There are N threads in a process. And each thread creates an Maxine VFX instance，and each thread pushes a different cuda context. An error is reported during operation, as shown in the following figure.
Must multiple Maxine VFX instances in the same process use the same cuda context?
