[CUDA] Fix issue about integer overflow

[mfbalin]

Fixes #6585.

[dgl-bot]

To trigger regression tests:

• @dgl-bot run [instance-type] [which tests] [compare-with-branch];
  For example: @dgl-bot run g4dn.4xlarge all dmlc/master or @dgl-bot run c5.9xlarge kernel,api dmlc/master

[dgl-bot]

Commit ID: 7e276c3

Build ID: 1

Status: ✅ CI test succeeded.

Report path: link

Full logs path: link

[mfbalin]

@Rhett-Ying could you review this quick fix?

[Rhett-Ying]

could you add a testcase for this issue?

[Rhett-Ying]

could you add a testcase for this issue?

It's probably not easy to add testcase for this issue in python code path as it needs to create a large graph.

[mfbalin]

could you add a testcase for this issue?

It's probably not easy to add testcase for this issue in python code path as it needs to create a large graph.

I agree. It may take too much memory and runtime.

[dgl-bot]

Commit ID: c9a7510d4c86c9a7292b9e161ebd149ee356304a

Build ID: 2

Status: ✅ CI test succeeded.

Report path: link

Full logs path: link

[dgl-bot]

Commit ID: c081f904c45115bf5415ac161376655e657da1f5

Build ID: 3

Status: ✅ CI test succeeded.

Report path: link

Full logs path: link

[Rhett-Ying]

@mfbalin could you help resolve below warning which is the side effect of this PR?

utils.h:49:73: warning: comparison of integer expressions of different signedness: ‘long long unsigned int’ and ‘const long int’ [-Wsign-compare]
   49 |   while (bits < static_cast<int>(sizeof(T) * 8) && (1ull << bits) < range) {