messages listlengths 2 2 | matrix_path stringlengths 36 69 | op_type stringclasses 1
value | K int64 0 0 | level stringclasses 3
values | matrix_name stringlengths 5 39 | num_rows int64 302 1M | num_cols int64 1 1.11M | nnz int64 358 16.2M |
|---|---|---|---|---|---|---|---|---|
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Arenas_email.bin | spgemm | 0 | level1_small | Arenas_email | 1,133 | 1,133 | 10,902 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Bai_rbsb480.bin | spgemm | 0 | level1_small | Bai_rbsb480 | 480 | 480 | 17,088 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Boeing_msc00726.bin | spgemm | 0 | level1_small | Boeing_msc00726 | 726 | 726 | 34,518 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Boeing_nasa1824.bin | spgemm | 0 | level1_small | Boeing_nasa1824 | 1,824 | 1,824 | 39,208 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Boeing_nasa2910.bin | spgemm | 0 | level1_small | Boeing_nasa2910 | 2,910 | 2,910 | 174,296 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/CPM_cz1268.bin | spgemm | 0 | level1_small | CPM_cz1268 | 1,268 | 1,268 | 12,786 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/DRIVCAV_cavity05.bin | spgemm | 0 | level1_small | DRIVCAV_cavity05 | 1,182 | 1,182 | 32,747 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/DRIVCAV_cavity10.bin | spgemm | 0 | level1_small | DRIVCAV_cavity10 | 2,597 | 2,597 | 76,367 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/DRIVCAV_cavity14.bin | spgemm | 0 | level1_small | DRIVCAV_cavity14 | 2,597 | 2,597 | 76,367 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/DRIVCAV_cavity15.bin | spgemm | 0 | level1_small | DRIVCAV_cavity15 | 2,597 | 2,597 | 76,367 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Dattorro_EternityII_A.bin | spgemm | 0 | level1_small | Dattorro_EternityII_A | 7,362 | 1 | 358 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Dattorro_EternityII_Etilde.bin | spgemm | 0 | level1_small | Dattorro_EternityII_Etilde | 10,054 | 1 | 512 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/FIDAP_ex20.bin | spgemm | 0 | level1_small | FIDAP_ex20 | 2,203 | 2,203 | 69,981 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/FIDAP_ex23.bin | spgemm | 0 | level1_small | FIDAP_ex23 | 1,409 | 1,409 | 43,703 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/FIDAP_ex24.bin | spgemm | 0 | level1_small | FIDAP_ex24 | 2,283 | 2,283 | 48,737 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/FIDAP_ex36.bin | spgemm | 0 | level1_small | FIDAP_ex36 | 3,079 | 3,079 | 53,843 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/FIDAP_ex4.bin | spgemm | 0 | level1_small | FIDAP_ex4 | 1,601 | 1,601 | 32,299 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/FIDAP_ex7.bin | spgemm | 0 | level1_small | FIDAP_ex7 | 1,633 | 1,633 | 54,543 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Goodwin_Goodwin_010.bin | spgemm | 0 | level1_small | Goodwin_Goodwin_010 | 1,182 | 1,182 | 32,282 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Grund_bayer08.bin | spgemm | 0 | level1_small | Grund_bayer08 | 3,008 | 3,008 | 27,576 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Grund_bayer09.bin | spgemm | 0 | level1_small | Grund_bayer09 | 3,083 | 3,083 | 21,216 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Grund_meg1.bin | spgemm | 0 | level1_small | Grund_meg1 | 2,904 | 2,904 | 58,142 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Gset_G29.bin | spgemm | 0 | level1_small | Gset_G29 | 2,000 | 2,000 | 39,980 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Gset_G39.bin | spgemm | 0 | level1_small | Gset_G39 | 2,000 | 2,000 | 23,556 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Gset_G5.bin | spgemm | 0 | level1_small | Gset_G5 | 800 | 800 | 38,352 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Gset_G54.bin | spgemm | 0 | level1_small | Gset_G54 | 1,000 | 1,000 | 11,832 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_bcsstk12.bin | spgemm | 0 | level1_small | HB_bcsstk12 | 1,473 | 1,473 | 34,241 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_bcsstk13.bin | spgemm | 0 | level1_small | HB_bcsstk13 | 2,003 | 2,003 | 83,883 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_bcsstm10.bin | spgemm | 0 | level1_small | HB_bcsstm10 | 1,086 | 1,086 | 22,092 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_bcsstm13.bin | spgemm | 0 | level1_small | HB_bcsstm13 | 2,003 | 2,003 | 21,943 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_cegb2802.bin | spgemm | 0 | level1_small | HB_cegb2802 | 2,802 | 2,802 | 277,362 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_lock2232.bin | spgemm | 0 | level1_small | HB_lock2232 | 2,232 | 2,232 | 80,352 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_lshp1561.bin | spgemm | 0 | level1_small | HB_lshp1561 | 1,561 | 1,561 | 10,681 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_mbeause.bin | spgemm | 0 | level1_small | HB_mbeause | 496 | 496 | 41,063 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_orani678.bin | spgemm | 0 | level1_small | HB_orani678 | 2,529 | 2,529 | 90,158 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_psmigr_1.bin | spgemm | 0 | level1_small | HB_psmigr_1 | 3,140 | 3,140 | 543,162 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_psmigr_2.bin | spgemm | 0 | level1_small | HB_psmigr_2 | 3,140 | 3,140 | 540,022 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/HB_psmigr_3.bin | spgemm | 0 | level1_small | HB_psmigr_3 | 3,140 | 3,140 | 543,162 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_CAG_CAG_mat1916.bin | spgemm | 0 | level1_small | JGD_CAG_CAG_mat1916 | 1,916 | 1,916 | 195,985 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_CAG_CAG_mat364.bin | spgemm | 0 | level1_small | JGD_CAG_CAG_mat364 | 364 | 364 | 13,585 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Forest_TF14.bin | spgemm | 0 | level1_small | JGD_Forest_TF14 | 2,644 | 3,160 | 29,862 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Groebner_HFE18_96_in.bin | spgemm | 0 | level1_small | JGD_Groebner_HFE18_96_in | 2,372 | 4,096 | 933,343 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Groebner_f855_mat9.bin | spgemm | 0 | level1_small | JGD_Groebner_f855_mat9 | 2,456 | 2,511 | 171,214 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Groebner_f855_mat9_I.bin | spgemm | 0 | level1_small | JGD_Groebner_f855_mat9_I | 2,456 | 2,511 | 171,214 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Groebner_robot24c1_mat5_J.bin | spgemm | 0 | level1_small | JGD_Groebner_robot24c1_mat5_J | 302 | 404 | 15,118 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Kocay_Trec13.bin | spgemm | 0 | level1_small | JGD_Kocay_Trec13 | 1,301 | 6,561 | 654,517 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_SL6_D_7.bin | spgemm | 0 | level1_small | JGD_SL6_D_7 | 1,270 | 971 | 12,714 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_SPG_EX3.bin | spgemm | 0 | level1_small | JGD_SPG_EX3 | 2,600 | 2,600 | 71,760 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Taha_abtaha1.bin | spgemm | 0 | level1_small | JGD_Taha_abtaha1 | 14,596 | 209 | 51,307 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Trefethen_Trefethen_2000.bin | spgemm | 0 | level1_small | JGD_Trefethen_Trefethen_2000 | 2,000 | 2,000 | 41,906 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/JGD_Trefethen_Trefethen_700.bin | spgemm | 0 | level1_small | JGD_Trefethen_Trefethen_700 | 700 | 700 | 12,654 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/LeGresley_LeGresley_2508.bin | spgemm | 0 | level1_small | LeGresley_LeGresley_2508 | 2,508 | 2,508 | 16,727 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/ML_Graph_YaleB_10NN.bin | spgemm | 0 | level1_small | ML_Graph_YaleB_10NN | 2,414 | 2,414 | 34,808 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/ML_Graph_dataset20mfeatpixel_10NN.bin | spgemm | 0 | level1_small | ML_Graph_dataset20mfeatpixel_10NN | 2,000 | 2,000 | 27,932 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/ML_Graph_yeast_30NN.bin | spgemm | 0 | level1_small | ML_Graph_yeast_30NN | 1,484 | 1,484 | 62,350 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Mallya_lhr02.bin | spgemm | 0 | level1_small | Mallya_lhr02 | 2,954 | 2,954 | 37,206 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/MathWorks_TS.bin | spgemm | 0 | level1_small | MathWorks_TS | 2,142 | 2,142 | 45,262 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/MathWorks_tomography.bin | spgemm | 0 | level1_small | MathWorks_tomography | 500 | 500 | 28,726 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Meszaros_aa5.bin | spgemm | 0 | level1_small | Meszaros_aa5 | 801 | 8,308 | 65,953 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Moqri_MISKnowledgeMap.bin | spgemm | 0 | level1_small | Moqri_MISKnowledgeMap | 2,427 | 2,427 | 57,022 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Morandini_rotor2.bin | spgemm | 0 | level1_small | Morandini_rotor2 | 791 | 791 | 10,685 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Mycielski_mycielskian11.bin | spgemm | 0 | level1_small | Mycielski_mycielskian11 | 1,535 | 1,535 | 134,710 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Mycielski_mycielskian12.bin | spgemm | 0 | level1_small | Mycielski_mycielskian12 | 3,071 | 3,071 | 407,200 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Mycielski_mycielskian9.bin | spgemm | 0 | level1_small | Mycielski_mycielskian9 | 383 | 383 | 14,542 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Negre_dendrimer.bin | spgemm | 0 | level1_small | Negre_dendrimer | 730 | 730 | 63,024 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Newman_polblogs.bin | spgemm | 0 | level1_small | Newman_polblogs | 1,490 | 1,490 | 19,025 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Norris_heart2.bin | spgemm | 0 | level1_small | Norris_heart2 | 2,339 | 2,339 | 682,797 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Norris_heart3.bin | spgemm | 0 | level1_small | Norris_heart3 | 2,339 | 2,339 | 682,797 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Oberwolfach_piston.bin | spgemm | 0 | level1_small | Oberwolfach_piston | 2,025 | 2,025 | 100,015 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Oberwolfach_spiral.bin | spgemm | 0 | level1_small | Oberwolfach_spiral | 1,434 | 1,434 | 18,228 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/QCD_conf5_0-4x4-26.bin | spgemm | 0 | level1_small | QCD_conf5_0-4x4-26 | 3,072 | 3,072 | 119,808 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/SNAP_CollegeMsg.bin | spgemm | 0 | level1_small | SNAP_CollegeMsg | 1,899 | 1,899 | 20,296 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/SNAP_email-Eu-core-temporal.bin | spgemm | 0 | level1_small | SNAP_email-Eu-core-temporal | 1,005 | 1,005 | 24,929 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/SNAP_email-Eu-core.bin | spgemm | 0 | level1_small | SNAP_email-Eu-core | 1,005 | 1,005 | 25,571 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Sandia_adder_dcop_03.bin | spgemm | 0 | level1_small | Sandia_adder_dcop_03 | 1,813 | 1,813 | 11,148 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Sandia_adder_dcop_18.bin | spgemm | 0 | level1_small | Sandia_adder_dcop_18 | 1,813 | 1,813 | 11,246 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Sandia_adder_dcop_31.bin | spgemm | 0 | level1_small | Sandia_adder_dcop_31 | 1,813 | 1,813 | 11,246 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Sandia_adder_dcop_37.bin | spgemm | 0 | level1_small | Sandia_adder_dcop_37 | 1,813 | 1,813 | 11,246 |
[
{
"role": "system",
"content": "You are an expert CUDA programmer specializing in sparse matrix kernels for NVIDIA GPUs (B200, Blackwell, sm_100). Write compact, correct, fast kernels. Respond with exactly one ```cuda code fence containing the complete contents of kernel.cu. The code must include any necess... | level_sparse/level1_small/Sandia_adder_dcop_64.bin | spgemm | 0 | level1_small | Sandia_adder_dcop_64 | 1,813 | 1,813 | 11,246 |
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Dr.Sparse SpGEMM training pool (562 matrices)
The complete RL / selector training pool of Dr.Sparse (branch v2): 562 SuiteSparse matrices in the
harness .bin layout (int32 rows, cols, nnz; int32 row_ptr; int32 col_ind; float32 values; float32 x),
laid out as level1_small/ (91), level2_medium/ (273), level3_large/ (198); the huge tier is deliberately left out of training and evaluation.
Every matrix has a cuSPARSE SpGEMM reference (C = AA, or AA^T when rectangular) on an H200;
the 27 that do not are listed in spgemm_excluded.txt and are not included. None of the OTF-81 test
matrices (level_sparse/otf_test_set.csv in the repo) is included. train.jsonl is the per-matrix
prompt file the RL launcher reads; manifest.csv lists rows, cols, nnz and level.
python level_sparse/extend_pool.py --hf-download KinGeorge/Dr.Sparse-RL-train-562 --out /big/disk/train562 --link-into level_sparse
- Downloads last month
- 43