Use correct assignedInterval for SPILL_COST calculation

[kunalspathak]

In #53853, I changed the SPILL_COST heuristic to also consider reload weight if the variable will be reloaded in future location rather than getting spilled at past location. However, while doing that, I incorrectly checking the wrong interval for isLocalVar. I wanted to use similar condition as used in

runtime/src/coreclr/jit/lsra.cpp

Line 3267 in d4b98b9

if (fromRefPosition->RegOptional() && !(interval->isLocalVar && fromRefPosition->IsActualRef()))

However, the interval I was using was that of the current refposition for which we are allocating the register rather than the interval to which the candidate register is already assigned.

Here is the sample diff of Algorithms.VectorDoubleRenderer:RenderSingleThreadedWithADT(float,float,float,float,float) that we can see: The high weight local variable gets assigned to the register.

With better selection of register, we can now eliminate unnecessary spilling/resolution of expensive variables:

Some of the other regressions seen will be fixed by #54345.

Fixes: #54352

[kunalspathak]

Code Size

Scenario	OS/arch	main	PR	Diff	Diff %
Benchmarks	Linux.x64	106830	106996	166	0.16%
Benchmarks	windows.x64	79433	79432	-1	0.00%
Benchmarks	windows.x86	213690	214110	420	0.20%
Libraries.crossgen2	Linux.x64	56998	57020	22	0.04%
Libraries.crossgen2	windows.x64	6972	6865	-107	-1.53%
Libraries.crossgen2	windows.x86	240319	240312	-7	0.00%
Libraries.pmi	Linux.x64	82219	82134	-85	-0.10%
Libraries.pmi	windows.x64	73971	74381	410	0.55%
Libraries.pmi	windows.x86	410698	411263	565	0.14%

Perf Score

OS/arch	main	PR	Diff	Diff %
Linux.x64	102397	100669.68	-1727.32	-1.69%
windows.x64	168025.98	164684.59	-3341.39	-1.99%
windows.x86	10261610.27	10258036.27	-3574	-0.03%
Linux.x64	107910.52	106060.97	-1849.55	-1.71%
windows.x64	26584.66	26107	-477.66	-1.80%
windows.x86	535206.26	535575.34	369.08	0.07%
Linux.x64	708552.68	706913.69	-1638.99	-0.23%
windows.x64	128077.33	126861.96	-1215.37	-0.95%
windows.x86	41500664980	40594662922	-906002058.5	-2.18%

Detail diffs are: https://gist.github.com/kunalspathak/49940d4690ef99f2e24b2ee8bfbac9e2

[kunalspathak]

@dotnet/jit-contrib

[kunalspathak]

Failures seems to be infra issues.