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

Provide improved mju_sqrMatTDSparse implementation that doesn't require dense memory allocation for sparse matrices.

Browse Source 
PiperOrigin-RevId: 516812783
Change-Id: Ieb43337831d8d3b2c7f18a7facd8e0a0f2b0eff5
This commit is contained in:
Kyle Bayes
2023-03-15 06:55:28 -07:00
committed by Copybara-Service
parent fe18e58ad2
commit 056e849273
7 changed files with 830 additions and 131 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/engine/engine_util_sparse.h
+2 -2
View File
@@ -66,8 +66,8 @@ MJAPI void mju_transposeSparse(mjtNum* res, const mjtNum* mat, int nr, int nc,
const int* rownnz, const int* rowadr, const int* colind);
// construct row supernodes
void mju_superSparse(int nr, int* rowsuper,
const int* rownnz, const int* rowadr, const int* colind);
MJAPI void mju_superSparse(int nr, int* rowsuper,
const int* rownnz, const int* rowadr, const int* colind);
// compute sparse M'*diag*M (diag=NULL: compute M'*M), res has uncompressed layout
MJAPI void mju_sqrMatTDSparse(mjtNum* res, const mjtNum* mat, const mjtNum* matT,