.gitignore

book/
book-before/
book-after/
clustermax/
clustermax_from_node2
clustermax*/
cluster/Stas*
third_party/TransformerEngine-src/
third_party/wheels/
third_party/cutlass/
artifacts_tmp/
queue_runner.pid
.cuda_env/
file_inventory.csv
third_party/wheels/torch-*.whl
third_party/wheels/transformer_engine-*.whl
third_party/wheels/flash_attn*.whl
third_party/wheels/vllm*.whl
third_party/vllm-src/
third_party/te_sm100a_env.sh
.cursor/.plan-and-implement/
.plan-and-implement/
.cursor/plans/
.env
.codex
.cache/
.env.local
.env.bak
.venv/
.venv*/
.tmp_pdfqa/
.tmp_pdf*/
report.html
mcp_tools_test_report.md
core/artifacts/ 
.hypothesis/
artifacts_chXX/
benchmark_profiles_chXX/
config/.env
config/.env.local
benchmark_test_results.json
benchmark_test_results.md
benchmark_events.jsonl
benchmark_events_*.jsonl
benchmark_status_table.json
benchmark_status_table.txt
benchmark_results/
checkpoints/
benchmark_test_results.md
.torch_inductor/
review_issues.json
tmp_bench/
results_all/
results_ch1/
results_ch20/
synthetic_moe_results.json
accuracy_test_results_*/
tmp_bench_ch20/
tmp_results/
tmp_profiles/
tmp_profile_*/
tmp_bench_*/
tmp_*/
gpt-oss-20b/
mixtral-8x7b/
phi-3.5-moe/
.kiro/
*-plan.md
benchmark_peak_results_*.json
benchmark_peak_manifest_*.json
benchmark_profiles/
benchmark_results_*.json
benchmark_results_*.json
benchmark_test_results_*.json
benchmark_test_results_*.json
DUPLICATE_ANALYSIS.md
HIGH_PRIORITY_DUPLICATES.md
DUPLICATE_ANALYSIS_UPDATED.md
proof_of_benefit.csv
reports/
tmp*.py
artifacts_tmp*.json
artifacts_tmp*.pid
comparison_*.html
profile_captures/

core/benchmark/blackwell_optimizations/blackwell_optimizations/test_all_features
core/benchmark/blackwell_optimizations/blackwell_optimizations/test_clusters
core/benchmark/blackwell_optimizations/blackwell_optimizations/test_dsmem
core/benchmark/blackwell_optimizations/blackwell_optimizations/test_fp8
core/benchmark/blackwell_optimizations/blackwell_optimizations/test_tma
core/benchmark/blackwell_optimizations/blackwell_optimizations/test_tmem
core/benchmark/blackwell_optimizations/blackwell_optimizations/test_warp_spec
ch19/baseline_fp4_hardware_kernel
ch19/optimized_fp4_hardware_kernel

# Python
__pycache__/
*.pyc
*.pyo
*.pyd
.Python
*.egg
*.egg-info/
dist/
build/
.pytest_cache/
.ipynb_checkpoints/
code/cluster_perf_suite/

# CUDA compiled binaries
*_sm100
*_sm121
*.o
*.so
*.a
*.cubin
*.fatbin
*.ptx

# Profiling outputs
profiles/
profiles_*/
profiling_results/
profiling_results_*/
profiling_results_new/
ch*/profiling_results/
code/profiles/
example_profiling/
profile_runs/
ch*/*_pytorch_profile/
*.nsys-rep
*.ncu-rep
*.sqlite
*.qdrep
nsys_profile_*
ncu_profile_*
hta_profile_*
perf.data
perf.data_*
comprehensive_*
*.pt.trace.json
ch*/hta_traces/
# Allow curated profiling captures for review/commit
!profile_captures/
!profile_captures/**
!profile_captures/**/*.nsys-rep
!profile_captures/**/*.ncu-rep
!profile_captures/**/*.qdrep
!profile_captures/**/*.sqlite

# Test and benchmark results
benchmark_results/
*/benchmark_results/
test_results/
test_results_*/
quick_test_results/
quick_test_results_*.tar.gz
all_chapters_results/
all_chapters_results_*/
validation_results/
validation_results_*/
quick_load_test_validation/

# Root-level benchmark and performance result files
all_results.json
all_results_output.txt
performance_results.json
performance_summary.txt
final_benchmark_results.txt
final_benchmark_run.log
BENCHMARK_PEAK_RESULTS_*.json
single_gpu_load_test.json

# Generated summary files
*summary.md
**/summary.md
**/BENCHMARK_PAIR_REVIEW_REPORT.md

# Expectations backups
*.json.bak

# LLM-generated analysis outputs
llm_analysis/
**/llm_analysis/
llm_patches/
**/llm_patches/
**/llm_analysis/
llm_patches_test/
**/llm_patches_test/
llm_explanations/
**/llm_explanations/

# Power and performance metrics
power_results/
power_metrics/
quick_power_test/
power_baseline*.json
nvlink_bandwidth_test.json
large_model_eager_8xb200*.json

# PyTorch compile caches
.torch_inductor/
.torch_extensions/
compiled_code_output/
torch_compile_debug/

# Build and run artifacts
artifacts/*
!artifacts/golden/
artifacts/20*/
ch*/*_build.log
ch*/*_profile.log
ch*/*_benchmark.log
ch*/*_output.log
*_build.log
*_profile.log
*_benchmark.log

# Chapter-specific generated files
ch*/*_trace.json
ch*/*_snapshot.json
ch*/synthetic_moe_results.json

# Log files
*.log
cufile.log
quick_retest.log
benchmark_run.log
final_benchmark.log

# Generated plots and outputs (in chapter directories)
ch*/roofline_plot.png
ch*/*.png
ch*/*.svg

# Chapter NVCC build products (extensionless executables)
# Match extensionless executables but exclude source files and config files
ch*/*_sm[0-9]*
ch*/*_sm[0-9][0-9]*
# Test executables (manually compiled test programs)
ch*/test_*
# Build targets without architecture suffix (base names)
ch*/warp_spec_cluster_test
ch*/warp_specialized_cluster_pipeline_multistream
ch*/warp_specialized_two_pipelines_multistream
# Explicitly allow Python and CUDA source files
!ch*/*.py
!ch*/*.cu
!ch*/Makefile

# Third-party downloads
third_party/cutlass/
third_party/llvm/
!ch*/README.md
!ch*/arch_config.py
!ch*/compare.py

# NFS files
.nfs*

# Bootcamp game data (player-specific)
bootcamp/data/game_state.json
bootcamp/data/*.json
!bootcamp/data/.gitkeep

# Generated evaluation datasets (keep README, ignore data files)
eval_datasets/*.txt
eval_datasets/*.json
!eval_datasets/README.md

# Editor and IDE
.vscode/
.idea/
*.swp
*.swo
*~
.DS_Store

# Node / frontend
node_modules/
dashboard/web/.next/
dashboard/web/.turbo/