5 ms·

OpenBenchmarking.org – A Platform for Benchmarking and Performance Analysis

56 points by rayascott 11y ago

LogicX 11y ago
It just doesn't do it for me, for my general benchmarking purposes. You really have to dig down to do comparisons, and they felt unfulfilling. Serverbear.com, despite having not been updated for years, still feels like the best UI, and best quick snapshot for performance, and makes it easy to compare performance. It's not perfect, by a long shot, but it's the closest to meeting my needs until I'm presented with something better.
- rosehill 11y ago
  How does result set compare?
- jepler 11y ago
  ditto. I would love an easy answer to the question: what CPU has the best performance as a compile-server. That answer has to be somewhere inside openbenchmarking.org, but I'll be darned if I can get it to tell me.
kmfrk 11y ago
Never understood why Valve didn't leverage Steam to build a database of performance benchmarks, especially to change the awful System Requirements system.
- davidgaleano 11y ago
  Well, at least they made their hardware and software survey public: http://store.steampowered.com/hwsurvey http://store.steampowered.com/hwsurvey
- justinclift 11y ago
  What do you reckon the right way for them do that that would be (in a practical sense)? Thinking about that a bit, it would probably be possible, but there would be downsides that might not make it worth it. For example, supporting people on the low side of hardware specs complaining that X game/software/etc isn't working well enough. Conversely, other people on the same hardware might not have an issue with it. (different levels of "acceptable" for different people) Not sure it'd be worth the effort (any time soon). There are other much simpler things Valve needs to fix first in order to increase sales. :)
- slavik81 11y ago
  It might be a good idea, but they've already got a lot on their plate for such a small company.
rixoff 11y ago
This site is a pile of confusion. Of all the benchmarks I'm interested, the HPC Challenge, the results are buried. No overview, and when you finally get into the results, the performances are all terrible. (Should be closer to 500GFLOPS per node, not 100!) This also appears geared towards consumer use, which is akin to adding throttle body spacers and tall spoilers to your car... not practical applications.
slavik81 11y ago
I like this. It could certainly use some focus on getting me from a question I need answered to the data I really care about. Still, the site looks nice, runs well, and seems to have a lot of useful information.

56 points by rayascott 11y ago

It just doesn't do it for me, for my general benchmarking purposes. You really have to dig down to do comparisons, and they felt unfulfilling. Serverbear.com, despite having not been updated for years, still feels like the best UI, and best quick snapshot for performance, and makes it easy to compare performance. It's not perfect, by a long shot, but it's the closest to meeting my needs until I'm presented with something better.

rosehill 11y ago
How does result set compare?
jepler 11y ago
ditto. I would love an easy answer to the question: what CPU has the best performance as a compile-server. That answer has to be somewhere inside openbenchmarking.org, but I'll be darned if I can get it to tell me.

kmfrk 11y ago

Never understood why Valve didn't leverage Steam to build a database of performance benchmarks, especially to change the awful System Requirements system.

davidgaleano 11y ago
Well, at least they made their hardware and software survey public: http://store.steampowered.com/hwsurvey http://store.steampowered.com/hwsurvey
justinclift 11y ago
What do you reckon the right way for them do that that would be (in a practical sense)? Thinking about that a bit, it would probably be possible, but there would be downsides that might not make it worth it. For example, supporting people on the low side of hardware specs complaining that X game/software/etc isn't working well enough. Conversely, other people on the same hardware might not have an issue with it. (different levels of "acceptable" for different people) Not sure it'd be worth the effort (any time soon). There are other much simpler things Valve needs to fix first in order to increase sales. :)
slavik81 11y ago
It might be a good idea, but they've already got a lot on their plate for such a small company.

rixoff 11y ago

This site is a pile of confusion. Of all the benchmarks I'm interested, the HPC Challenge, the results are buried. No overview, and when you finally get into the results, the performances are all terrible. (Should be closer to 500GFLOPS per node, not 100!) This also appears geared towards consumer use, which is akin to adding throttle body spacers and tall spoilers to your car... not practical applications.

slavik81 11y ago

I like this. It could certainly use some focus on getting me from a question I need answered to the data I really care about. Still, the site looks nice, runs well, and seems to have a lot of useful information.