feat: Intel GPU Max (Ponte Vecchio) OpenMP target offload support#1445

Draft

sbryngelson wants to merge 35 commits into

Commits on May 15, 2026

improve: Intel GPU Max (Ponte Vecchio) OpenMP target offload support

sbryngelson
and
claude
committed
fix: ifx SPIR64 inliner ICE in m_compute_levelset (split loops + -fno-inline)

sbryngelson
and
claude
committed
docs: chemistry/pyrometheus Intel GPU verification and warning note

sbryngelson
and
claude
committed
fix: ifx SPIR64 InvalidArraySize — assumed-shape + local VLA fixes for OpenMP device routines
sbryngelson
committed
fix: ifx SPIR64 Level Zero link failures for Intel GPU Max (Ponte Vecchio)
sbryngelson
committed
perf: Intel GPU Max Level Zero tuning in run toolchain
sbryngelson
committed
fix: restore ifort (Intel classic) CPU build support in CMakeLists.txt
sbryngelson
committed
feat: add GT CRNCH RoboGator (crnch) module entry for Intel GPU Max
sbryngelson
committed
fix: find clang-offload-bundler via ifx PATH when FC=mpiifx
sbryngelson
committed
fix: LAPACK PIE link error + crnch module Intel MPI paths
sbryngelson
committed
fix: LAPACK FortranCInterface PIE link failure on Ubuntu 22.04 with ifx
sbryngelson
committed
fix: build FFTW from source for all non-Cray compilers including IntelLLVM
sbryngelson
committed
Merge remote-tracking branch 'origin/master' into intel-gpu
sbryngelson
committed
fix: crnch module — add I_MPI_FABRICS=shm for Intel MPI on single-node
sbryngelson
committed

Commits on May 18, 2026

Commits on May 19, 2026

Commits on May 20, 2026

Merge branch 'master' into intel-gpu
sbryngelson
authored