rocm-systems

Автор	SHA1	Сообщение	Дата
Siu Chi Chan	f6837a4e7f	minor workaround for lazy binding [ROCm/clr commit: `b5045af7e9`]	2019-04-02 17:28:06 -04:00
Siu Chi Chan	f7817c48d5	move triple_to_hsa_isa into the header [ROCm/clr commit: `30053f6443`]	2019-03-11 19:51:44 +00:00
Yaxun (Sam) Liu	733c5b98a0	Fix build failure in code_object_bundle.cpp [ROCm/clr commit: `e6d155bcd5`]	2018-07-10 16:49:59 -04:00
Aaron Enye Shi	2071f27bb0	Clean up triple_to_hsa_isa naming (#403 ) [ROCm/clr commit: `a004c8dfa2`]	2018-04-17 06:47:55 +05:30
Aaron Enye Shi	5d295abb55	Fix bug when using older rocr naming (#399 ) [ROCm/clr commit: `21382e361c`]	2018-04-14 15:47:46 +05:30
Aaron Enye Shi	03877a5ed2	Fix new gfx target naming convention (#397 ) This was introduced in recent xnack changes naming changes which affected our offload bundler. This patch will fix issues in HIP samples, ROCR tests, and others. [ROCm/clr commit: `199cff2365`]	2018-04-13 16:24:39 +05:30
Maneesh Gupta	46ddefedee	Apply .clangformat to all repo source files Change-Id: I7e79c6058f0303f9a98911e3b7dd2e8596079344 [ROCm/clr commit: `9e47fccc89`]	2018-03-12 11:29:03 +05:30
Alex Voicu	aa48cc7b55	This introduces LipoProteinLipase (lpl), a simple tool for creating fat binaries. It represents a direct replacement of the creaky hccgenco.sh script, which had various issues. The format it uses is that of a code object bundle, generated by the Clang Offload Bundler. The output is always suffixed with the ".adipose" extension. It is shared with HCC. The hipcc script and associated tests are modified to use lpl. Help can be obtained by invoking lpl --help. A more computer-sciency / corporate friendly name is likely to be beneficial, which is a reason for choosing easily searchable/replaceable names such as lpl or adipose. [ROCm/clr commit: `4e0739c68a`]	2017-12-08 04:22:57 +00:00
Alex Voicu	b738633bcb	This switches HIP from its currently convoluted macro + pfe based dispatch mechanism to a more natural one partially based on the existing module API. The basic idea is that HCC will always correctly emit __global__ functions: as empty-bodied stubs, on host, and as kernels, on device. It then becomes trivial to obtain the mangled name on host, at dispatch, from the function's address, and then to use the mangled name to retrieve the kernel. This should address all problems stemming from serialisation, dubious mismatches due to the manufactured functor, macro-isms et al. It also immediately enables support for generalised globals as a consequence of that being available in the module API. Finally, it will make debug much easier, since the actual names of the __global__ functions will automatically be used in traces etc. One detail is that due to how dispatch works now (hipLaunchKernel and hipLaunchKernelGGL are themselves variadic function templates which deduce the function type of the callee), in certain cases it may be necesssary to insert explicit casts to ensure that the variadic argument list selects a viable overload - this can be observed in some unit tests. Eventually we may be able to remove this limitation, but for now it does not appear terribly onerous. The code is not extremely HIPpie, nor is it fully optimised, but rather is intended as a starting point for the HIP team to make its own. [ROCm/clr commit: `28f87f7d2e`]	2017-11-01 15:09:59 +00:00

9 Коммитов