User Details
- User Since
- Jan 18 2017, 9:43 AM (279 w, 2 d)
Aug 19 2020
I'll add a test to it. Please give me some time as I am busy with other stuffs.
Jul 27 2020
Jul 26 2020
Your understanding above is exactly right. It should be equivalent to
#pragma omp target data map(tofrom: c) #pragma omp target data map(tofrom: c.a[0:NUM]) #pragma omp target teams distribute parallel for for (int i = 0; i < NUM; i++) { ++c.a[i]; }
Jul 25 2020
Jul 24 2020
Thanks George for looking into this, and sorry for the late response.
The problem you had is probably not related to this patch. A lot of libomptarget tests are not working for nvptx now. I suppose someone is going to fix them.
Jul 23 2020
Can you check if https://reviews.llvm.org/D84470 fixes the problem as this test doesn't work on my machine? If yes, please accept and upstream it since I don't have the permission.
Jul 22 2020
I couldn't reproduce this problem because all commands are ignored on my machine. Could you send the exact compiling and running commands?
In the IR files you sent me before, any idea where segfault happens?
Thanks. Another weird place is it passes with nvprof. Not sure why using nvprof makes a difference here.
I can reproduce this. When running the test itself, Sum=1024.
When running the test with nvprof, Sum=2048. Combining with that you said Sum=2048 when LIBOMPTARGET_DEBUG=1, I suspect the GPU offloading is disabled in the above case. Any idea what happened recently to libomptarget which can potentially cause this problem? I didn't follow the recent development so have no idea.
Jul 20 2020
Jul 16 2020
Thanks George!
Fix is at https://reviews.llvm.org/D83959
@Meinersbur Michael, could you please help upstream the patch above? Thanks
Jul 15 2020
Include the llvm part
@grokos Could you upstream this patch and the runtime patch neck to neck? Upstreaming one of them will break the OpenMP offloading. It will be nice if you can test this patch locally with the runtime patch. Thanks
Update diff and pass test. Please accept
Jul 14 2020
I'll update the diff and please check and accept after that
Also, not sure if https://reviews.llvm.org/D71782 is still needed. Please check
The codegen patch needs to land before this one to pass these test programs
Mar 30 2020
Feb 14 2020
Test cases will be uploaded in another patch when the Clang patch is upstreamed. That Clang patch depends on this (https://reviews.llvm.org/D67833). So I think the order is this patch, clang patch, test patch.
Dec 23 2019
Address Jon's comments
Rebase and rediff with the nfc version
Dec 21 2019
Yes, this is the NFC part of the declare mapper patch. I've tested it. There is not much test in libomptarget though.
Dec 20 2019
Dec 18 2019
Change the function name
Address Alexey's comments to change mapper function size and refactor code
Dec 17 2019
Dec 16 2019
Alexey, thanks for the review
I'll split the patch into 2 later
Rebase and address comments
Nov 25 2019
Thanks for your reviews. Hope this looks better.
Nov 15 2019
Nov 13 2019
Rebase
Thanks Alexey and Jon for your review. Fixed the issues and rebased
Sep 27 2019
Alexey and George: This is a big decision to make. We need to have most people's consents. I'll send it to the mailing list later.
Sorry, you are right. I didn't think about the case to always clean up the mappers after finishing using it.
Please review when you have time, thanks
Sep 26 2019
Rebase
Thanks Alexey! Please check the other 2 mapper patches at https://reviews.llvm.org/D67833 and https://reviews.llvm.org/D68100 when you have time. They should be the last mapper patches.
Fix mapper type checking
Sep 25 2019
HI Alexey, the ast print test is already there. Because I didn't check the mapper for array type before, such code will always not report any error, and ast print test is correct. Codegen test belongs to the other patch I released. It fits that patch much better.
Sep 24 2019
Without this patch, it cannot recognize array with mapper, for instance, #pragma omp target map(mapper(a),to: arr[0:2]) won't work without this patch.
Sep 20 2019
Aug 5 2019
Fix declare mapper codegen test when the function argument has name attached.
Aug 2 2019
Hi David,
Aug 1 2019
@Meinersbur Hi Michael, could you help me commit this patch? Thanks!
Jul 31 2019
Add a test
Jul 30 2019
Change mapper function argument checking
Jul 29 2019
Thanks Alexey! Could you look into the runtime patch D60972 then?
Ping and rebase
Jul 26 2019
Make emitUDMapperArrayInitOrDel private