Lines Matching +full:vm +full:- +full:pointer

11 as happens for heap-allocated objects, frames are allocated
12 contiguously in a per-thread stack. 
14 * It reduces allocation overhead to a pointer comparison and increment.
19 can be linked into the per-thread stack so as to not impact performance too much.
27 * Specials: The per-frame object references needed by the VM: globals dict,
29 * Linkage: Pointer to the previous activation record, stack depth, etc.
41 It needs the interpreter to hold two pointers, a frame pointer and a stack pointer.
53 location for the parameters. However, it requires the VM to maintain
54 an extra pointer for the locals, which can hurt performance.
58 is numbered `N-1`.
59 This allows the locals, specials and linkage to accessed from the frame pointer.
81 The pointer to the function is not strictly required, but it is cheaper to
92 The `PyFrameObject` may outlive a stack-allocated `_PyInterpreterFrame`.
97 This mechanism provides the appearance of persistent, heap-allocated
106 The embedded frame is linked into the per-thread frame when iterated or