Arvind Sudarsanam

> > I'd like to get some architecture and code-organzation level comments on my SYCL thinLTO prototype. > > The feature isn't ready for code-review, so I'd appreciate keeping the...

One of the things I see in the testcase is this: ` static const char* __DEVICE_CONSTANT__ format = FORMAT; ` According to the OpenCL spec, the format string must reside...

Interesting. I tried with the intel/llvm clang++ compiler and that did seem to work. Let me try with OneAPI compiler. Thanks

> Is the intel/llvm clang++ compiler better than using the oneapi compiler? I used the intel/llvm compiler built using sources and it is expected to be 'newer' than OneAPI compiler....

> I think we are going to new remove this tool soon. We are refactoring the compilation process for offload code and new approach won't use this tool or similar...

I am looking into this now. Thanks

Such info can be found in specification documents. https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_API.html#_memory_model https://registry.khronos.org/SYCL/specs/sycl-2020/html/sycl-2020.html#sec:memory.model https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#memory-model Hope that gets you started. Thanks

Hi @wangzy0327 I tried to compile your code using 'clang++ -fsycl test.cpp'. Hope that is the right way. I ran into a few issues. When I looked closer at your...

Hi @Pennycook I also tried --save-temps and looked at the report. The report is huge. But the expected 8 lines are at the very end of the report and can...

Code format failure is in a location which is not touched by this change. Thanks.