Reducing the _X64WriteBarrierCardTableManager::_cardTable memory reservation

Hi, we are trying to reduce the memory reservation for the `_X64WriteBarrierCardTableManager::_cardTable`.
(for windows x64 with address space of 128TB it is reserving 32GB).
The default page size is 4KB and we are trying to make it 64 times bigger, i.e. 256KB.
We changed the following constants:
- [RecyclerWriteBarrierManager::s_WriteBarrierPageSize](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Memory/RecyclerWriteBarrierManager.h#L214), [RecyclerWriteBarrierManager::s_BytesPerCard](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Memory/RecyclerWriteBarrierManager.h#L217), [AutoSystemInfo::PageSize](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Core/SysInfo.h#L65), [CompileAssert in SysInfo.h](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Core/SysInfo.h#L138), [__in_ecount_pagesize](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Core/SysInfo.h#L139) and
[__in_ecount_twopagesize](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Core/SysInfo.h#L140) multiplied by 64.
- [HeapBlockMap::PagesPer4GB changed to 1 << 14](https://github.com/Microsoft/ChakraCore/blob/master/lib/Common/Memory/HeapBlockMap.h#L292)

Is it safe to do it and is there a better way of doing it?
Thanks in advance.

Start by reviewing the constants and CompileAssert locations named in RecyclerWriteBarrierManager.h, SysInfo.h, and HeapBlockMap.h, then trace their usages in the memory-management code. Done means establishing whether the proposed page-size and card-table changes are safe and identifying a supported approach, with validation of the resulting memory reservation.

Written by the indexing model from the issue text.