TPCE Benchmark: Broker Table

[le1nux]

Hey guys,

according to the TPCE-Benchmark specifications the broker table is a scaling table. You put it to the growing tables. Did you do this on purpose? https://github.com/apavlo/h-store/blob/master/src/benchmarks/edu/brown/benchmark/tpce/TPCEConstants.java#L319

http://www.tpc.org/tpce/spec/v1.12.0/tpce-v1.12.0.pdf (It's stated on page 70 I guess.)

Best wishes from Germany!

[apavlo]

Our TPC-E implementation is horribly broken and I don't know what the status is for the code in the master branch.

I do know that there is a better, more simplified version somewhere for the S-Store project.

How interested are you in getting this working?

Cook13 [email protected] wrote:

Hey guys,

according to the TPCE-Benchmark specifications the broker table is a scaling table. You put it to the growing tables. Did you do this on purpose? https://github.com/apavlo/h-store/blob/master/src/benchmarks/edu/brown/benchmark/tpce/TPCEConstants.java#L319

http://www.tpc.org/tpce/spec/v1.12.0/tpce-v1.12.0.pdf (It's stated on page 70 I guess.)

Best wishes from Germany!

— Reply to this email directly or view it on GitHub.

[le1nux]

As an undergraduate student I'm currently implementing the TPC-E benchmark for my professor's database benchmark framework. The official documentation does not say much about the initial population of the database and instead suggests to use their EGen tool. That's why I was really glad when I found your project with an EGen implementation in Java and used it for inspiration purposes. At some points I already thought that your implementation might not work exactly as predefined in the specifications. For instance the amount of tuples in the growing tables does not correspond to the amount defined in the specs. I do know that they are approximations but a deviation of up to tenfold seems quite a lot to me.

I searched for the S-Store project and couldn't find it on GitHub. Is it maybe a private repository?

[jmeehan16]

The S-Store TPC-E implementation is also in a rather messy, non-functioning state, though it is most likely closer to completion than the original H-Store version. You are welcome to take a look at it if you would like. The S-Store repository is currently located in the "InsertInto" branch of the H-Store repository (long story, bad name). The benchmark name is "TPCE-B." Hopefully that helps.

John

On Mon, Aug 4, 2014 at 5:49 AM, Cook13 [email protected] wrote:

As an undergraduate student I' currently implementing the TPCE benchmark for my professor's database benchmark framework. The official documentaton does not say much about the initial population of the database and instead suggests to use their EGen tool. That's why I was really glad when I found your project with an EGen implementation in Java and used it for inspiration purposes. At some points I already thought that your implemenation might not work exaclty as predefined in the specifications. For instance the amount of tuples in the scaling tables does not correspond to the amount defined in the specs.

I searched for the S-Store project and couldn't find it on github. Is it maybe a private repository?

— Reply to this email directly or view it on GitHub https://github.com/apavlo/h-store/issues/169#issuecomment-51038605.

[le1nux]

Thanks John! I'll have a look at it.