docs: expand profiling guide with JVM and async-profiler coverage

[andygrove]

Summary

Add detailed profiling guide - rendered version

Changes

• Renames profiling_native_code.md to profiling.md and expands it into a comprehensive profiling guide
• Adds a tool comparison table (async-profiler vs JFR vs cargo-flamegraph) with guidance on when to use each
• Adds async-profiler section: installation, attaching to running Spark apps via asprof, java agent usage, event types, output formats, platform notes
• Adds Java Flight Recorder section: spark-submit flags, jcmd dynamic recording, viewer options, useful JFR events for Comet debugging
• Adds Tips for Profiling Comet section: wall-clock profiling for JNI overhead, alloc profiling around Arrow FFI, isolating Rust issues, correlating JVM/native frames
• Preserves all existing cargo bench / cargo-flamegraph content
• Links to benchmarks/tpc/README.md for integrated benchmark profiling

[comphead]

Thanks @andygrove we might want to include a full working example

For example I use

TPCH Local Q4 (Profiler)
JAVA_HOME=/opt/homebrew/opt/openjdk@17/libexec/openjdk.jdk/Contents/Home SPARK_HOME=/opt/homebrew/Cellar/apache-spark/3.5.5/libexec spark-submit --master="local[*]" --conf spark.driver.memory=8G --conf spark.executor.instances=1 --conf spark.executor.cores=4 --conf spark.cores.max=4 --conf spark.executor.memory=16g --conf spark.memory.offHeap.enabled=true --conf spark.memory.offHeap.size=16g --conf spark.comet.explainFallback.enabled=true --conf spark.comet.logFallbackReasons.enabled=true --conf spark.eventLog.enabled=true --jars $COMET_JAR --driver-class-path $COMET_JAR --conf spark.driver.extraClassPath=$COMET_JAR --conf spark.executor.extraClassPath=$COMET_JAR --conf spark.plugins=org.apache.spark.CometPlugin --conf spark.shuffle.manager=org.apache.spark.sql.comet.execution.shuffle.CometShuffleManager --conf spark.comet.enabled=true --conf spark.comet.exec.shuffle.enableFastEncoding=true --conf spark.comet.exec.shuffle.fallbackToColumnar=true --conf spark.comet.exec.replaceSortMergeJoin=true --conf spark.comet.cast.allowIncompatible=true --conf spark.comet.scan.impl=native_iceberg_compat --conf "spark.executor.extraJavaOptions=-agentpath:/Users/xxx/Downloads/async-profiler-4.3-macos/lib/libasyncProfiler.dylib=start,event=cpu,file=profile-executor.html,tree" --conf "spark.driver.extraJavaOptions=-agentpath:/Users/xxx/Downloads/async-profiler-4.3-macos/lib/libasyncProfiler.dylib=start,event=cpu,file=profile-driver.html,tree" dev/benchmarks/tpcbench.py --name comet --benchmark tpch --data ../datafusion-benchmarks/tpch/data --queries ../datafusion-benchmarks/tpch/queries --output . --iterations 1 --query 4

And this flag helps a bit file=profile-driver.html,tree which builds an output as a tree hierarchy