Richard Membarth

Test case: ```rust fn cuda_clock() -> u32 { let mut cycle_count:u32; asm("mov.u32 %0, %clock;" : "=r"(cycle_count) ::: "volatile"); cycle_count } ``` ```bash # > impala test.impala impala: anydsl/impala/src/impala/parser.cpp:1439: std::string impala::Parser::parse_str():...

Example: `static foo = bar();`

LLVM annotates the private address space for amdgpu (but not for nvptx). This triggers the assertion "Calling a function with a bad signature!" when generating amdgpu code for the following...

Partial evaluator too eager cont.

3

The following example should specialize ```specialize```, but it doesn't since PE is too eager. ```rust extern "thorin" { fn pe_info[T](&[u8], T) -> (); } enum Num { one, two }...

top-level continuation 'name' got free def

6

We get this assertion when compiling the traversal for GPU (nvvm & amdgpu). When commenting out unused code in the ```intrinsics_amdgpu.impala``` (amdgcn_atomic_* lines 81 - 93), the code compiles for...

thorin::Node_TypeParam not handled in CodeGen::convert(Type type)

4

Thorin segfaults since `thorin::Node_TypeParam` is nod handled in `CodeGen::convert(Type type)`, see commit AnyDSL/impala@93237d5

clEnqueueSVMMemcpy SegFault

4

Running `clEnqueueSVMMemcpy(queue, CL_TRUE, dst, src, size, 0, NULL, NULL);` with dst allocated with `clSVMAlloc` and src allocated by the system (e.g. posix_memalign) triggers a segmentation fault: ```bash Thread 4 "Command...

The `master` branch misses the `tools/gen_cpp.py` for building sleef bc from source.