Re: multi-CPU optimization inside a distribution

2016-07-02 Thread Solomon Peachy
On Fri, Jul 01, 2016 at 02:08:48PM -0400, Josh Boyer wrote: > So the thing is, I'm genuinely curious why you would want to run a > desktop OS on either architecture. For the same reason Fedora puts out releases targeting those architechures? :) I'm not wedded to any particular hardware archite

Re: multi-CPU optimization inside a distribution

2016-07-01 Thread Richard W.M. Jones
On Fri, Jul 01, 2016 at 01:42:40PM -0400, Solomon Peachy wrote: > With very few exceptions (eg Gigabyte's mostly-unobtainable MP30-AR0 or > some locked-down appliance-type servers) they've all been ultra-dense > high-end boxes (dozens of cores per U) priced into the stratosphere or > glorified c

Re: multi-CPU optimization inside a distribution

2016-07-01 Thread Josh Boyer
On Fri, Jul 1, 2016 at 1:42 PM, Solomon Peachy wrote: > On Fri, Jul 01, 2016 at 11:35:27AM -0500, Bruno Wolff III wrote: >> If I could afford it, I'd pay a significant premium to get a Talos board. >> But as it is, I'm going to be stuck using second hand, pretty old computers >> for the foreseeabl

Re: multi-CPU optimization inside a distribution

2016-07-01 Thread Solomon Peachy
On Fri, Jul 01, 2016 at 11:35:27AM -0500, Bruno Wolff III wrote: > If I could afford it, I'd pay a significant premium to get a Talos board. > But as it is, I'm going to be stuck using second hand, pretty old computers > for the foreseeable future. And even though "significant premium" is a polite

Re: multi-CPU optimization inside a distribution

2016-07-01 Thread Bruno Wolff III
On Fri, Jul 01, 2016 at 00:34:02 -0400, Gerald Henriksen wrote: To get developers they need to get a CPU/motherboard combination out for under $1000, likely closer to $500, as these will be purchases as a secondary system to "play with" because there is no direct way to make money as a POWER d

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Gerald Henriksen
On Thu, 30 Jun 2016 22:07:24 +0200, you wrote: >On Thu, 30 Jun 2016 15:59:38 -0400 >Solomon Peachy wrote: > >> On Thu, Jun 30, 2016 at 08:04:57PM +0100, Richard W.M. Jones wrote: >> > I will just snipe here that the situation would be better if POWER >> > > 4 hardware was generally available to b

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Gerald Henriksen
On Thu, 30 Jun 2016 18:01:38 +0200, you wrote: >> Realistically, it isn't about either specifically. Each iteration of >> POWER tends to require tuning specifically for that generation if you >> want to get the most performance out of your software. > >Latest CPU features get utilized mostly only

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Dan Horák
On Thu, 30 Jun 2016 15:59:38 -0400 Solomon Peachy wrote: > On Thu, Jun 30, 2016 at 08:04:57PM +0100, Richard W.M. Jones wrote: > > I will just snipe here that the situation would be better if POWER > > > 4 hardware was generally available to buy. > > The upcoming Raptor Talos POWER8 board [1] lo

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Solomon Peachy
On Thu, Jun 30, 2016 at 08:04:57PM +0100, Richard W.M. Jones wrote: > I will just snipe here that the situation would be better if POWER > 4 > hardware was generally available to buy. The upcoming Raptor Talos POWER8 board [1] looks promising, but at $3700 a pop just for the motherboard and an en

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Richard W.M. Jones
On Thu, Jun 30, 2016 at 12:12:37PM +0200, Dan Horák wrote: > For ppc64 (the big endian POWER) the base is set by the toolchain > default which is Power4/ppc970. When Power7 came we were asked what are > the options to take the advantage of these CPUs, 3 generations newer > than the base. The soluti

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Josh Boyer
On Thu, Jun 30, 2016 at 12:01 PM, Jan Kratochvil wrote: > On Thu, 30 Jun 2016 16:13:59 +0200, Josh Boyer wrote: >> On Thu, Jun 30, 2016 at 9:53 AM, Jan Kratochvil >> wrote: >> > But now there is ppc64le for >=Power8. So <=Power7 can use ppc64 but newer >> >>=Power8 hardware does use ppc64le whi

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Peter Robinson
On Thu, Jun 30, 2016 at 5:01 PM, Jan Kratochvil wrote: > On Thu, 30 Jun 2016 16:13:59 +0200, Josh Boyer wrote: >> On Thu, Jun 30, 2016 at 9:53 AM, Jan Kratochvil >> wrote: >> > But now there is ppc64le for >=Power8. So <=Power7 can use ppc64 but newer >> >>=Power8 hardware does use ppc64le whic

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Jan Kratochvil
On Thu, 30 Jun 2016 16:13:59 +0200, Josh Boyer wrote: > On Thu, Jun 30, 2016 at 9:53 AM, Jan Kratochvil > wrote: > > But now there is ppc64le for >=Power8. So <=Power7 can use ppc64 but newer > >>=Power8 hardware does use ppc64le which is already Power8 optimized. > > So isn't this problem (some

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Peter Robinson
On Thu, Jun 30, 2016 at 2:53 PM, Jan Kratochvil wrote: > On Thu, 30 Jun 2016 12:12:37 +0200, Dan Horák wrote: >> For ppc64 (the big endian POWER) the base is set by the toolchain >> default which is Power4/ppc970. When Power7 came we were asked what are >> the options to take the advantage of thes

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Josh Boyer
On Thu, Jun 30, 2016 at 9:53 AM, Jan Kratochvil wrote: > On Thu, 30 Jun 2016 12:12:37 +0200, Dan Horák wrote: >> For ppc64 (the big endian POWER) the base is set by the toolchain >> default which is Power4/ppc970. When Power7 came we were asked what are >> the options to take the advantage of thes

Re: multi-CPU optimization inside a distribution

2016-06-30 Thread Jan Kratochvil
On Thu, 30 Jun 2016 12:12:37 +0200, Dan Horák wrote: > For ppc64 (the big endian POWER) the base is set by the toolchain > default which is Power4/ppc970. When Power7 came we were asked what are > the options to take the advantage of these CPUs, 3 generations newer > than the base. The solution was

multi-CPU optimization inside a distribution

2016-06-30 Thread Dan Horák
Hi all, we were talking about this item for some time, so let's start a thread for it to have the discussion and hopefully also a solution documented. This is meant as discussion initiation based on the situation in Fedora on POWER. I would like to ask the bigger experts than me to fill the missin