devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Replace memset with fill to work for non-trivial scalars.

Browse Source 
For custom scalars, zero is not necessarily represented by
a zeroed-out memory block (e.g. gnu MPFR). We therefore
cannot rely on `memset` if we want to fill a matrix or tensor
with zeroes. Instead, we should rely on `fill`, which for trivial
types does end up getting converted to a `memset` under-the-hood
(at least with gcc/clang).

Requires adding a `fill(begin, end, v)` to `TensorDevice`.

Replaced all potentially bad instances of memset with fill.

Fixes #2245.
This commit is contained in:
Antonio Sanchez
2021-05-11 09:52:00 -07:00
committed by Rasmus Munk Larsen
parent e9c9a3130b
commit 1e6c6c1576
18 changed files with 229 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
bench/tensors/tensor_benchmarks.h
View File

@@ -564,9 +564,9 @@ for (int iter = 0; iter < 10; ++iter) {
// Initialize the content of the memory pools to prevent asan from
// complaining.
device_.memset(a_, 12, m_ * k_ * sizeof(T));
device_.memset(b_, 23, k_ * n_ * sizeof(T));
device_.memset(c_, 31, m_ * n_ * sizeof(T));
device_.fill(a_, a_ + m_ * k_, T(12));
device_.fill(b_, b_ + k_ * n_, T(23));
device_.fill(c_, c_ + m_ * n_, T(31));
}

6
bench/tensors/tensor_contract_sycl_bench.cc
View File

@@ -56,9 +56,9 @@ void contraction(const Device& device_, TensorIndex num_iters, TensorIndex m_, T
// Initialize the content of the memory pools to prevent asan from
// complaining.
device_.memset(a_, 12, m_ * k_ * sizeof(T));
device_.memset(b_, 23, k_ * n_ * sizeof(T));
device_.memset(c_, 31, m_ * n_ * sizeof(T));
device_.fill(a_, m_ * k_, T(12));
device_.fill(b_, k_ * n_, T(23));
device_.fill(c_, m_ * n_, T(31));
Eigen::array<TensorIndex, 2> sizeA;
sizeA[0] = m_;