5
\$\begingroup\$

For questions, TIO has been a superb resource. However, for challenges there is no equivalent tool available.

I believe we could explore working with one or more 3rd parties to build such a tool; it would be in the clear interests of performance analysers, cloud providers, Microsoft, chipmakers etc to build something and have visibility with a fairly top flight group of developers.

A previous question on whether TIO was adequate for performance challenges was fairly conclusive that it wasn't, and that there wasn't an easy alternative. Having a communal tool would also resolve the gnarly problems of performance competitions.

In terms of function, I would imagine at least initially a lang/VM-specific challenge format, e.g. single-threaded Python, parallel C++/native via xyz libraries, or JVM; it may be sensible for different 3rd parties to provide different environments matching their commercial focus. As a minimum, a company could in theory provide for a specific scenario as a microservice to TIO if that was acceptable to everyone.

If there is support here, I am happy to make contact with essentially anyone on the perf analysis radar in that light. If they could work with (maybe sponsor?) TIO as well, that might help pay back a debt to Dennis and simultaneously make the tool relevant, given how perfect the TIO fit is for PPCG questions.

\$\endgroup\$
5
\$\begingroup\$

I don't think so

I disagree primarily with this:

it would be in the clear interests of performance analysers, cloud providers, Microsoft, chipmakers etc to build something and have visibility with a fairly top flight group of developers.

For parties in the business of performance, they already have highly-skilled developers that know how to write and test for performance. Furthermore, these parties likely have specific features or parts they are looking to test. There's no guarantee that we would test those parts, so why would they fund this?

\$\endgroup\$
  • \$\begingroup\$ Not as a method of testing anything, purely for marketing purposes, in the same way that MS provide developer resources anyway. Dyalog APL sponsor TIO, that kind of thing. If there were a community of people competing on performance grounds in the same way that we compete in code golf, would those not be exactly the right audience? The perf tool providers have the right infrastructure and knowledge to support a tool that tests the performance of code snippets, so it would seem an area of great overlap and a straightforward marketing opportunity for them. \$\endgroup\$ – Phil H May 3 at 12:25
  • \$\begingroup\$ True, they could offer it for free to market it to developers in the perf business. However, the chances of an outsider telling a company how to do their marketing sounds far fetched. We are consumers. This means that the best way that we can speak to businesses is by spending money. Unless that's on the table, I don't really see how much influence we really have. \$\endgroup\$ – Nathan Merrill May 3 at 13:22
  • \$\begingroup\$ Many companies spend a significant amount of money on gathering opinions from existing and prospective customers. It also costs us a negligible amount to make a suggestion \$\endgroup\$ – trichoplax May 9 at 6:28
1
\$\begingroup\$

We are a fairly small community for a global company to target, and those of us who compete in speed optimisation challenges are an even smaller group. I don't see any reason not to ask, but I'd be surprised if there was interest.

Would there be interest from the community in building such a tool ourselves? Would this inevitably cost a lot to run? Are there ways around this? Is it worth drawing up a set of minimal requirements for such a tool?

\$\endgroup\$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .