Improve performance of sdk.Dec ImmutOpInt64 by avoiding newint

[sunnya97]

Based on @ValarDragon TODO from #11004:

We should use the already allocated operand bigint to avoid newint each time. Need to add a mutex for race conditions.

---

For Admin Use

• Not duplicate issue
• Appropriate labels applied
• Appropriate contributors tagged
• Contributor assigned/self-assigned

[elias-orijtech]

Shared state protected by a mutex is a heavy weight optimization. Is this issue only based on a TODO or is there evidence Dec.ImmutOpInt64 is performance sensitive in realistic workloads?

[ValarDragon]

Unclear if the mutex would be faster, but in Osmosis we do a lot of ~all decimal math operations.

E.g. this one is used in exponentiation math, which we do a lot of

[elias-orijtech]

Ok, it sounds like we should wait for a benchmark that clearly shows the performance hit solved by optimizing ImmutOpInt64, correct? If so, I'm inclined to close this as won't fix and leave the TODO in code. Let me know if closing this kind of issue is inappropriate.

[ValarDragon]

This sounds good to me