Testing some singleton implementations

[Jerry]

"Stanislav Blinov" <stanislav.blinov at gmail.com> writes:

> On Wednesday, 5 February 2014 at 00:11:58 UTC, Jerry wrote:
>
>> Here's the best and worst times I get on my linux laptop.  These are
>> with 2.064.2 dmd and gdc 4.9 with 2.064.2
>>
>> On Ubuntu x86_64:
>>
>> ~/dmd2/linux/bin64/dmd -O -release -inline -noboundscheck -unittest
>> singleton.d
>>
>> Test 2 time for SyncSingleton: 753.547 msecs.
>> Test 2 time for AtomicSingleton: 22290.3 msecs.
>>
>> Test 3 time for SyncSingleton: 254.968 msecs.
>> Test 3 time for AtomicSingleton: 22903.3 msecs.
>>
>> Test 6 time for SyncSingleton: 510.118 msecs.
>> Test 6 time for AtomicSingleton: 23970.9 msecs.
>>
>> Test 8 time for SyncSingleton: 480.175 msecs.
>> Test 8 time for AtomicSingleton: 12827.9 msecs.
>
> Whoah, those times for AtomicSingleton are way high. What kind of machine is
> your laptop?

Core 2 Due T9400.  The gdc times were much better for AtomicSingleton -
about 4x slower than SyncSingleton.

> Perhaps we need to repost the test with the latest implementation of
> AtomicSingleton.

I downloaded the test program yesterday.