Kimi-K2.6 · popcorn · l12_default

L1 P1 1_Conv1d_Depthwise.py

correct 1.27x 1.0μs

L1 P2 2_GraphEdgeSoftmaxCSR.py

error —

L1 P3 3_CSRSpMMMessagePassing.py

error —

L1 P4 4_GJKRobotEnvironmentIntersectio

correct 679.75x 0.0μs

L1 P5 5_Conv1d_Causal.py

error —

L1 P6 6_EdgeSoftmaxMultiHeadCSR.py

correct 1053.33x 0.0μs

L1 P7 7_Conv1d_FFT.py

error —

L1 P8 8_SegmentTopKCSR.py

error —

L1 P9 9_AssociativeScan.py

error —

L1 P10 10_SampledDenseDenseMatmulEdges.

error —

L1 P11 11_DegreeNormalizedAggregation.p

correct 2978.45x 0.0μs

L1 P12 12_DeepSeekMoEDispatchPermute.py

correct 1121.02x 0.1μs

L1 P13 13_FarthestPointSampling3D.py

error —

L1 P14 14_GaussianProcessRBFKernel.py

error —

L1 P15 15_COOScatterAddNodeFeatures.py

error —

L1 P16 16_DeepSeekMoECombineScatter.py

error —

L1 P17 17_CSRMaxAggregation.py

correct 1802.03x 0.0μs

L1 P18 18_CSRMultiHeadSpMM.py

error —

L1 P19 19_PairwiseDistanceMatrix.py

correct 7.51x 0.0μs

L1 P20 20_RotaryPositionEmbeddingBio.py

correct 6.90x 0.0μs

L1 P21 21_RadialBasisFunctionExpansion.

error —

L1 P22 22_SE3InvariantLinear.py

correct 4.88x 0.0μs

L1 P23 23_virtual_all_reduce_mean.py

correct 1.18x 0.0μs

L1 P24 24_VoxelGridPooling.py

correct 276.88x 0.0μs

L1 P25 25_virtual_reduce_scatter_masked

correct 2.12x 0.0μs

L1 P26 26_ImportanceSampling.py

error —

L1 P27 27_permutation_all_to_all.py

correct 4.90x 0.0μs

L1 P28 28_DirichletMultinomialLogLikeli

error —

L1 P29 29_logsumexp_across_virtual_rank

correct 5.23x 0.0μs

L1 P30 30_broadcast_masked_from_row0.py

correct 1.18x 0.0μs

L1 P31 31_bitmask_or_popcount.py

correct 82.19x 0.0μs

L1 P32 32_argmax_tiebreak_smallest_rank

error —

L2 P1 1_DeepSeekMLALoRAExpansion.py

error —

L2 P3 3_FusedCrossAttentionRoPEDropout

error —

L2 P4 4_GFVLA_FusedSceneGraphUpdate.py

correct 2.03x 0.2μs

L2 P5 5_LayerNorm_GELU_DilatedConv1d_R

error —

L2 P6 6_TriangularAttention.py

correct 1.75x 0.2μs

L2 P7 7_Conv1d_Depthwise_Separable.py

correct 1.48x 0.4μs

L2 P8 8_DeepSeekMoEGroundedTop2Routing

correct 2.99x 0.0μs

L2 P9 9_MetropolisHastingsStep.py

correct 8.32x 0.0μs

L2 P10 10_GFVLA_FusedPairwiseNodeAggreg

correct 0.25x 0.5μs

L2 P12 12_BatchNorm_GELU_DilatedConv1d_

error —

L2 P13 13_QuantVLA_SelectiveLayoutFusio

correct 1.51x 0.1μs

L2 P14 14_TriangularMultiplicativeUpdat

error —

L2 P15 15_GQAKVHeadExpansionAttention.p

correct 6289.31x 0.0μs

L2 P16 16_HamiltonianMonteCarloStep.py

correct 42.97x 0.0μs

L2 P17 17_FP8ScaledAttention.py

correct 940.53x 0.0μs

L2 P19 19_DepthwiseSeparableConv1D_GELU

error —

L2 P20 20_GatedDeltaNetLinearAttention.

error —

L2 P21 21_TriangularMultiplicativeUpdat

correct 2.03x 0.1μs

L2 P22 22_FusedBallQuery3D.py

error —

L2 P23 23_GibbsSamplingStep.py

correct 9.05x 0.4μs

L2 P24 24_KimiDeltaAttentionChannelwise

error —

L2 P25 25_OuterProductMean.py

error —

L2 P26 26_RoPEKVCacheUpdate.py

correct 225.71x 0.0μs

L2 P28 28_FusedGroupNormSiLU.py

correct 1.34x 0.0μs

L2 P29 29_MSARowAttention.py

correct 2.37x 0.2μs

L2 P30 30_SelectiveScan_Mamba.py

correct 56.22x 2.3μs

L2 P31 31_FusedMLAAttention.py

correct 3441.40x 0.1μs

L2 P32 32_MSAColumnAttention.py

correct 3.83x 0.1μs

L2 P33 33_VectorizedPoseTransformFused.

correct 45.79x 0.0μs

L2 P35 35_FusedConv1dGroupNorm.py

error —

L2 P36 36_InvariantPointAttention.py

correct 3.95x 0.2μs

L2 P37 37_CSRFusedAttentionValue.py

error —

L2 P39 39_Fused1DTemporalConvolution.py

correct 0.07x 1.5μs

L2 P40 40_SmithWatermanDPScore.py

error —

L2 P41 41_VariationalELBO.py

error —

L2 P42 42_BatchedForwardKinematics.py

correct 3.68x 0.0μs

L2 P43 43_ContactMapPrediction.py

error —

L2 P44 44_ReparameterizationTrick.py

error —

L2 P45 45_AxialAttention.py

error —

L2 P46 46_SteinVariationalGradient.py

correct 4.29x 0.2μs

L2 P47 47_virtual_all_gather_concat_lin

correct 1.12x 0.0μs

L2 P48 48_ParticleFilter.py

error —

L2 P49 49_sum_gelu_across_virtual_ranks

correct 1.44x 0.0μs

L2 P50 50_HiddenMarkovForward.py

error —