How can I setup HeavyDB on locally to benchmark
Hi everyone, I have researched a topic about optimizing query performance using GPU and I want to use HeavyDB to benchmark using TPC-DS and TPC-H data sample. However, the facilities at my university don't meet the requirements of HeavyDB. Are there any ways to run HeavyDB locally to benchmark? My laptop specs:
CPU: Core i7 13620H
GPU: NVIDIA® GeForce RTX™
4060 Laptop GPU Discrete 8GB
GDDR6
-
Subject: Installation and Setup Instructions
Hi,
You need to install Ubuntu 22.04 and the NVIDIA drivers on your system. If you are using Windows, you'll need to re-partition your hardware to create some space for the operating system and its data. Once you've done this, please follow the installation instructions in our documentation at https://docs.heavy.ai/installation-and-configuration/installation.
Given the limited memory available (assuming you have 16 GB of system RAM), the scale factor (SF) should not exceed 20 or 30. I can share optimized DDLs and queries for TPC-H, TPC-DS, and SSB benchmarks.
Alternatively, you can request an instance with a free license on a major cloud service provider at https://www.heavy.ai/product/downloads.
Another option is to subscribe to the marketplace with a Bring Your Own License (BYOL) using a free license or explore other available options at https://aws.amazon.com/marketplace/search/results?searchTerms=heavyai.
In any case, I can provide support on setting up the system or conducting the benchmarks.
Let me know if you need any assistance.
Best,
Candido
-
Can u share me the optimized DDLs and queries for TPC-DS, TPC-H. And can you help me know how to benchmark HeavyDB by myself on my local machine. Can I follow this repos: OmniSciDB: Benchmark and Correctness Testing between HeavyDB and PostgreSQL databases
Please sign in to leave a comment.
Comments
2 comments