Bridgman Is No Longer "The AMD Open-Source Guy"

[twriter]

Hi Tim,

welcome on board!

Is there anything you want to share with us concerning your vision on the AMD open source strategy?

Earlier in the thread, John described some of the behind the scenes work of the last year. My initial goal is to get that released and to continue with our efforts to have launch-time open source support.

As for a longer term vision, I think our Executive Team has outlined a strategy that requires more and better Linux and open source support. Technologies like HSA and markets like cloud and embedded can all benefit from open source so our work supports that vision.

On a personal note, I'm a Linux head. I've been running and developing on Linux since 1993. Prior to joining AMD, I worked as an independent Linux consultant in the mid '90s and, in 1998, co-founded a well known Linux consulting company based in Toronto. For a time, I was active in TLUG (Toronto Linux Users Group) and CLUE (Canadian Linux Users Exchange). The company I co-founded was instrumental in the creation of LPI and I was one of the first, perhaps the first, LPIC-2's in Canada. I'm posting this from Firefox running on Ubuntu 12.04 (with XFCE), I have an Android phone, and I avoid running Windows because it just gets in my way. All that is to say that I consider myself a strong Linux and open source advocate with a good understanding of why businesses and individuals choose Linux and open source. While I'm not in position to define AMD's strategy, I do what I can to ensure that Linux and open source are given proper consideration.

Tim

[crazycheese]

"Manager Embedded Linux and Open Source Graphics."

I don´t know if its bad or good...

Manager "Opensource Graphics" would be WHOLE lot better. The "embedded" thing is making me think, the effort will stay unchanged - ie very minimalistic.

"Opensource driver usable only on low-end cards and its fine" strategy.
"Opensource driver is good for me. Nothing of non-technical to improve." strategy.

I don´t know how good Tim is and what management tools he has at his disposition, but even "embeded" thing on AMD is flawed due to:
1) inefficient powermanagement
2) absence of any popular video acceleration

And then, there are those Intel guys, which opensource driver actually performs and works as it should (I am not calling them saints).
But they refuse to join effort with radeon and nouveau working under gallium..
And somehow they allow newest hardware to work with their driver, so users just buy their graphics for use with opensource driver. Without needs to make choices of driver or different manufacturer.

These are things I´d like to see improved. The general PR-direction of last years of AMD opensource "effort", the way I catch it, is very stagnating. Could be characterized as "broken/flawed everywhere".
The programmers behind it, are no way "broken/flawed" however.

I also hope Mr. Writer won´t be mis-used in form of the responsible person for the "pre-calculated" crash.

Hope, he will be able to jump in and push the airplane out of the stall spirale. That he will be able to find and use hidden potential within opensource direction.

But the "embedded" thing confuses me a LOT, in not a good way...

--- edit

I'm posting this from Firefox running on Ubuntu 12.04 (with XFCE), I have an Android phone, and I avoid running Windows because it just gets in my way.
Tim

Thats very cool to hear! However, this is my personal opinion, the problem is not within "like/dislike".. The problem is really within "not being allowed".
People should be running the OS´es they prefer. But when some OS is pre-destined to be only possible way, and not by internal qualities, but by market-share from-above decisions.. here lies the source of protests.

[PsynoKhi0]

Thanks for the hard work John, best wishes for your future endeavors! Hope you get to stick around and chime in every now and then.
And congratulations Tim on the scapeg.... ermm new title!

[89c51]

As for a longer term vision, I think our Executive Team has outlined a strategy that requires more and better Linux and open source support. Technologies like HSA and markets like cloud and embedded can all benefit from open source so our work supports that vision.

Tim

Tim is the Executive team aware of the problems that we have with the open drivers right now??

[bridgman]

My turn for a rant

Remember that the embedded business is funding 2 of the open source graphics developers, so one might consider saying nice things about embedded and being glad that Tim is covering both areas.

Crazycheese, you are looking at an evolving work in process (where some things take longer than others), assuming that the current state is all that will ever happen, using that assumption as the basis for a hypothetical strategy, claiming that the strategy you made up is ours, then criticizing us for "our" choice of strategy.

If you want progress on power management and UVD that means we need to divert effort from highly visible tasks to internal work where you won't see results until the end.

If you want progress on performance that means we need to divert effort from incremental tweaks to building a new compiler stack from ground up (again, large amounts of work before you see any useful results).

If you want progress on launch-time support that means we need to start working on new products a year before release and long before you can see visible results in public (ditto).

I can understand complaining about there being a bit less visible progress recently. I can understand complaining about current state of performance, power management and launch-time support.

What I don't understand is why the connection between the two isn't obvious, given everything that has been said here.

[AnonymousCoward]

I'm posting this from Firefox running on Ubuntu 12.04 (with XFCE), I have an Android phone, and I avoid running Windows because it just gets in my way. All that is to say that I consider myself a strong Linux and open source advocate with a good understanding of why businesses and individuals choose Linux and open source. While I'm not in position to define AMD's strategy, I do what I can to ensure that Linux and open source are given proper consideration.

Tim

Cool. Btw what GPU are you using on your Linux box, Intel or Nvidia like Bridgman?

[djdoo]

@Bridgman:
John it was a pleasure for me to talk with you here in phoronix...
I wish only successes in your new position, always stay a hardcore opensource engineer and don't leave the place now that Tim takes the wheel !

@Tim:
Welcome man and cool runnings!!
We will always be around to make your head much heavier than it was before you meet us with our questions

[twriter]

Cool. Btw what GPU are you using on your Linux box, Intel or Nvidia like Bridgman?

It's a notebook with ATI Mobility Radeon HD 4200.

Tim

[crazycheese]

My turn for a rant

It wasn´t rant.

Remember that the embedded business is funding 2 of the open source graphics developers, so one might consider saying nice things about embedded and being glad that Tim is covering both areas.

Crazycheese, you are looking at an evolving work in process (where some things take longer than others), assuming that the current state is all that will ever happen, using that assumption as the basis for a hypothetical strategy, claiming that the strategy you made up is ours, then criticizing us for "our" choice of strategy.

If you want progress on power management and UVD that means we need to divert effort from highly visible tasks to internal work where you won't see results until the end.

If you want progress on performance that means we need to divert effort from incremental tweaks to building a new compiler stack from ground up (again, large amounts of work before you see any useful results).

If you want progress on launch-time support that means we need to start working on new products a year before release and long before you can see visible results in public (ditto).

I can understand complaining about there being a bit less visible progress recently. I can understand complaining about current state of performance, power management and launch-time support.

What I don't understand is why the connection between the two isn't obvious, given everything that has been said here.

Thats because logic is for connecting things, not for empowering them.

[RussianNeuroMancer]

John, thank you for talking with us on this forum and for all your work!

Tim, congratulations! I hope you don't forget about XBMC-XvBA developers