Dataset Viewer
Auto-converted to Parquet Duplicate
model
stringclasses
1 value
dataset
stringclasses
3 values
num_tables
int64
3
26
tables
listlengths
3
26
num_batches
int64
182
299k
batch_size
int64
128
128
mode
stringclasses
1 value
keys_per_page
int64
16
16
total_accesses
int64
606k
221M
DLRM
criteo_kaggle_large
26
[ { "id": 0, "num_rows": 1104, "embedding_dim": 64 }, { "id": 1, "num_rows": 163, "embedding_dim": 64 }, { "id": 2, "num_rows": 222301, "embedding_dim": 64 }, { "id": 3, "num_rows": 112811, "embedding_dim": 64 }, { "id": 4, "num_rows": 244, "...
182
128
inference
16
605,696
DLRM
avazu
21
[ { "id": 0, "num_rows": 7, "embedding_dim": 64 }, { "id": 1, "num_rows": 7, "embedding_dim": 64 }, { "id": 2, "num_rows": 4737, "embedding_dim": 64 }, { "id": 3, "num_rows": 7745, "embedding_dim": 64 }, { "id": 4, "num_rows": 26, "embedding_...
82,299
128
inference
16
221,219,712
DLRM
avazu
21
[ { "id": 0, "num_rows": 7, "embedding_dim": 64 }, { "id": 1, "num_rows": 7, "embedding_dim": 64 }, { "id": 2, "num_rows": 4737, "embedding_dim": 64 }, { "id": 3, "num_rows": 7745, "embedding_dim": 64 }, { "id": 4, "num_rows": 26, "embedding_...
66,501
128
inference
16
178,754,688
DLRM
avazu
21
[ { "id": 0, "num_rows": 7, "embedding_dim": 64 }, { "id": 1, "num_rows": 7, "embedding_dim": 64 }, { "id": 2, "num_rows": 4737, "embedding_dim": 64 }, { "id": 3, "num_rows": 7745, "embedding_dim": 64 }, { "id": 4, "num_rows": 26, "embedding_...
62,902
128
inference
16
169,080,576
DLRM
taobao
3
[ { "id": 0, "num_rows": 987991, "embedding_dim": 64 }, { "id": 1, "num_rows": 4161138, "embedding_dim": 64 }, { "id": 2, "num_rows": 9437, "embedding_dim": 64 } ]
239,140
128
inference
16
91,829,760
DLRM
taobao
3
[ { "id": 0, "num_rows": 987991, "embedding_dim": 64 }, { "id": 1, "num_rows": 4161138, "embedding_dim": 64 }, { "id": 2, "num_rows": 9437, "embedding_dim": 64 } ]
299,214
128
inference
16
114,898,176

CXL-SSD Archetype Routing — Generated Traces

Stage-00 snapshot (2026-05-02) of processed access traces used in the CXL-SSD page-oriented embedding lookup paper.

These are MQSim/MaxEmbed-format traces derived from public recommendation datasets (MovieLens, Criteo, Avazu, Taobao, Amazon, Yelp, …) and the Qwen KV-Cache traces. They feed into the Cylon (FEMU) and MQSim simulators.

Companion repos

Layout

Each subdirectory corresponds to one source dataset and contains traces of the form <layout>_<ratio>.trace plus the matching MQSim workload.xml / config snippets.

The trace ASCII format (per MQSim convention):

<arrival_time_ns> <device_num=0> <LSN_byte_addr> <size_sectors=8> <type 0=W|1=R>

Page-oriented variant: LSN is the byte address of a 4 KB page; size = 8 sectors (4 KB).

Provenance

Generated from /research_data/raw/<dataset>/ via the pipelines in MERCI_page_aware/analysis/, trace_converter/, and MaxEmbed/scripts/. See MANIFEST.md for per-dataset preprocessing parameters.

License

Trace files derived from each upstream dataset inherit the upstream licence. This collection (the bundle layout + associated metadata) is released CC-BY-NC-4.0 for academic reproducibility.

Do not redistribute Criteo/Avazu/Taobao/Amazon raw fields outside academic context.

Downloads last month
30