Marcin Hajder
Marcin Hajder
> Note, the barrier test is currently using a single producer and a single consumer. Ideally we would have two producers and two consumers instead... @bashbaug I will refactor the...
> Any chance you could "git mv" the prior test_external_semaphore_sync_fd.cpp to test_external_semaphore_handle_types.cpp? I think that will let us see the deltas between the two files vs. treating them as entirely...
> We would like to run this test since it is a new builtin being tested but could you rebase to tip first? @lakshmih done
> Can we update the reference algorithm? It produces wrong output for this result: > > Input -> -1.92592994438723585306e-34 Reference value - Wolfram -> -5.19229685853482762853e+33 Reference value - CTS ->...
> We’re now able to compute the correct reference values. However, we had a question regarding the required precision for the host reference: > > Is 64-bit precision sufficient for...
> We’ve noticed that using 128-bit significantly increases execution time, so we’d like to confirm whether that level of precision is truly required. @lakshmih as discussed in teleconference here is...
I've tested this PR with linux and windows, currently I am not able to launch the test on MacOS - I will appreciate confirmation from this platform. GLES version does...
@bashbaug thanks for your feedback. The confusion around `cl_context_properties` comes from the differencies in creating context between platforms. Please note that for OSX and X11 existing setup does not use...
@lakshmih similar situation to #1996, I compared all reference values with equivalents computed with single precision - there are no differences but still I've created the PR to use single...