Benchmarks

The bench/ directory hosts example programs, micro-benchmarks, and development utilities built on top of ORQ.

Sub-directories:

queries/ – End-to-end analytical queries and demonstrations.
- tpch/ - TPC-H queries
- secretflow/ - comparisons with SecretFlow
- other/ - all other queries from competitors
micro/ – Small micro-benchmarks targeting specific primitives.

For more information on the queries, see the C++ source files. Below, documentation for the data-generation classes is provided.

TPCH Database Generator.

Reference: Ch. 1 of the TPC-H Specification, https://www.tpc.org/TPC_Documents_Current_Versions/pdf/TPC-H_v3.0.1.pdf

Naming semantics:

Table names are all capitalized
Column names should use PascalCase
Column names should not use table prefixes, unless required for disambiguation in a query. In that case, a single-character table identifer can be prepended, with an underscore (e.g., O_Comment)

Defines

CUSTOMERS_MULTIPLIER

ORDERS_MULTIPLIER

PART_MULTIPLIER

SUPPLIER_MULTIPLIER

PARTSUPP_MULTIPLIER

NATION_SIZE

REGION_SIZE

PARTSUPP_PER_PART

CUSTOMER_ORDER_FRACTION

COMMENT_BITS

TPCH_DATE_THRESHOLD

TPCH_DATE_INTERVAL

TPCH_DISCOUNT_THRESHOLD

TPCH_QUANTITY_THRESHOLD

template<typename T = int> class TPCDatabase

#include <tpch_dbgen.h>

Public Functions

inline TPCDatabase(double sf)

inline TPCDatabase(double sf, sqlite3 *sqlite_db)

inline size_t customersSize()

inline size_t ordersSize()

inline size_t lineitemsSize()

inline size_t partSize()

inline size_t partSuppSize()

inline size_t supplierSize()

inline size_t nationSize()

inline size_t regionSize()

inline EncodedTable<T> getCustomersTable()

inline EncodedTable<T> getLineitemTable(bool proactive_sharing = false)

Get the Lineitem Table object.

Parameters:: proactive_sharing – whether to use proactive sharing: also return in the table secret-shared values with constant thresholds subtracted off already.
Returns:: EncodedTable<T>

inline EncodedTable<T> getOrdersTable()

inline EncodedTable<T> getRegionTable()

inline EncodedTable<T> getNationTable()

inline EncodedTable<T> getPartTable()

inline EncodedTable<T> getSupplierTable()

inline EncodedTable<T> getPartSuppTable()

inline EncodedTable<T> getLineitemTableSecretFlow(int num_rows)

Get the Lineitem Table object with the SecretFlow SCQL data distribution.

Returns:: EncodedTable<T>

inline EncodedTable<T> getPartTableSecretFlow(int num_rows)

inline EncodedTable<T> getOrdersTableSecretFlow(int num_rows)

Public Members

const double scaleFactor

sqlite3 *sqlite_db

size_t lineItemsSize = 0

Protected Functions

inline Vector<T> randomColumn(size_t size, T min_val, T max_val)

Generate a random vector of type T (possibly signed) from [min_val, max_val).

Parameters:

size –
min_val – the minimum value (inclusive)
max_val – the maximum value (exclusive)

Returns:

Vector<T>

inline Vector<T> counterColumn(size_t size)

Generate a counter column, suitable for primary keys, etc. Will count from 0 to size - 1.

Parameters:: size –
Returns:: Vector<T>

Defines

SECRECY_R_SIZE

SECRECY_S_SIZE

SECRECY_T1_SIZE

SECRECY_T2_SIZE

SECRECY_T3_SIZE

DIAGNOSIS_SIZE

MEDICATION_SIZE

STANDARD_SIZE

template<typename T = int> class SecrecyDatabase : public TPCDatabase<int>

#include <secrecy_dbgen.h>

Public Functions

inline SecrecyDatabase(double sf)

inline SecrecyDatabase(double sf, sqlite3 *sqlite_db)

inline size_t secrecyRSize()

inline size_t secrecySSize()

inline size_t secrecyT1Size()

inline size_t secrecyT2Size()

inline size_t secrecyT3Size()

inline EncodedTable<T> getSecrecyRTable()

inline EncodedTable<T> getSecrecySTable()

inline EncodedTable<T> getDiagnosisTable()

inline EncodedTable<T> getMedicationTable()

inline EncodedTable<T> getPasswordTable()

inline EncodedTable<T> getTaxiTable()

inline EncodedTable<T> getSecrecyT1Table()

inline EncodedTable<T> getSecrecyT2Table()

inline EncodedTable<T> getSecrecyT3Table()

inline EncodedTable<T> getCreditScoreTable()

Public Members

const size_t MAX_TIME = 10000

const int N_DISEASE = 100

const int N_DISEASE_CLASS = 16

const int MAX_COST = 10000