Hi! On Fri, 6 Nov 2015 12:03:25 +0100, Bernd Schmidt <bschm...@redhat.com> wrote: > On 11/06/2015 11:30 AM, Richard Biener wrote: > > On Fri, 6 Nov 2015, Bernd Schmidt wrote: > >> > >> Realistically we're probably not going to reject this work, but I still > >> want > >> to ask whether the approach was acked by the community before you started. > >> I'm > >> really not exactly thrilled about having two different classes of backends > >> in > >> the compiler, and two different ways of handling offloading. > > > > Realistically the other approaches werent acked either (well, implicitely > > by review). > > I think the LTO approach was discussed beforehand. As far as I remember > (and Jakub may correct me) it was considered for intelmic, and Jakub had > considerable input on it. I heard that it came up at the 2013 Cauldron. > Writing an rtl backend is the default thing to do for gcc and I would > expect any other approach to be discussed beforehand. > > > Not doing an RTL backend for NVPTX would have simplified > > your life as well. > > I'm not convinced about this. At least I just had to turn off the > register allocator, not write a new one.
From the notes of the Accelerator BoF at the GNU Tools Cauldron 2013, <http://news.gmane.org/find-root.php?message_id=%3C1375103926.7129.7694.camel%40triegel.csb%3E>: | The main issue we discussed in the backend category was how to target | more than one ISA when generating code (i.e., we need code in the host's | ISA and in the accelerator(s)' (virtual) ISA(s)). Multi-target support | in GCC might be one option, but would probably need quite some time and | thus depending on it would probably delay the accelerator efforts. It | might be simpler to stream code several times to different backends | using the LTO infrastructure. [...] A third | option that SuSE is experimenting with is not writing a new backend but | instead generating code right after the last GIMPLE pass; however, HSAIL | needs register allocation, so it was noted that writing a light-weight | backend might be | easier. Grüße Thomas
signature.asc
Description: PGP signature
