Skip to content

Commit

Permalink
Completed updates for 0.3.2 release.
Browse files Browse the repository at this point in the history
  • Loading branch information
jlparkI committed Apr 2, 2024
1 parent 40a450e commit 1dd8404
Show file tree
Hide file tree
Showing 4 changed files with 29 additions and 1 deletion.
7 changes: 7 additions & 0 deletions HISTORY.md
Original file line number Diff line number Diff line change
Expand Up @@ -186,3 +186,10 @@ padding we use is automatically masked if user so desires.
### Version 0.3.1

Fixed bug in static layers module.

### Version 0.3.2

Extensive modifications to wrapped C++ code for CPU convolution kernels
specifically, achieving a 25% speedup and a large reduction in memory
usage, which is now constant irrespective of sequence length. Fixed
bug in sequence length normalization.
20 changes: 20 additions & 0 deletions test/fht_operations_tests/basic_fht_functions_test.py
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,16 @@ def test_3d_array_transform(self):
self.assertTrue(outcome_d)
self.assertTrue(outcome_f)

dim = (124,36,2)
outcome_f, outcome_d = run_fht_test(dim)
self.assertTrue(outcome_d)
self.assertTrue(outcome_f)

dim = (124,36,32)
outcome_f, outcome_d = run_fht_test(dim)
self.assertTrue(outcome_d)
self.assertTrue(outcome_f)

dim = (3001, 5, 1024)
outcome_f, outcome_d = run_fht_test(dim)
self.assertTrue(outcome_d)
Expand All @@ -55,6 +65,16 @@ def test_2d_array_transform(self):
self.assertTrue(outcome_d)
self.assertTrue(outcome_f)

dim = (124, 2)
outcome_f, outcome_d = run_fht_2d_test(dim)
self.assertTrue(outcome_d)
self.assertTrue(outcome_f)

dim = (124, 32)
outcome_f, outcome_d = run_fht_2d_test(dim)
self.assertTrue(outcome_d)
self.assertTrue(outcome_f)

dim = (3001, 1024)
outcome_f, outcome_d = run_fht_2d_test(dim)
self.assertTrue(outcome_d)
Expand Down
2 changes: 1 addition & 1 deletion xGPR/fitting_toolkit/cg_fitting_toolkit.py
Original file line number Diff line number Diff line change
Expand Up @@ -113,7 +113,7 @@ def cg_fit_lib_ext(kernel, dataset, cg_tol = 1e-5, max_iter = 500,
cg_operator = CPU_CGLinearOperator(dataset, kernel,
verbose)
weights, convergence = CPU_CG(A = cg_operator, b = z_trans_y,
M = preconditioner, rtol = cg_tol, atol = 0, maxiter = max_iter)
M = preconditioner, tol = cg_tol, atol = 0, maxiter = max_iter)


if convergence != 0:
Expand Down
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
#ifndef HADAMARD_TRANSFORM_OPERATIONS_H
#define HADAMARD_TRANSFORM_OPERATIONS_H

#define MIN(a,b) ((a) < (b) ? (a) : (b))

template <typename T>
void transformRows(T __restrict xArray[], int startRow, int endRow,
Expand Down

0 comments on commit 1dd8404

Please sign in to comment.