DPCT1085

Intel® DPC++ Compatibility Tool Developer Guide and Reference

Download PDF

ID 768918

Date 6/24/2024

Version

Public

A newer version of this document is available. Customers should click here to go to the newest version.

Visible to Intel only — GUID: GUID-969BEFDE-8A2F-4B29-8631-9E0CCF5463DD

View Details

DPCT1085

Message

The function <function name> requires sub-group size to be <size>, while other sub-group functions in the same SYCL kernel require a different sub-group size. You may need to adjust the code.

Detailed Help

Each kernel can only be decorated with one sub-group size. This warning is emitted when a kernel requires different sub-group sizes. Check if the sub-group size can be unified into one value, and if it cannot be unified, redesign the code logic.

For example, this original CUDA* code:

 __global__ void kernel(int* data1, int* data2) {
   typedef cub::WarpScan<int> WarpScan;
   typedef cub::WarpScan<int, 16> WarpScan16;

   typename WarpScan::TempStorage temp1;
   typename WarpScan16::TempStorage temp2;

   int input = data1[threadIdx.x];
   int output1 = 0;
   int output2 = 0;
   WarpScan(temp1).InclusiveSum(input, output1);
   data1[threadIdx.x] = output1;
   WarpScan16(temp2).InclusiveSum(input, output2);
   data2[threadIdx.x] = output1;
 }

 void foo(int* data1, int* data2) {
   kernel<<<1, 32>>>(data1, data2);
 }

results in the following migrated SYCL* code:

 void kernel(int* data1, int* data2, const sycl::nd_item<3> &item_ct1) {

   int input = data1[item_ct1.get_local_id(2)];
   int output1 = 0;
   int output2 = 0;
   output1 = sycl::inclusive_scan_over_group(item_ct1.get_sub_group(), input,
                                             sycl::plus<>());
   data1[item_ct1.get_local_id(2)] = output1;
   /*
   DPCT1085:0: The function inclusive_scan_over_group requires sub-group size to
   be 16, while other sub-group functions in the same SYCL kernel require a
   different sub-group size. You may need to adjust the code.
   */
   output2 = sycl::inclusive_scan_over_group(item_ct1.get_sub_group(), input,
                                             sycl::plus<>());
   data2[item_ct1.get_local_id(2)] = output1;
 }

 void foo(int* data1, int* data2) {
   dpct::get_in_order_queue().parallel_for(
       sycl::nd_range<3>(sycl::range<3>(1, 1, 32), sycl::range<3>(1, 1, 32)),
       [=](sycl::nd_item<3> item_ct1) [[intel::reqd_sub_group_size(32)]] {
         kernel(data1, data2, item_ct1);
       });
 }

 void foo(int* data) {
   dpct::get_in_order_queue().parallel_for(
       sycl::nd_range<3>(sycl::range<3>(1, 1, 32), sycl::range<3>(1, 1, 32)),
       [=](sycl::nd_item<3> item_ct1) [[intel::reqd_sub_group_size(32)]] {
         kernel(data, item_ct1);
       });
 }

which is manually adjusted to:

 void kernel(int* data1, int* data2, const sycl::nd_item<3> &item_ct1) {

   int input = data1[item_ct1.get_local_id(2)];
   int output1 = 0;
   int output2 = 0;
   output1 = sycl::inclusive_scan_over_group(item_ct1.get_sub_group(), input,
                                             sycl::plus<>());
   data1[item_ct1.get_local_id(2)] = output1;
   output2 = sycl::inclusive_scan_over_group(item_ct1.get_sub_group(), input,
                                             sycl::plus<>());
   data2[item_ct1.get_local_id(2)] = output1;
   item_ct1.barrier();
   if (item_ct1.get_local_id(2) % 32 >= 16) {
     int warp_id = item_ct1.get_local_id(2) / 32;
     data2[item_ct1.get_local_id(2)] -= data2[warp_id * 32 + 15];
   }
 }

 void foo(int* data1, int* data2) {
   dpct::get_in_order_queue().parallel_for(
       sycl::nd_range<3>(sycl::range<3>(1, 1, 32), sycl::range<3>(1, 1, 32)),
       [=](sycl::nd_item<3> item_ct1) [[intel::reqd_sub_group_size(32)]] {
         kernel(data1, data2, item_ct1);
       });
 }

Suggestions to Fix

Code requires manual fix. Rewrite the code manually.

Select Your Language

Using Intel.com Search

Quick Links

Recent Searches

Advanced Search

Only search in

Intel® DPC++ Compatibility Tool Developer Guide and Reference

DPCT1085

Message

Detailed Help

Suggestions to Fix