Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tuning gpairs implementation params #341

Merged
merged 1 commit into from
Feb 23, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 7 additions & 7 deletions dpbench/benchmarks/gpairs/gpairs_numba_dpex_k.py
Original file line number Diff line number Diff line change
Expand Up @@ -37,11 +37,11 @@ def count_weighted_pairs_3d_intel_no_slm_ker(
lws0 = nd_item.get_local_range(0)
lws1 = nd_item.get_local_range(1)

n_wi = 20
n_wi = 32

dsq_mat = dpex.private.array(shape=(20 * 20), dtype=dtype)
w0_vec = dpex.private.array(shape=(20), dtype=dtype)
w1_vec = dpex.private.array(shape=(20), dtype=dtype)
dsq_mat = dpex.private.array(shape=(32 * 32), dtype=dtype)
w0_vec = dpex.private.array(shape=(32), dtype=dtype)
w1_vec = dpex.private.array(shape=(32), dtype=dtype)

offset0 = gr0 * n_wi * lws0 + lid0
offset1 = gr1 * n_wi * lws1 + lid1
Expand Down Expand Up @@ -81,7 +81,7 @@ def count_weighted_pairs_3d_intel_no_slm_ker(

# update slm_hist. Use work-item private buffer of 16 tfloat elements
for k in range(0, slm_hist_size, private_hist_size):
private_hist = dpex.private.array(shape=(16), dtype=dtype)
private_hist = dpex.private.array(shape=(32), dtype=dtype)
for p in range(private_hist_size):
private_hist[p] = 0.0

Expand Down Expand Up @@ -133,8 +133,8 @@ def gpairs(
rbins,
results,
):
n_wi = 20
private_hist_size = 16
n_wi = 32
private_hist_size = 32
lws0 = 16
lws1 = 16

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@ sycl::event gpairs_impl(sycl::queue q,
FpTy *hist)
{

const unsigned int n_wi = 20, private_hist_size = 16, lws0 = 16, lws1 = 16;
const unsigned int n_wi = 32, private_hist_size = 32, lws0 = 16, lws1 = 16;
const size_t m0 = static_cast<size_t>(n_wi) * static_cast<size_t>(lws0);
const size_t m1 = static_cast<size_t>(n_wi) * static_cast<size_t>(lws1);
const size_t n_groups0 = ceiling_quotient(n, m0);
Expand Down
Loading