CTS for subgroupAdd and subgroupMul

[alan-baker]

Issue: #

---

Requirements for PR author:

• [ ] All missing test coverage is tracked with "TODO" or .unimplemented().
• [ ] New helpers are /** documented */ and new helper files are found in helper_index.txt.
• [ ] Test behaves as expected in a WebGPU implementation. (If not passing, explain above.)
• [ ] Test have be tested with compatibility mode validation enabled and behave as expected. (If not passing, explain above.)

Requirements for reviewer sign-off:

• [ ] Tests are properly located in the test tree.
• [ ] Test descriptions allow a reader to "read only the test plans and evaluate coverage completeness", and accurately reflect the test code.
• [ ] Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
• [ ] Helpers and types promote readability and maintainability.

When landing this PR, be sure to make any necessary issue status updates.

[alan-baker]

I checked the the Add tests. The coverage looks good. It's nearly there.

The accuracy result should say why we're only spot checking two positions: 1. to ensure we don't overflow and hit indeterminate value, and 2. to reuse the accuracy checking machinery.

(On 1. actually, can't we hit an overflow value by accident? E.g. max value + max value?)

By reusing the intervals do we not properly handle the overflow?

[dneto0]

I checked the the Add tests. The coverage looks good. It's nearly there. The accuracy result should say why we're only spot checking two positions: 1. to ensure we don't overflow and hit indeterminate value, and 2. to reuse the accuracy checking machinery. (On 1. actually, can't we hit an overflow value by accident? E.g. max value + max value?)

By reusing the intervals do we not properly handle the overflow?

Yes we do, but by only doing two multiplies (or adds) it's much less likely that we'll hit infinity. It's about making a high fraction of the cases provide meaningful things to check. Sorry should have been more clear.