Improve (and test) memoryview reference counting

[da-woods]

1. Avoid needing an indirection to align the atomic types. They should be aligned anyway, and this causes a notable slow-down.
2. Don't use "volatile" typedef - volatile doesn't guarantee anything about atomicity and isn't needed here.
3. Use explicit memory ordering. Won't make much difference on x86/x64 but may be faster on other architectures.
4. Add a stress-test to try to make sure we haven't messed up.

Closes #5509

[scoder]

Some internal warnings regarding utility code were moved:

[2] compiling (c/cy2) pure_warnings ... 
=== Expected: ===
12:10: Unknown type declaration 'Bar' in annotation, ignoring
15:16: Unknown type declaration 'stdint.bar' in annotation, ignoring
18:17: Unknown type declaration in annotation, ignoring
19:15: Unknown type declaration in annotation, ignoring
20:17: Unknown type declaration in annotation, ignoring
25:10: 'cpdef_method' redeclared
33:14: Unknown type declaration 'Bar' in annotation, ignoring
36:10: 'cpdef_cname_method' redeclared
37:20: Unknown type declaration 'stdint.bar' in annotation, ignoring
980:29: Ambiguous exception value, same as default return value: 0
1021:46: Ambiguous exception value, same as default return value: 0
1111:29: Ambiguous exception value, same as default return value: 0
=== Got: ===
12:10: Unknown type declaration 'Bar' in annotation, ignoring
15:16: Unknown type declaration 'stdint.bar' in annotation, ignoring
18:17: Unknown type declaration in annotation, ignoring
19:15: Unknown type declaration in annotation, ignoring
20:17: Unknown type declaration in annotation, ignoring
25:10: 'cpdef_method' redeclared
33:14: Unknown type declaration 'Bar' in annotation, ignoring
36:10: 'cpdef_cname_method' redeclared
37:20: Unknown type declaration 'stdint.bar' in annotation, ignoring
963:29: Ambiguous exception value, same as default return value: 0
1004:46: Ambiguous exception value, same as default return value: 0
1094:29: Ambiguous exception value, same as default return value: 0
FAIL

[4] compiling (c/cy2) fused_redeclare_T3111 ... ok
=== Expected: ===
25:10: 'cpdef_method' redeclared
36:10: 'cpdef_cname_method' redeclared
980:29: Ambiguous exception value, same as default return value: 0
980:29: Ambiguous exception value, same as default return value: 0
1021:46: Ambiguous exception value, same as default return value: 0
1021:46: Ambiguous exception value, same as default return value: 0
1111:29: Ambiguous exception value, same as default return value: 0
1111:29: Ambiguous exception value, same as default return value: 0
=== Got: ===
25:10: 'cpdef_method' redeclared
36:10: 'cpdef_cname_method' redeclared
963:29: Ambiguous exception value, same as default return value: 0
963:29: Ambiguous exception value, same as default return value: 0
1004:46: Ambiguous exception value, same as default return value: 0
1004:46: Ambiguous exception value, same as default return value: 0
1094:29: Ambiguous exception value, same as default return value: 0
1094:29: Ambiguous exception value, same as default return value: 0
FAIL

[scoder]

Given the speed improvement that you report, I consider this a candidate for 3.0rc1. What do you think?

[da-woods]

Given the speed improvement that you report, I consider this a candidate for 3.0rc1. What do you think?

The speed improvement is ~20% on a very artificial test-case (I think it's the kind of code that people tend not to write in reality, because it's a bit slow). So it's decent, but won't really change anyone's calculations about what to write. #5508 will give a much more dramatic speed increase for the same case (and complements this rather than replaces it).

I think this PR is fairly low risk (to the extent that anyone really understands atomics), while the other PR needs a more careful look and a bit of polishing. So I'm personally inclined to merge this one.

If we do see issues with this, it'll be on non-x86 platforms, which I don't think we currently test.