Intel® oneAPI Data Parallel C++
Support for Intel® oneAPI DPC++ Compiler, Intel® oneAPI DPC++ Library, Intel ICX Compiler , Intel® DPC++ Compatibility Tool, and GDB*
581 Discussions

DPC++ item's id error in the kernel function

shenuo
Employee
616 Views

gpu_selector gs;
cpu_selector cs;
queue Q(cs);
queue Q2(gs);

event ev = Q.submit([&](handler& h) {
auto out = sycl::stream(1024, 1024 * 2, h);

h.parallel_for(nd_range(range{960, 540}, range{ 2,2 }), [=](nd_item<2> it) {
out << it.get_global_id()[0] << " " << it.get_global_id()[1] << "\n";
});
});

ev2.wait();

 

when use Q2, the id output is not reasonable, while Q1 is correct. it seems the has relationship with gpu.

could any one tell me what's wrong here.

my gpu is UHD630, cpu is Intel(R) Core(TM) i7-10610U CPU

OS is win10, IDE is VS2019.

 

0 Kudos
4 Replies
HemanthCH_Intel
Moderator
577 Views

Hi,


Thanks for posting in Intel Communities.


Could you please provide the screenshots of expected results and observed results?


Thanks & Regards,

Hemanth


0 Kudos
shenuo
Employee
570 Views

thanks for your reply.

just give you one piece of the result :

2488645 0
2488645 1
2488646 0
2488646 1
1251528 2
1251528 3
1251529 2
1251529 3
37765 2
15048 0
37765 3
37766 2
15048 1
37766 3
15049 0
15049 1

you can see the value first dimension is wrong, for the max value is just 960 !

0 Kudos
shenuo
Employee
567 Views

Hi,

  I think this must be related to the graphics or driver problem.

  I find it works fine on another machine

0 Kudos
HemanthCH_Intel
Moderator
542 Views

Hi,


Glad to know that your issue is resolved. If you need any additional information, please post a new question as this thread will no longer be monitored by Intel.


Thanks & Regards,

Hemanth


0 Kudos
Reply