feat: use cached max trace heights in leaf/internal aggregation

[stephenh-axiom-xyz]

Resolves INT-4375.

• Hardcodes NativeConfig max trace heights (derived from the kitchen-sink benchmark) and uses them during leaf/internal verification.
• Adds metering to phantom chip, removes metering from public values chip (since estimation is derived from config)

[codspeed-hq]

CodSpeed WallTime Performance Report

Merging #1839 will not alter performance

_{Comparing feat/remove-single-segment-e2 (022cb4a) with feat/new-execution (a7d1125)}

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Summary

✅ 20 untouched benchmarks

[shuklaayush]

looks okay but some tests seem to be failing

[jonathanpwang]

LGTM after fixing the sdk bug.

I would have preferred if we deleted E2 from SingleSegmentExecutor directly, but I am rewriting it in feat/new-exec-device anyways so we can hold off

[shuklaayush]

lgtm

[github-actions]

group	app.proof_time_ms	app.cycles	app.cells_used	leaf.proof_time_ms	leaf.cycles	leaf.cells_used
verify_fibair	(+14 [+1.1%]) 1,261	322,648	17,270,578	-	-	-
fibonacci	(+40 [+1.7%]) 2,374	1,500,277	50,589,231	3,445	1,248,040	70,094,918
regex	(-11 [-0.2%]) 6,926	4,165,432	166,449,586	10,598	3,349,024	230,129,839
ecrecover	(+50 [+3.7%]) 1,392	137,283	8,179,549	10,502	2,934,914	242,451,548
pairing	(-14 [-0.3%]) 4,005	1,862,964	102,530,853	5,269	2,010,483	134,816,207
fib_e2e	17,348	12,000,277	404,296,287	17,128	7,596,379	434,396,860
kitchen_sink	15,317	154,595	903,326,712	23,907	7,991,164	755,626,485

Commit: 022cb4a

Benchmark Workflow