The New Linux Hardware, Software Playground

[phoronix]

Phoronix: The New Linux Hardware, Software Playground

Here's some new features and other upcoming work that may excite Linux enthusiasts in the Phoronix software/hardware playground...

http://www.phoronix.com/vr.php?view=MTAwNTM

[Laughing1]

Nice job, I'm impressed.

[Kivada]

Phoronix: The New Linux Hardware, Software Playground

Here's some new features and other upcoming work that may excite Linux enthusiasts in the Phoronix software/hardware playground...

http://www.phoronix.com/vr.php?view=MTAwNTM

Will one of the improvements be implementing whatever methods used by CPU-Z and GPU-Z to detect the hardware more accurately as is done by every Windows review site?

[zeroepoch]

Missing Features

Maybe I'm not the expected audience, but I find fundamental things missing or at least I can't find them. The most obvious feature to me is being able to do comparisons between results. If some guy with an i7 990x ran the apache benchmark and another guy ran the same benchmark on his atom machine why can't I just give the website two IDs or result names and get the results compared. All the data is there but there is no way to mine the data. If I can't run a simplified query a filter UI would be helpful instead. Maybe this is intentional so we rely on you to think of everything we'd want to search for. Think of newegg and searching for products and comparing them. Seems to fit right into your ideas as well.

Also why did I install python-fpdf then there is no option from the command line to save the results as PDF? Also uploading results just fails with some generic error message like "upload failed".

You have some great software and a great website but then the polish or user friendliness seems to be missing. Last time I checked the bug where you need to login to the forum first and return to the home page each time to get phoronix premium working is still there.

[Michael]

Will one of the improvements be implementing whatever methods used by CPU-Z and GPU-Z to detect the hardware more accurately as is done by every Windows review site?

PTS does support some of them for better RAM detection, etc. It's done automatically, but it requires running PTS as root. Not many of the Linux users normally run their applications as root.

[Michael]

Maybe I'm not the expected audience, but I find fundamental things missing or at least I can't find them. The most obvious feature to me is being able to do comparisons between results. If some guy with an i7 990x ran the apache benchmark and another guy ran the same benchmark on his atom machine why can't I just give the website two IDs or result names and get the results compared. All the data is there but there is no way to mine the data. If I can't run a simplified query a filter UI would be helpful instead. Maybe this is intentional so we rely on you to think of everything we'd want to search for. Think of newegg and searching for products and comparing them. Seems to fit right into your ideas as well.

It works by either inputting the same OpenBenchmarking.org ID when anyone else wants to compare their system to yours. Or if on the left column of OpenBenchmarking.org it should detect all 'matching' results in the Compare Results menu and there's also a search item there. Or the manual way is to just append the extra OpenBenchmarking.org IDs your talking about delimited by a comma to the URL string of /results/

Also why did I install python-fpdf then there is no option from the command line to save the results as PDF?

It should say it right in the documentation... But it's the result-file-to-pdf option generates the PDF.

You have some great software and a great website but then the polish or user friendliness seems to be missing. Last time I checked the bug where you need to login to the forum first and return to the home page each time to get phoronix premium working is still there.

Yes there's always room to make it more user-friendly. Unfortunately as I am basically the only one writing all of the content on Phoronix, and all of the code behind the Phoronix Test Suite and Openbenchmarking.org, among other work, I don't have much time for polishing.

[madjr]

would be cool if we could see some sort of integration or data exchange with the ubuntu friendly programme in the future.

[Michael]

would be cool if we could see some sort of integration or data exchange with the ubuntu friendly programme in the future.

They don't seem to be interested in collaboration. I've offered multiple times with Canonical/Ubuntu to collaborate on various test related matters, and they just tend to seem to be not interested and/or NIH.

[curaga]

That's some mighty good plans. I hope there'll be some caching, so not every page load at openbenchmarking causes a hit to various bugzillas, MLs, and git repos?

[Michael]

I hope there'll be some caching, so not every page load at openbenchmarking causes a hit to various bugzillas, MLs, and git repos?

Of course. Otherwise it would be much too slow for users and too much server load. There's already various caching mechanisms in place seeing as already the internally collected OB result data is in excess of a GB that's routinely scanned.