TPC-H SF1000 (1TB): TiDB Cloud Lake Warehouse Size Benchmark
This page compares Small, Medium, and Large TiDB Cloud Lake Warehouses on the same TPC-H SF1000 workload. SF1000 is commonly used to represent about 1TB of TPC-H data.
Dataset Scale
TPC-H Scale Factor 1000 (SF1000) represents approximately 1TB of generated data. The dataset contains about 6 billion rows across the 8 standard TPC-H tables.
Table
Rows
customer
150,000,000
lineitem
6,000,000,000
nation
25
orders
1,500,000,000
part
200,000,000
partsupp
800,000,000
region
5
supplier
10,000,000
Note
TPC Benchmark™ and TPC-H™ are trademarks of the Transaction Processing Performance Council (TPC). This test is inspired by TPC-H, but it is not an official TPC-H result.
Summary
Warehouse Size
Total Time
Speedup vs. Small
Speedup vs. Previous Size
Small
1173.32s
1.00x
—
Medium
537.93s
2.18x
2.18x
Large
285.96s
4.10x
1.88x
Medium is about 2.18x faster than Small. Large is about 4.10x faster than Small and completes the full 22-query workload in under 5 minutes.
Query Details
Unit: seconds. Lower is better.
Query
Small
Medium
Large
Small → Medium
Medium → Large
Q1
31.61
19.93
10.33
1.59x
1.93x
Q2
10.00
7.15
5.52
1.40x
1.30x
Q3
73.07
24.50
17.75
2.98x
1.38x
Q4
177.60
17.22
16.39
10.31x
1.05x
Q5
300.78
17.69
11.22
17.00x
1.58x
Q6
13.93
4.46
2.19
3.12x
2.04x
Q7
33.08
18.08
9.78
1.83x
1.85x
Q8
31.37
17.81
10.90
1.76x
1.63x
Q9
102.01
45.41
29.52
2.25x
1.54x
Q10
40.84
31.18
23.05
1.31x
1.35x
Q11
5.91
3.59
2.20
1.65x
1.63x
Q12
23.74
11.81
8.70
2.01x
1.36x
Q13
57.78
34.52
23.78
1.67x
1.45x
Q14
38.45
9.84
5.18
3.91x
1.90x
Q15
13.22
8.18
4.82
1.62x
1.70x
Q16
4.77
3.25
2.27
1.47x
1.43x
Q17
20.77
11.29
5.87
1.84x
1.92x
Q18
90.78
158.89
17.24
0.57x
9.22x
Q19
20.36
10.86
8.22
1.87x
1.32x
Q20
25.34
10.07
4.94
2.52x
2.04x
Q21
52.85
64.57
61.19
0.82x
1.06x
Q22
5.06
7.64
4.90
0.66x
1.56x
Total
1173.32
537.93
285.96
2.18x
1.88x
Notes
The overall workload scales clearly as the Warehouse size increases. Some individual queries may not scale linearly due to query shape, execution plan, scheduling, and cache behavior.
