[llvm] MTM: improve operand latency when missing sched info (PR #101389)

[Michael Maitland via llvm-commits]

michaelmaitland wrote:

> In practice, this improvement of one crude estimate by offseting it with another crude estimate leads to better codegen on average, and yields huge gains on standard benchmarks.

Do you have any data to support this claim?

https://github.com/llvm/llvm-project/pull/101389