summaryrefslogtreecommitdiff
path: root/csrc/kernels.cuh
diff options
context:
space:
mode:
authorTim Dettmers <tim.dettmers@gmail.com>2022-07-26 19:38:01 -0700
committerTim Dettmers <tim.dettmers@gmail.com>2022-07-26 19:38:01 -0700
commit5737f2b027a1e0ec8540a3aa914632d44ad9c62d (patch)
treeb288c905eaba75dc6b43a8bcebc82720c16e4816 /csrc/kernels.cuh
parent47a73d94c3d3284f6073b0ff189ed5bc9e3a8762 (diff)
parentdc8c9efdb33130f960adc864916b67d0cb744dbb (diff)
Merge branch 'patch_merge' into extract_outliers
Diffstat (limited to 'csrc/kernels.cuh')
-rw-r--r--csrc/kernels.cuh2
1 files changed, 1 insertions, 1 deletions
diff --git a/csrc/kernels.cuh b/csrc/kernels.cuh
index eda5ba0..2447494 100644
--- a/csrc/kernels.cuh
+++ b/csrc/kernels.cuh
@@ -107,7 +107,7 @@ template<typename T, int BLOCK_SIZE, int NUM_VALS> __global__ void kPercentileCl
__global__ void kHistogramScatterAdd2D(float* histogram, int *index1, int *index2, float *src, const int maxidx1, const int n);
-template <typename T, int SPMM_ITEMS, int BITS> __global__ void kspmm_coo_very_sparse_naive(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, T *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
+template <typename T, int SPMM_ITEMS, int BITS> __global__ void kspmm_coo_very_sparse_naive(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, T *B, half *out, float * __restrict__ const dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template <int ITEMS_PER_THREAD, int SUBTILE_ROWS, int THREADS>__global__ void kdequant_mm_int32_fp16(
int *__restrict__ const A, float *__restrict__ const rowStats, float *__restrict__ const colStats,