On 22 April 2018 at 21:35, Richard Henderson richard.henderson@linaro.org wrote:
I'll note that a lulesh binary that I generated for HKG18 does not work on the branch, but does under FoundationModel. Therefore there must be remaining emulation errors. However, it's a really complicated binary that is non-debuggable. Adding printfs interferes with vectorization and gdb does not yet let us print all of the relevant registers.
I have had some luck in this sort of situation with using both the fast model's instruction tracing and QEMU's and then doing enough post-processing of both traces to get a list of PC values which can then be diffed to find out where execution diverges.
You are probably right to start by fixing the bugs in the smaller test cases first, though.
thanks -- PMM