Re: Contribute a RISC-V 64 JIT backend
Logan Chien <[email protected]> Mon, 19 Feb 2024 20:08:22 -0800
| Newsgroups | gmane.comp.python.pypy |
|---|---|
| Message-ID | <CALQyFuCp5Q0iY+YxXNSg5W0-2HpZXKKE6iwo7NvZx7Rf42jbeQ@mail.gmail.com> |
--===============1720811988393932531== Content-Type: multipart/alternative; boundary="00000000000013ea2b0611c85d86" --00000000000013ea2b0611c85d86 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Hi Armin, Thank you for the reply. > This should just be #defined to do nothing with Boehm, maybe in rpython/translator/c/src/mem.h With this change and a few RISC-V backend fixes (related to self.cpu.vtable_offset), I can build and run a JIT+BoehmGC PyPy. This configuration (JIT+BoehmGC) can pass test_tokenize and test_zipfile64 (from lib_python_tests.py). Thus, my next step will focus on the differences between JIT+BoehmGC and JIT+IncminimarkGC. I think the differences are: 1. call_malloc_nursery_* (fast/slow paths) 2. shadow stack push/pop updates 3. gcmap 4. write barriers (fast/slow paths) 5. realloc_frame Regards, Logan On Sun, Feb 18, 2024 at 11:06=E2=80=AFPM Armin Rigo <[email protected]> = wrote: > Hi Logan, > > On Mon, 19 Feb 2024 at 05:02, Logan Chien <[email protected]> > wrote: > > 2890 | OP_GC_INCREASE_ROOT_STACK_DEPTH(l_v498959, /* > nothing */); > > Ah, yet another missing macro. This should just be #defined to do > nothing with Boehm, maybe in rpython/translator/c/src/mem.h in the > section "dummy version of these operations, e.g. with Boehm". > > > Just to be sure, is the following command correct? > > > > python2.7 ./pytest.py rpython/jit/backend/test/test_zll_stress_0.py -s = -v > > Yes, that's correct. > > > A bient=C3=B4t, > Armin > --00000000000013ea2b0611c85d86 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div>Hi Armin,</div><div><br></div><div>Thank you for the = reply.<br></div><div><br></div><div>> This should just be #defined to do= nothing with Boehm, maybe in rpython/translator/c/src/mem.h</div><div><br>= </div><div>With this change and a few RISC-V backend fixes (related to self= .cpu.vtable_offset), I can build and run a JIT+BoehmGC PyPy.</div><div><br>= </div><div>This configuration (JIT+BoehmGC) can pass test_tokenize and test= _zipfile64 (from lib_python_tests.py).</div><div><br></div><div>Thus, my ne= xt step will focus on the differences between JIT+BoehmGC and JIT+Incminima= rkGC.</div><div><br></div><div>I think the differences are:</div><div><br><= /div><div>1. call_malloc_nursery_* (fast/slow paths)<br></div><div>2. shado= w stack push/pop updates</div><div>3. gcmap</div><div>4. write barriers (fa= st/slow paths)</div><div>5. realloc_frame</div><div><br></div><div>Regards,= </div><div>Logan<br></div><div><br></div></div><br><div class=3D"gmail_quot= e"><div dir=3D"ltr" class=3D"gmail_attr">On Sun, Feb 18, 2024 at 11:06=E2= =80=AFPM Armin Rigo <<a href=3D"mailto:[email protected]">armin.rigo@= gmail.com</a>> wrote:<br></div><blockquote class=3D"gmail_quote" style= =3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding= -left:1ex">Hi Logan,<br> <br> On Mon, 19 Feb 2024 at 05:02, Logan Chien <<a href=3D"mailto:tzuhsiang.c= [email protected]" target=3D"_blank">[email protected]</a>> wrote:<= br> >=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 2890 |=C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0OP_GC_INCREASE_ROOT_STACK_DEPTH(l_v498959, /* nothing */);<br> <br> Ah, yet another missing macro.=C2=A0 This should just be #defined to do<br> nothing with Boehm, maybe in rpython/translator/c/src/mem.h in the<br> section "dummy version of these operations, e.g. with Boehm".<br> <br> > Just to be sure, is the following command correct?<br> ><br> > python2.7 ./pytest.py rpython/jit/backend/test/test_zll_stress_0.py -s= -v<br> <br> Yes, that's correct.<br> <br> <br> A bient=C3=B4t,<br> Armin<br> </blockquote></div> --00000000000013ea2b0611c85d86-- --===============1720811988393932531== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ pypy-dev mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/pypy-dev.python.org/ Member address: [email protected] --===============1720811988393932531==--