[PATCH i-g-t v2 0/5] Replace loop in WMTP tests

[Zbigniew Kempczyński]

Current solution where we run loop with some number of iterations
is problematic because we tweak this number for different platforms.

Start to migrate to loop stoppable from cpu side, where we can
immediately stop long running job via simple memory write instead
of waiting until loop will complete all iterations.

v2: - adjust test number of children according to available ram
    - add compute-preempt-many-all-ram subtest

Cc: Francois Dugast <francois.dugast at intel.com>
Cc: Priyanka Dandamudi <priyanka.dandamudi at intel.com>

Zbigniew Kempczyński (5):
  lib/intel_compute: add support for stoppable loop
  lib/intel_compute_square_kernels: use stoppable loop for LNL/BMG
  lib/intel_compute_square_kernel: add loop shader binary for PTL
  tests/xe_compute_preempt: adjust number of children according to ram
    size
  tests/xe_compute_preempt: consume all ram for wmtp

 lib/intel_compute.c                | 39 ++++++++++++---
 lib/intel_compute.h                |  2 +
 lib/intel_compute_square_kernels.c | 77 ++++++++++++++++++++++++++++++
 opencl/loop.cl                     |  9 ++++
 tests/intel/xe_compute_preempt.c   | 54 ++++++++++++++++++---
 5 files changed, 168 insertions(+), 13 deletions(-)
 create mode 100644 opencl/loop.cl

-- 
2.34.1