Re: About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-10 Thread Jerome Kieffer
On Thu, 9 Feb 2017 16:06:56 +0100 Michael Banck wrote: > > After spending a week or two of my PhD on that, I compared the performances > > with the one from the stock debian build and it turned out to be within > > 1% my costly build. Very disappointing to me. Good job you

Re: About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-09 Thread Michael Banck
On Thu, Feb 09, 2017 at 02:13:24PM +0100, Jerome Kieffer wrote: > On Thu, 9 Feb 2017 13:19:09 +0100 > Julian Taylor wrote: > > > These numbers are very surprising to me. > > I do agree with Julian ... > Once upon a time, I have been a computational chemist and

Re: About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-09 Thread Jerome Kieffer
On Thu, 9 Feb 2017 13:19:09 +0100 Julian Taylor wrote: > These numbers are very surprising to me. I do agree with Julian ... Once upon a time, I have been a computational chemist and believed I could get this 2-6x speed up by recompiling the whole scientific stack

Re: About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-09 Thread Julian Taylor
On 09.02.2017 07:23, lumin wrote: > >> So do you think we are doing a bad service to our users by striping >> -march=native? Could you please provide some numbers? > > No, we are not doing bad. Nobody is wrong. We cannot gain compatibility > and performance at the same time. I don't remember

Re: About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-08 Thread lumin
> So do you think we are doing a bad service to our users by striping > -march=native?  Could you please provide some numbers? No, we are not doing bad. Nobody is wrong. We cannot gain compatibility and performance at the same time. I don't remember the exact numbers of those experiments

Re: About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-08 Thread Julien Puydt
Le 08/02/2017 à 18:10, Andreas Tille a écrit : > I wonder whether we could invent some mechanism that is rebuilding a > package in postinst and installs the result on the machine instead of a > pre-build binary. Or we could provide some toolset which enables > scientists to download a set of

About the sense of removing -march=native (Was: Is theano worth saving?)

2017-02-08 Thread Andreas Tille
Hi Lumin, thanks for your long explanation. While I have a divergent opinion in some points I do not feel competent enough to discuss this here. However, I stumbled upon one point which I consider a valuable topic to move to debian-de...@lists.debian.org. On Wed, Feb 08, 2017 at 03:33:16PM