https://bugs.llvm.org/show_bug.cgi?id=43444
Bug ID: 43444
Summary: When double hyperbolic and trigonometric functions are
combined in a single kernel the compiler hangs
Product: clang
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: OpenCL
Assignee: unassignedclangb...@nondot.org
Reporter: kola...@mail.ru
CC: anastasia.stul...@arm.com, llvm-bugs@lists.llvm.org
Created attachment 22570
--> https://bugs.llvm.org/attachment.cgi?id=22570&action=edit
test.cl
When compiling a kernel dealing to complex math I had encountered a problem
that the compiler hangs. It hangs when llvm-based runtimes are used, no matter
cpu or gpu based. The problem is reproduced (on my machine) in both llvm-8 and
llvm-10 based runtimes, more precisely cpu and gpu llvm-8 Mesa Clover runtimes
installed from Ubuntu disco packages and llvm-10-based (llvm-10 and clang-10
are installed from apt.llvm.org) self-compiled pocl (with the patches making it
compatible to llvm 10).
On Windows the kernels have been built successfully using AMD APP SDK (I
haven't tried pocl).
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
