bcui2 commited on
Commit
4c9ee3a
·
verified ·
1 Parent(s): cecd0b0

Serve the input-archive manifest to the Dataset Viewer as Parquet

Browse files

Adds index/input_archives.parquet (a Parquet copy of manifest.csv) and points the input_archives viewer config to it; manifest.csv itself is unchanged.

Files changed (2) hide show
  1. README.md +2 -2
  2. index/input_archives.parquet +3 -0
README.md CHANGED
@@ -25,7 +25,7 @@ configs:
25
  - config_name: golden_outputs
26
  data_files: index/golden_outputs.parquet
27
  - config_name: input_archives
28
- data_files: manifest.csv
29
  ---
30
 
31
  # HPC-Bench
@@ -50,7 +50,7 @@ This dataset holds everything needed to run the benchmark that is not source cod
50
  | Benchmark index | `index/benchmarks.parquet` | 208 rows | One row per (setting, benchmark) |
51
  | Input-size index | `index/inputs.parquet` | 1,040 rows | One row per (setting, benchmark, input size) |
52
  | Reference-output index | `index/golden_outputs.parquet` | 1,023 rows | One row per reference-output file |
53
- | Archive manifests | `manifest.csv`, `baseline_code/manifest.csv` | | Path, size, and SHA-256 of every archive |
54
  | Croissant metadata | `croissant.json` | | Croissant 1.1 with NeurIPS RAI fields |
55
 
56
  Settings: **EX1** serial CPU optimization (`gcc -O3`), **EX2** OpenMP CPU parallelization (`gcc -O3 -fopenmp`, thread count set by `OMP_NUM_THREADS`), **EX3** CUDA GPU optimization (`nvcc`). EX1 and EX2 share the same 92 benchmarks and the same inputs. EX3 contains 24 benchmark directories: the 22 CUDA benchmarks reported in the paper, plus `backprop` and `nw`, which have CUDA sources and inputs but are not part of the paper's CUDA benchmark table (`in_paper_benchmark_set = false`). `nn` generates its input at run time and has no input archive.
 
25
  - config_name: golden_outputs
26
  data_files: index/golden_outputs.parquet
27
  - config_name: input_archives
28
+ data_files: index/input_archives.parquet
29
  ---
30
 
31
  # HPC-Bench
 
50
  | Benchmark index | `index/benchmarks.parquet` | 208 rows | One row per (setting, benchmark) |
51
  | Input-size index | `index/inputs.parquet` | 1,040 rows | One row per (setting, benchmark, input size) |
52
  | Reference-output index | `index/golden_outputs.parquet` | 1,023 rows | One row per reference-output file |
53
+ | Archive manifests | `manifest.csv`, `baseline_code/manifest.csv` | | Path, size, and SHA-256 of every archive (`index/input_archives.parquet` is a Parquet copy of `manifest.csv` for the Dataset Viewer) |
54
  | Croissant metadata | `croissant.json` | | Croissant 1.1 with NeurIPS RAI fields |
55
 
56
  Settings: **EX1** serial CPU optimization (`gcc -O3`), **EX2** OpenMP CPU parallelization (`gcc -O3 -fopenmp`, thread count set by `OMP_NUM_THREADS`), **EX3** CUDA GPU optimization (`nvcc`). EX1 and EX2 share the same 92 benchmarks and the same inputs. EX3 contains 24 benchmark directories: the 22 CUDA benchmarks reported in the paper, plus `backprop` and `nw`, which have CUDA sources and inputs but are not part of the paper's CUDA benchmark table (`in_paper_benchmark_set = false`). `nn` generates its input at run time and has no input archive.
index/input_archives.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:917db0132d83bf761a1051f357bc7ae3b28eacd139504b91ad46a5b99f8451b5
3
+ size 21751