chenlin/Mujoco_WASM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor mju_cholUpdateSparse to use a dense accumulator, 1.5-2x function speedup.

Browse Source 
PiperOrigin-RevId: 846754379
Change-Id: Ie599658db2907a26095072d97d07b99a693050dc
This commit is contained in:
Yuval Tassa
2025-12-19 08:56:43 -08:00
committed by Copybara-Service
parent 2935af23f0
commit 849a650143
4 changed files with 236 additions and 41 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/engine/engine_util_solve.h
+2 -2
View File
@@ -62,9 +62,9 @@ void mju_cholSolveSparse(mjtNum* res, const mjtNum* mat, const mjtNum* vec, int
// sparse reverse-order Cholesky rank-one update: L'*L +/i x*x'; return rank
// x is sparse, change in sparsity pattern of mat is not allowed
MJAPI int mju_cholUpdateSparse(mjtNum* mat, mjtNum* x, int n, int flg_plus,
MJAPI int mju_cholUpdateSparse(mjtNum* mat, const mjtNum* x, int n, int flg_plus,
const int* rownnz, const int* rowadr, const int* colind,
int x_nnz, int* x_ind, mjData* d);
int x_nnz, const int* x_ind, mjData* d);
// band-dense Cholesky decomposition
// returns minimum value in the factorized diagonal, or 0 if rank-deficient