Hi,
PD currently doesn't support Hyper-Threading. I've heard all the contradictory about what's the best to do about it.
Maybe nothing, or maybe disabling Hyper-Threading on BIOS setup and using a "non smp" kernel.
Help? Thanks! renf
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
But Pd is actually two processes, the gui and the core, so you should
be able to run each on separate CPUs. On Windows, the gui will show
up as "wish.exe" and the core as "pd.exe".
Anyone done this?
.hc
On May 29, 2006, at 12:40 AM, Renato Fabbri wrote:
Hi,
PD currently doesn't support Hyper-Threading. I've heard all the contradictory about what's the best to do about it.
Maybe nothing, or maybe disabling Hyper-Threading on BIOS setup and using a "non smp" kernel.
Help? Thanks! renf
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
News is what people want to keep hidden and everything else is
publicity.
- Bill Moyers
Nice. Thanks.
How do you make the gui run on one cpu and the core on the other under linux?
--- Hans-Christoph Steiner hans@eds.org wrote:
But Pd is actually two processes, the gui and the core, so you should
be able to run each on separate CPUs. On Windows, the gui will show
up as "wish.exe" and the core as "pd.exe".Anyone done this?
.hc
On May 29, 2006, at 12:40 AM, Renato Fabbri wrote:
Hi,
PD currently doesn't support Hyper-Threading. I've heard all the contradictory about what's the best
to
do about it.
Maybe nothing, or maybe disabling Hyper-Threading
on
BIOS setup and using a "non smp" kernel.
Help? Thanks! renf
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam
protection around
PD-list@iem.at mailing list UNSUBSCRIBE and account-management ->
listinfo/pd-list
News is what people want to keep hidden and everything else is
publicity.- Bill Moyers
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
On Mon, 2006-05-29 at 00:55 +0200, Hans-Christoph Steiner wrote:
But Pd is actually two processes, the gui and the core, so you should
be able to run each on separate CPUs. On Windows, the gui will show
up as "wish.exe" and the core as "pd.exe".
hyperthreading limits the cpu utilization by simulating two processors with 50% of the speed of the whole cpu. it is very unlikely that the gui process takes as much cpu as the pd process. so in a not very unlikely case, one virtual processor, running the main pd process, will have a utilization of 100%, while the other virtual processor is more or less sleeping ...
hyperthreading does make sense for multithreaded/multiprocess applications, where there are two sets of threads/processes, which take more or less the same amount of cpu speed, which is definitely not the case with pd.
the first thing i was doing, when using pd on a hyperthreading p4 for an installation, was to switch off hyperthreading.
hth ... tim
-- TimBlechmann@gmx.de ICQ: 96771783 http://www.mokabar.tk
The composer makes plans, music laughs. Morton Feldman
On Monday 29 May 2006 09:36, Tim Blechmann wrote:
hyperthreading limits the cpu utilization by simulating two processors with 50% of the speed of the whole cpu. it is very unlikely that the gui process takes as much cpu as the pd process. so in a not very unlikely case, one virtual processor, running the main pd process, will have a utilization of 100%, while the other virtual processor is more or less sleeping ...
This is not true. A single thread is allowed to use the full power of the CPU. It's just that if/when it gets pipeline stalled, misses cache etc., another thread can use up its 'wasted' cycles/execution units.
HT can though slow a system down because of various pipeline wierdnesses and an operating system's SMP code not being up to scratch.
http://en.wikipedia.org/wiki/Simultaneous_multithreading
robert.
--- Tim Blechmann TimBlechmann@gmx.net wrote:
On Mon, 2006-05-29 at 00:55 +0200, Hans-Christoph Steiner wrote:
But Pd is actually two processes, the gui and the
core, so you
should
be able to run each on separate CPUs. On Windows,the gui will show
up as "wish.exe" and the core as "pd.exe".
hyperthreading limits the cpu utilization by simulating two processors with 50% of the speed of the whole cpu. it is very unlikely that the gui process takes as much cpu as the pd process. so in a not very unlikely case, one virtual processor, running the main pd process, will have a utilization of 100%, while the other virtual processor is more or less sleeping ...
hyperthreading does make sense for multithreaded/multiprocess applications, where there are two sets of threads/processes, which take more or less the same amount of cpu speed, which is definitely not the case with pd.
the first thing i was doing, when using pd on a hyperthreading p4 for an installation, was to switch off hyperthreading.
hth ... tim
-- TimBlechmann@gmx.de ICQ: 96771783 http://www.mokabar.tk
The composer makes plans, music laughs. Morton Feldman
Ok, makes sense to me.
But, when using PD with the smp kernel and watching a system monitor, all I see is both processors alternating on the tasks making some quasi-quadratic envelopes from zero up, with overlapping.
Where is the system splitting the tasks here??
So, Tim, your advice is to use a non smp kernel and desable hyper-threading at bios setup?
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
On Mon, 2006-05-29 at 10:09 -0700, Renato Fabbri wrote:
But, when using PD with the smp kernel and watching a system monitor, all I see is both processors alternating on the tasks making some quasi-quadratic envelopes from zero up, with overlapping.
Where is the system splitting the tasks here??
not sure ... possibly after sleeping, the pd process will be moved to another processor ...
So, Tim, your advice is to use a non smp kernel and desable hyper-threading at bios setup?
i guess, it's no problem to run an smp kernel, but you should disable hyperthreading. the only exception would be to run two instances of pd, with a similar cpu load, running on top of a jackdmp server ...
tim
-- TimBlechmann@gmx.de ICQ: 96771783 http://www.mokabar.tk
Linux is like a wigwam: no windows, no gates, apache inside, stable.
Hi,
sorry for taking over this thread. I know this is sort of OT but I don't quite understand the topic.
Does this all mean a) it is no advantage for PD to run with/on a Core Duo processor (on WinXP or linux) b) it is a disadvantage c) it is a disadvantage unless I deliberately make _no_ use of the virtual multiprocessor?
If this thread isn't related to Core Duo then I'd appreciate any comments on my questions (since I'm trying to find the right "next laptop").
Thanks Urs
Tim Blechmann schrieb:
On Mon, 2006-05-29 at 10:09 -0700, Renato Fabbri wrote:
But, when using PD with the smp kernel and watching a system monitor, all I see is both processors alternating on the tasks making some quasi-quadratic envelopes from zero up, with overlapping.
Where is the system splitting the tasks here??
not sure ... possibly after sleeping, the pd process will be moved to another processor ...
So, Tim, your advice is to use a non smp kernel and desable hyper-threading at bios setup?
i guess, it's no problem to run an smp kernel, but you should disable hyperthreading. the only exception would be to run two instances of pd, with a similar cpu load, running on top of a jackdmp server ...
tim
-- TimBlechmann@gmx.de ICQ: 96771783 http://www.mokabar.tk
Linux is like a wigwam: no windows, no gates, apache inside, stable.
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
CoreDuo has two full CPU cores on each die. It is not HyperThreading technology. Pd runs on a Core CPU just as well as any other recent Intel CPU.
On 6/2/06, Urs Liska pd@ursliska.de wrote:
Hi,
sorry for taking over this thread. I know this is sort of OT but I don't quite understand the topic.
Does this all mean a) it is no advantage for PD to run with/on a Core Duo processor (on WinXP or linux) b) it is a disadvantage c) it is a disadvantage unless I deliberately make _no_ use of the virtual multiprocessor?
If this thread isn't related to Core Duo then I'd appreciate any comments on my questions (since I'm trying to find the right "next laptop").
Thanks Urs
Test it and find out. Computer hardware and operating systems are so
complex that no one can really know how software is going to perform
without testing it.
If you do test it, please report back on this list.
.hc
On Jun 2, 2006, at 9:03 PM, Urs Liska wrote:
Hi,
sorry for taking over this thread. I know this is sort of OT but I don't quite understand the topic.
Does this all mean a) it is no advantage for PD to run with/on a Core Duo processor
(on WinXP or linux) b) it is a disadvantage c) it is a disadvantage unless I deliberately make _no_ use of the
virtual multiprocessor?If this thread isn't related to Core Duo then I'd appreciate any
comments on my questions (since I'm trying to find the right "next
laptop").Thanks Urs
Tim Blechmann schrieb:
On Mon, 2006-05-29 at 10:09 -0700, Renato Fabbri wrote:
But, when using PD with the smp kernel and watching a system monitor, all I see is both processors alternating on the tasks making some quasi-quadratic envelopes from zero up, with overlapping.
Where is the system splitting the tasks here??
not sure ... possibly after sleeping, the pd process will be moved to another processor ...
So, Tim, your advice is to use a non smp kernel and desable hyper-threading at bios setup?
i guess, it's no problem to run an smp kernel, but you should disable hyperthreading. the only exception would be to run two instances of pd, with a
similar cpu load, running on top of a jackdmp server ... tim -- TimBlechmann@gmx.de ICQ: 96771783 http://www.mokabar.tk Linux is like a wigwam: no windows, no gates, apache inside, stable.
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
-- Urs Liska Glümerstr. 5 D-79102 Freiburg
www.graft-music.com www.suonomobile.de
[Pd 0.39.2, WinXP]
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
"Computer science is no more related to the computer than astronomy
is related to the telescope."
-Edsger
Dykstra
On Fri, 2006-06-02 at 21:03 +0200, Urs Liska wrote:
a) it is no advantage for PD to run with/on a Core Duo processor (on WinXP or linux)
no
b) it is a disadvantage
you won't be able to use all cpu power for dsp processing
pd does work on multiprocessor machines, but it is designed for single-processor machines, so it will gracefully ignore any other processors ...
it's like a car with two engines... instead of both engines driving the wheels, one engine is driving the wheels and one the window lifts and windscreen wipers :)
hth ... tim
-- TimBlechmann@gmx.de ICQ: 96771783 http://www.mokabar.tk
Most of the trouble in this world has been caused by folks who can't mind their own business, because they have no business of their own to mind, any more than a smallpox virus has. William S. Burroughs
pd does work on multiprocessor machines, but it is designed for single-processor machines, so it will gracefully ignore any other processors ...
http://sourceforge.net/projects/jmax/
AvS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
` |Schreck Ensemble . . . . . . . . . . . . . . . . . . . . +
` |# -laboratory for live electro-acoustic music- # |
| http://www.schreck.nl/ |
| http://www.xs4all.nl/~schreck/ |
` *===========================================================++
` |Compositions http://www.xs4all.nl/~schreck/html/compo.html |
` |Samples http://www.xs4all.nl/~schreck/html/samp.html |
` |Patches http://www.xs4all.nl/~schreck/html/pat.html |
` |Videos http://www.xs4all.nl/~schreck/html/video.html |
` |Scores http://www.xs4all.nl/~schreck/html/scores.html |
*===========================================================++
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
hi all,
Am Samstag, 3. Juni 2006 00:25 schrieb Arie van Schutterhoef:
pd does work on multiprocessor machines, but it is designed for single-processor machines, so it will gracefully ignore any other processors ...
hahaha, im sure tim's statement was not geared towards a gui using most of the cpu time, even if it can spread dsp & cpu across serveral cpu's .... instead it was about having the dsp running on seperate cpu's simultanously .... one thing that jmax an not do as well ..
AvS
greets,
chris
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
` |Schreck Ensemble . . . . . . . . . . . . . . . . . . . . +
` |# -laboratory for live electro-acoustic music- # | | http://www.schreck.nl/ | | http://www.xs4all.nl/~schreck/ | ` *===========================================================++ ` |Compositions http://www.xs4all.nl/~schreck/html/compo.html | ` |Samples http://www.xs4all.nl/~schreck/html/samp.html | ` |Patches http://www.xs4all.nl/~schreck/html/pat.html | ` |Videos http://www.xs4all.nl/~schreck/html/video.html | ` |Scores http://www.xs4all.nl/~schreck/html/scores.html | *===========================================================++. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
about having the dsp running on seperate cpu's simultanously http://music.york.ac.uk/rimm/tech.htm
AvS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
` |Schreck Ensemble . . . . . . . . . . . . . . . . . . . . +
` |# -laboratory for live electro-acoustic music- # |
| http://www.schreck.nl/ |
| http://www.xs4all.nl/~schreck/ |
` *===========================================================++
` |Compositions http://www.xs4all.nl/~schreck/html/compo.html |
` |Samples http://www.xs4all.nl/~schreck/html/samp.html |
` |Patches http://www.xs4all.nl/~schreck/html/pat.html |
` |Videos http://www.xs4all.nl/~schreck/html/video.html |
` |Scores http://www.xs4all.nl/~schreck/html/scores.html |
*===========================================================++
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
-
Arie van Schutterhoef -
chris clepper -
Christian Klippel -
Hans-Christoph Steiner -
Renato Fabbri -
Robert Scott -
Tim Blechmann -
Urs Liska