Standardize printf

### Extension Type

New Feature

### Feature Description

> PS: It would be nice if you would help echoing in the SYCL committee that `printf` is something that there is significant demand for downstream. @masterleinad was recently reporting how painful it was w/o it when he attempted to run the Kokkos SYCL backend on AMD accelerators. On the Kokkos user side, we also have strong demand for it and now provide it in our API https://kokkos.org/kokkos-core-wiki/API/core/utilities/printf.html

_Originally posted by @dalg24 in [#396](https://github.com/kokkos/mdspan/issues/396#issuecomment-2779453493)_

Both Intel and AdaptiveCPP support some flavor of `printf`. Without thinking too much, it seem to be a good khr idea, and potentially adding a querry to check for device support ( I guess FPGA will not like `printf` 🤷🏽 )

### Related Functionality in C++

printf / std::format

### Related Functionality in Other Languages

print

### Related SYCL Extensions

_No response_

Review the proposal to standardize printf for SYCL, including the related C++ printf/std::format functionality and existing Intel and AdaptiveCPP support. Assess whether device-support queries are needed, including FPGA considerations; done would require a decided specification design rather than an implementation-only change.

Written by the indexing model from the issue text.