0.3.31.188.0 blas=1; libscipy_openblas 0.3.30 blas=1; libgomp ? openmp=1; numpy=2.4.6; reference_id=python_iriv_numpy_port; reference_library=iriv_numpy_port; reference_version=1.0.0; timing_schema=adaptive-v1

Methodology

• Gate 1 reference: cpp cell at 1 thread (libn4m via ctypes), or python_tier1 when cpp is unavailable for an algorithm

• Gate 2 reference: the registry-declared canonical external reference for the method

• Gate 1 tolerance: 1e-6 max-abs-diff; Gate 2 tolerance: MethodSpec.rmse_rel_tol (also emitted as reference_parity_tolerance by newer CSVs)

• All backends read the same orchestrator-generated CSV dataset (benchmarks/cross_binding/data/data_<n>x<p>_seed<seed>.csv)

• Adaptive timing: warmstart run #1; if there is any subsequent run, run #1 is excluded from the score. 2 total runs report run #2 alone; 3 total runs report the mean of runs #2-#3; 10/20/40 total runs report the median after the warmstart.

• Per-cell timeout guard: 24 h

• Thread control via OMP_NUM_THREADS = OPENBLAS_NUM_THREADS = MKL_NUM_THREADS = BLIS_NUM_THREADS set in the subprocess env, plus Context.num_threads for Python pls4all and maxNumCompThreads() for Octave

• pls4all libn4m build: build/blas-omp/cpp/src/libn4m.so (BLAS + OpenMP enabled)