Skip to content

test(adam): reset iterations so all tests pass with flang-new #179

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.

Sign up for GitHub

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

Merged
rouson merged 2 commits into from
Jul 19, 2024
Merged

test(adam): reset iterations so all tests pass with flang-new #179

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
72b7b00
test(adam): increase iterations with flang-new
rouson Jul 18, 2024
e8f1aa5
refac(test): reduce iterations with flang
rouson Jul 19, 2024
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
12 changes: 7 additions & 5 deletions test/trainable_engine_test_m.F90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -348,14 +348,16 @@ function xor_gate_with_random_weights() result(test_passes)
type(trainable_engine_t) trainable_engine
real(rkind), parameter :: tolerance = 1.E-02_rkind
real(rkind), allocatable :: harvest(:,:,:)
#ifndef __INTEL_COMPILER
#ifdef __flang__
!! Reducing num_iterations yields a less robust test, but moving away from local minima by
!! increasing num_iterations causes this test to crash when compiled with the flang or ifx compilers.
integer, parameter :: num_inputs=2, mini_batch_size = 1, num_iterations=50000
#elif defined __INTEL_COMPILER
integer, parameter :: num_inputs=2, mini_batch_size = 1, num_iterations=49000
#else
integer, parameter :: num_inputs=2, mini_batch_size = 1, num_iterations=500000
!! Depending on where in the random-number sequence the weights start, this test can pass for lower
!! numbers of iterations, e.g., 400000. Using more iterations gives more robust convergence.
#else
integer, parameter :: num_inputs=2, mini_batch_size = 1, num_iterations=49000
!! Reducing num_iterations yields a less robust test, but increasing num_iterations causes this
!! test to crash when compiled with the Intel ifx compiler.
#endif
integer batch, iter

Expand Down