We've been observing some of the gfx942 runners with errors on test initialization that then resolve when retried on a different host. Using this issue to log the errors and runner as it happens to keep track of the pattern:
Runner name: 'linux-gfx942-1gpu-core42-ossci-rocm-jj2tj-runner-58ww6' : INVALID_ARGUMENT: runtime/src/iree/hal/drivers/amdgpu/util/vmem.c:194: INVALID_ARGUMENT; [hsa_amd_vmem_address_reserve_align] HSA_STATUS_ERROR_INVALID_ARGUMENT: One of the actual arguments does not meet a precondition stated in the documentation of the corresponding formal argument.
We've been observing some of the gfx942 runners with errors on test initialization that then resolve when retried on a different host. Using this issue to log the errors and runner as it happens to keep track of the pattern:
Runner name: 'linux-gfx942-1gpu-core42-ossci-rocm-jj2tj-runner-58ww6' :
INVALID_ARGUMENT: runtime/src/iree/hal/drivers/amdgpu/util/vmem.c:194: INVALID_ARGUMENT; [hsa_amd_vmem_address_reserve_align] HSA_STATUS_ERROR_INVALID_ARGUMENT: One of the actual arguments does not meet a precondition stated in the documentation of the corresponding formal argument.