trixirt

conformance_buffers_fill mostly works now on jetson, before it failed when the size was 8+ failures now are on the flags wanting to use host ptr.

yes i believe so, that cu api is a value not a pointer. I must have been getting lucky on the conformance test. I will rework.

Can you explain why a write and copies would not work ? The test i am fixing is conformance_buffers_fill

I will add a macro I am really looking at llvm 15 which causes many more similar problems. In 15 some of the functions are still available but assert at...

I see this error so I will handle. Handling these old releases is becoming more difficult, it would be good if the oldest (6) could be knocked off when the...

I am thinking looping over args can be handled better, here's where i am going so the #idef opaque ptr can be hidden in the macro. https://github.com/trixirt/pocl/commit/e9a23b52d77449908a538b05545edd821844cee1

I am in favor of doing just the recent llvms Fedora and RHEL are always moving to the latest llvm, that's 14 now and why I am doing these changes....

AFAIK it is ready but I have no idea how to test the TCE side. Where is the TCE toolchain etc. coming from ?

A user specific config may be necessary because the user may not be using the default compiler or some other knob syclcc.json controls. A user will not have write access...

CUDA is not open source, it can not be packaged, so it can not be used to build other packages. This will not prevent an user from building their own....