Simplify bump allocator for single-threaded PVM

[pgherveou]

Summary

• Replace AtomicUsize + CAS loop with Cell<usize> in BumpAllocator — PVM contracts are single-threaded so atomics are unnecessary overhead
• Panic on OOM instead of returning null pointer to surface allocation bugs early
• Add alloc_zeroed shortcut that delegates to alloc since the heap is always zero-initialized
• Add regenerate-reports.sh script for automated benchmark report generation
• Regenerate encoding benchmark report with fresh numbers

Test plan

• cargo test -p pvm-bump-allocator — unit tests updated to expect panics on OOM
• cargo clippy --workspace --all-targets --all-features
• CI passes

[github-actions]

Benchmark Size Comparison (vs main)

Artifact	Baseline	Current	Delta	Status
fibonacci_builder-dsl.debug.polkavm	36592	36592	+0.00%	OK
fibonacci_builder-dsl.release.polkavm	1194	1194	+0.00%	OK
fibonacci_no-alloc.debug.polkavm	35834	35834	+0.00%	OK
fibonacci_no-alloc.release.polkavm	469	469	+0.00%	OK
fibonacci_with-alloc.debug.polkavm	47358	45094	-4.78%	OK
fibonacci_with-alloc.release.polkavm	584	448	-23.29%	OK
multi_builder-dsl.debug.polkavm	53691	53691	+0.00%	OK
multi_builder-dsl.release.polkavm	3568	3568	+0.00%	OK
multi_no-alloc.debug.polkavm	55550	55550	+0.00%	OK
multi_no-alloc.release.polkavm	3419	3419	+0.00%	OK
multi_with-alloc.debug.polkavm	66896	64626	-3.39%	OK
multi_with-alloc.release.polkavm	3700	3864	+4.43%	OK
mytoken_builder-dsl.debug.polkavm	49965	49965	+0.00%	OK
mytoken_builder-dsl.release.polkavm	4097	4097	+0.00%	OK
mytoken_no-alloc.debug.polkavm	51973	51973	+0.00%	OK
mytoken_no-alloc.release.polkavm	3751	3751	+0.00%	OK
mytoken_with-alloc.debug.polkavm	65115	62849	-3.48%	OK
mytoken_with-alloc.release.polkavm	4378	4301	-1.76%	OK
----------------------------------------------------	------------	------------	----------	--------
Total	488134	481285	-1.40%

OK: All artifacts within 5% regression threshold

_{Run 22391459972 | 580082a | 2026-02-25 10:15 UTC}