| Summary: | ASSERTION FAILED: sel.hasDoubleType() | ||
|---|---|---|---|
| Product: | Beignet | Reporter: | Chris Cummins <chrisc.101> |
| Component: | Beignet | Assignee: | Zhigang Gong <zhigang.gong> |
| Status: | RESOLVED NOTABUG | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | unspecified | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: | Debug log | ||
This is expected (though the error message could be more explicit): double precision is not supported on Haswell (and is disabled by default even on hardware that does support it). Sure, I can agree that the code may not be expected to compile. The reason I filed a bug however is because a compiler assertion, whether or not it contains a meaningful assertion message, does not seem like the appropriate response to an input. An error diagnostic and CL_BUILD_PROGRAM_FAILURE is the expected outcome. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
Created attachment 133652 [details] Debug log Compilation of the following kernel fails with assertion 'sel.hasDoubleType()': kernel void A(global double2* a) { a[0] = get_global_id(0); } ASSERTION FAILED: sel.hasDoubleType() at file ../backend/src/backend/gen_insn_selection.cpp, function void gbe::ConvertInstructionPattern::convertSmallIntsToDouble(gbe::Selection::Opaque&, const gbe::ir::ConvertInstruction&, bool&) const, line 6366 Device: Intel(R) HD Graphics Haswell GT2 Desktop Beignet version: 1.3 Full debug log attached.