Re: Odd performance results

2016-07-13 Thread Paul E. McKenney
On Wed, Jul 13, 2016 at 09:18:17AM +0200, Ingo Molnar wrote: > > * Peter Zijlstra wrote: > > > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > > On 07/12/16 08:05, Paul E. McKenney wrote: > > > The CPU in question (and /proc/cpuinfo should show this)

Re: Odd performance results

2016-07-13 Thread Paul E. McKenney
On Wed, Jul 13, 2016 at 09:18:17AM +0200, Ingo Molnar wrote: > > * Peter Zijlstra wrote: > > > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > > On 07/12/16 08:05, Paul E. McKenney wrote: > > > The CPU in question (and /proc/cpuinfo should show this) has four cores > > >

Re: Odd performance results

2016-07-13 Thread Henrique de Moraes Holschuh
On Wed, 13 Jul 2016, Ingo Molnar wrote: > > The ordering Paul has, namely 0,1 for core0,smt{0,1} is not something > > I've ever seen on an Intel part. AMD otoh does enumerate their CMT stuff > > like what Paul has. > > That's more the natural 'direct' mapping from CPU internal topology to CPU >

Re: Odd performance results

2016-07-13 Thread Henrique de Moraes Holschuh
On Wed, 13 Jul 2016, Ingo Molnar wrote: > > The ordering Paul has, namely 0,1 for core0,smt{0,1} is not something > > I've ever seen on an Intel part. AMD otoh does enumerate their CMT stuff > > like what Paul has. > > That's more the natural 'direct' mapping from CPU internal topology to CPU >

Re: Odd performance results

2016-07-13 Thread Peter Zijlstra
On Wed, Jul 13, 2016 at 09:27:48AM -0300, Henrique de Moraes Holschuh wrote: > On Wed, 13 Jul 2016, Ingo Molnar wrote: > > > The ordering Paul has, namely 0,1 for core0,smt{0,1} is not something > > > I've ever seen on an Intel part. AMD otoh does enumerate their CMT stuff > > > like what Paul

Re: Odd performance results

2016-07-13 Thread Peter Zijlstra
On Wed, Jul 13, 2016 at 09:27:48AM -0300, Henrique de Moraes Holschuh wrote: > On Wed, 13 Jul 2016, Ingo Molnar wrote: > > > The ordering Paul has, namely 0,1 for core0,smt{0,1} is not something > > > I've ever seen on an Intel part. AMD otoh does enumerate their CMT stuff > > > like what Paul

Re: [CRM114spam]: Re: Odd performance results

2016-07-13 Thread Ingo Molnar
* Paul E. McKenney wrote: > > i7-3520M CPU @ 2.90GHz (Dual core with hyperthread, Thinkpad t530, fedora > > 24) > > Glad that it is not just me! ;-) Got one too: aldebaran:~> grep -i . /sys/devices/system/cpu/cpu*/topology/thread_siblings_list | sort -t u -k

Re: [CRM114spam]: Re: Odd performance results

2016-07-13 Thread Ingo Molnar
* Paul E. McKenney wrote: > > i7-3520M CPU @ 2.90GHz (Dual core with hyperthread, Thinkpad t530, fedora > > 24) > > Glad that it is not just me! ;-) Got one too: aldebaran:~> grep -i . /sys/devices/system/cpu/cpu*/topology/thread_siblings_list | sort -t u -k +3 -n

Re: Odd performance results

2016-07-13 Thread Ingo Molnar
* Peter Zijlstra wrote: > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > On 07/12/16 08:05, Paul E. McKenney wrote: > > The CPU in question (and /proc/cpuinfo should show this) has four cores > > with a total of eight threads. The "siblings" and "cpu

Re: Odd performance results

2016-07-13 Thread Ingo Molnar
* Peter Zijlstra wrote: > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > On 07/12/16 08:05, Paul E. McKenney wrote: > > The CPU in question (and /proc/cpuinfo should show this) has four cores > > with a total of eight threads. The "siblings" and "cpu cores" fields in > >

Re: [CRM114spam]: Re: Odd performance results

2016-07-12 Thread Paul E. McKenney
On Tue, Jul 12, 2016 at 08:10:49PM +0100, Dr. David Alan Gilbert wrote: > * Peter Zijlstra (pet...@infradead.org) wrote: > > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > > On 07/12/16 08:05, Paul E. McKenney wrote: > > > The CPU in question (and /proc/cpuinfo should show

Re: [CRM114spam]: Re: Odd performance results

2016-07-12 Thread Paul E. McKenney
On Tue, Jul 12, 2016 at 08:10:49PM +0100, Dr. David Alan Gilbert wrote: > * Peter Zijlstra (pet...@infradead.org) wrote: > > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > > On 07/12/16 08:05, Paul E. McKenney wrote: > > > The CPU in question (and /proc/cpuinfo should show

Re: [CRM114spam]: Re: Odd performance results

2016-07-12 Thread Dr. David Alan Gilbert
* Peter Zijlstra (pet...@infradead.org) wrote: > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > On 07/12/16 08:05, Paul E. McKenney wrote: > > The CPU in question (and /proc/cpuinfo should show this) has four cores > > with a total of eight threads. The "siblings" and "cpu

Re: [CRM114spam]: Re: Odd performance results

2016-07-12 Thread Dr. David Alan Gilbert
* Peter Zijlstra (pet...@infradead.org) wrote: > On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > > On 07/12/16 08:05, Paul E. McKenney wrote: > > The CPU in question (and /proc/cpuinfo should show this) has four cores > > with a total of eight threads. The "siblings" and "cpu

Re: Odd performance results

2016-07-12 Thread Peter Zijlstra
On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > On 07/12/16 08:05, Paul E. McKenney wrote: > The CPU in question (and /proc/cpuinfo should show this) has four cores > with a total of eight threads. The "siblings" and "cpu cores" fields in > /proc/cpuinfo should show the same

Re: Odd performance results

2016-07-12 Thread Peter Zijlstra
On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > On 07/12/16 08:05, Paul E. McKenney wrote: > The CPU in question (and /proc/cpuinfo should show this) has four cores > with a total of eight threads. The "siblings" and "cpu cores" fields in > /proc/cpuinfo should show the same

Re: Odd performance results

2016-07-12 Thread Paul E. McKenney
On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > On 07/12/16 08:05, Paul E. McKenney wrote: > > On Tue, Jul 12, 2016 at 04:55:51PM +0200, Peter Zijlstra wrote: > >> On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: > >>> On Sun, Jul 10, 2016 at 07:17:19AM +0200,

Re: Odd performance results

2016-07-12 Thread Paul E. McKenney
On Tue, Jul 12, 2016 at 10:49:58AM -0700, H. Peter Anvin wrote: > On 07/12/16 08:05, Paul E. McKenney wrote: > > On Tue, Jul 12, 2016 at 04:55:51PM +0200, Peter Zijlstra wrote: > >> On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: > >>> On Sun, Jul 10, 2016 at 07:17:19AM +0200,

Re: Odd performance results

2016-07-12 Thread H. Peter Anvin
On 07/12/16 08:05, Paul E. McKenney wrote: > On Tue, Jul 12, 2016 at 04:55:51PM +0200, Peter Zijlstra wrote: >> On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: >>> On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: On 10 July 2016 06:26:39 CEST, "Paul

Re: Odd performance results

2016-07-12 Thread H. Peter Anvin
On 07/12/16 08:05, Paul E. McKenney wrote: > On Tue, Jul 12, 2016 at 04:55:51PM +0200, Peter Zijlstra wrote: >> On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: >>> On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: On 10 July 2016 06:26:39 CEST, "Paul

Re: Odd performance results

2016-07-12 Thread Paul E. McKenney
On Tue, Jul 12, 2016 at 04:55:51PM +0200, Peter Zijlstra wrote: > On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: > > On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: > > > > > > > > > On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" > > >

Re: Odd performance results

2016-07-12 Thread Paul E. McKenney
On Tue, Jul 12, 2016 at 04:55:51PM +0200, Peter Zijlstra wrote: > On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: > > On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: > > > > > > > > > On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" > > > wrote: > > > >Hello!

Re: Odd performance results

2016-07-12 Thread Peter Zijlstra
On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: > On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: > > > > > > On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" > > wrote: > > >Hello! > > > > > >So I ran a quick benchmark which

Re: Odd performance results

2016-07-12 Thread Peter Zijlstra
On Sun, Jul 10, 2016 at 07:43:27AM -0700, Paul E. McKenney wrote: > On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: > > > > > > On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" > > wrote: > > >Hello! > > > > > >So I ran a quick benchmark which showed stair-step results. I > >

Re: Odd performance results

2016-07-10 Thread Paul E. McKenney
On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: > > > On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" > wrote: > >Hello! > > > >So I ran a quick benchmark which showed stair-step results. I > >immediately > >thought "Ah, this is due to CPU 0 and

Re: Odd performance results

2016-07-10 Thread Paul E. McKenney
On Sun, Jul 10, 2016 at 07:17:19AM +0200, Peter Zijlstra wrote: > > > On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" > wrote: > >Hello! > > > >So I ran a quick benchmark which showed stair-step results. I > >immediately > >thought "Ah, this is due to CPU 0 and 1, 2 and 3, 4 and 5, and 6

Re: Odd performance results

2016-07-09 Thread Peter Zijlstra
On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" wrote: >Hello! > >So I ran a quick benchmark which showed stair-step results. I >immediately >thought "Ah, this is due to CPU 0 and 1, 2 and 3, 4 and 5, and 6 and 7 >being threads in a core." Then I thought "Wait,

Re: Odd performance results

2016-07-09 Thread Peter Zijlstra
On 10 July 2016 06:26:39 CEST, "Paul E. McKenney" wrote: >Hello! > >So I ran a quick benchmark which showed stair-step results. I >immediately >thought "Ah, this is due to CPU 0 and 1, 2 and 3, 4 and 5, and 6 and 7 >being threads in a core." Then I thought "Wait, this is an x86!" >Then I

Odd performance results

2016-07-09 Thread Paul E. McKenney
Hello! So I ran a quick benchmark which showed stair-step results. I immediately thought "Ah, this is due to CPU 0 and 1, 2 and 3, 4 and 5, and 6 and 7 being threads in a core." Then I thought "Wait, this is an x86!" Then I dumped out cpu*/topology/thread_siblings_list, getting the following:

Odd performance results

2016-07-09 Thread Paul E. McKenney
Hello! So I ran a quick benchmark which showed stair-step results. I immediately thought "Ah, this is due to CPU 0 and 1, 2 and 3, 4 and 5, and 6 and 7 being threads in a core." Then I thought "Wait, this is an x86!" Then I dumped out cpu*/topology/thread_siblings_list, getting the following: