Re: [boinc_dev] CPU throttling and GPU apps

2013-07-09 Thread McLeod, John
The only problem is that there is no real standard for determining the 
temperature of the CPU (every manufacturer seems to have done it differently.  
Speed fan is keeping up with this, it is probably not something that BOINC 
developers really should be attempting to do.

-Original Message-
From: Charles Elliott [mailto:elliott...@verizon.net] 
Sent: Tuesday, July 09, 2013 7:15 AM
To: McLeod, John; 'David Anderson'; boinc_dev@ssl.berkeley.edu
Subject: RE: [boinc_dev] CPU throttling and GPU apps

Three of my computers are in a small second floor bedroom with an asphalt
roof and a box fan placed in a window close to the ground for its only
ventilation.  During the afternoons room temperatures are 95 F or higher,
too much for the computers.

If heat is the problem, it makes more sense to measure CPU/GPU temperatures
and reduce the load on a device if the temperature rises above a set point,
and increase it if the temps go below a lower set point.

I wrote a Java program that does exactly that.  It accesses temperatures by
parsing a SpeedFan logging file on each computer, and affects Boinc by
changing the ncpusN/ncpus line in cc_config.xml and issuing a
read_cc_config RPC.  Something similar might be possible for the GPUs.

It seems to work.  Yesterday, its first day of operation, during the
afternoon it lowered the CPUs used of the three computers from 8 to 6, and
then last night, starting about 10:30, it slowly raised them back to 8,
where it is now.

Charles Elliott

 -Original Message-
 From: boinc_dev [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf
 Of McLeod, John
 Sent: Monday, July 8, 2013 11:21 AM
 To: David Anderson; boinc_dev@ssl.berkeley.edu
 Subject: Re: [boinc_dev] CPU throttling and GPU apps
 
 How about changing it to not throttle apps that use less than the
 current throttling value?  E.g. if throttling is set at .9, don't
 throttle a task that uses .8.
 
 -Original Message-
 From: boinc_dev [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf
 Of David Anderson
 Sent: Friday, July 05, 2013 11:50 PM
 To: boinc_dev@ssl.berkeley.edu
 Subject: Re: [boinc_dev] CPU throttling and GPU apps
 
 I changed it not throttle apps that use  .5 CPUs
 -- David
 
 On 04-Jul-2013 2:38 PM, Eric J Korpela wrote:
  The only pro I can think of would be to reduce GPU use to keep
  temperature or power use down, but that would be better implemented
 as
  GPU throttling.
 
  On Thu, Jul 4, 2013 at 5:52 AM, Bernd Machenschalk
  bernd.machensch...@aei.mpg.de wrote:
  On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:
 
  I guess there are several pros and cons, e.g.:
 
  cons:
  - one one hand, GPU apps (depending on the CPU share?) get a
 higher OS
  prio (in terms of niceness) to prevent the GPU being starved.
 Throttling
  the CPU might very well cause this starvation
  - if a GPU app has a rather low CPU runtime share in the first
 place,
  further CPU throttling does not seem too useful.
  - in order to avoid GPU load to interfere with the user doing
 non-BOINC
  related stuff, there is already the setting Suspend GPU work while
  computer is in use.
 
 
  Here's one more:
 
  When not synchronized with GPU-CPU communication (kernel launches,
 data
  transfer) throtteling an App can break any running GPU task. I'm not
 sure
  whether the throtteling implementations of all BOINC Clients that
 are being
  used properly honor critical sections, nor am I that all GPU apps of
 all
  projects make proper use of these.
 
 
  pros:
  I can't think about many
 
 
  Actually I can't think about any.
 
  Best,
  Bernd
 
 
  ___
  boinc_dev mailing list
  boinc_dev@ssl.berkeley.edu
  http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
  To unsubscribe, visit the above URL and
  (near bottom of page) enter your email address.
  ___
  boinc_dev mailing list
  boinc_dev@ssl.berkeley.edu
  http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
  To unsubscribe, visit the above URL and
  (near bottom of page) enter your email address.
 
 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.
 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.

___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


Re: [boinc_dev] CPU throttling and GPU apps

2013-07-09 Thread David Anderson

Charles:
If this program is ready for public use, we could list it here:
http://boinc.berkeley.edu/addons.php
-- David

On 09-Jul-2013 4:15 AM, Charles Elliott wrote:

Three of my computers are in a small second floor bedroom with an asphalt
roof and a box fan placed in a window close to the ground for its only
ventilation.  During the afternoons room temperatures are 95 F or higher,
too much for the computers.

If heat is the problem, it makes more sense to measure CPU/GPU temperatures
and reduce the load on a device if the temperature rises above a set point,
and increase it if the temps go below a lower set point.

I wrote a Java program that does exactly that.  It accesses temperatures by
parsing a SpeedFan logging file on each computer, and affects Boinc by
changing the ncpusN/ncpus line in cc_config.xml and issuing a
read_cc_config RPC.  Something similar might be possible for the GPUs.

It seems to work.  Yesterday, its first day of operation, during the
afternoon it lowered the CPUs used of the three computers from 8 to 6, and
then last night, starting about 10:30, it slowly raised them back to 8,
where it is now.

Charles Elliott


-Original Message-
From: boinc_dev [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf
Of McLeod, John
Sent: Monday, July 8, 2013 11:21 AM
To: David Anderson; boinc_dev@ssl.berkeley.edu
Subject: Re: [boinc_dev] CPU throttling and GPU apps

How about changing it to not throttle apps that use less than the
current throttling value?  E.g. if throttling is set at .9, don't
throttle a task that uses .8.

-Original Message-
From: boinc_dev [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf
Of David Anderson
Sent: Friday, July 05, 2013 11:50 PM
To: boinc_dev@ssl.berkeley.edu
Subject: Re: [boinc_dev] CPU throttling and GPU apps

I changed it not throttle apps that use  .5 CPUs
-- David

On 04-Jul-2013 2:38 PM, Eric J Korpela wrote:

The only pro I can think of would be to reduce GPU use to keep
temperature or power use down, but that would be better implemented

as

GPU throttling.

On Thu, Jul 4, 2013 at 5:52 AM, Bernd Machenschalk
bernd.machensch...@aei.mpg.de wrote:

On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:


I guess there are several pros and cons, e.g.:

cons:
 - one one hand, GPU apps (depending on the CPU share?) get a

higher OS

prio (in terms of niceness) to prevent the GPU being starved.

Throttling

the CPU might very well cause this starvation
 - if a GPU app has a rather low CPU runtime share in the first

place,

further CPU throttling does not seem too useful.
 - in order to avoid GPU load to interfere with the user doing

non-BOINC

related stuff, there is already the setting Suspend GPU work while
computer is in use.



Here's one more:

When not synchronized with GPU-CPU communication (kernel launches,

data

transfer) throtteling an App can break any running GPU task. I'm not

sure

whether the throtteling implementations of all BOINC Clients that

are being

used properly honor critical sections, nor am I that all GPU apps of

all

projects make proper use of these.



pros:
 I can't think about many



Actually I can't think about any.

Best,
Bernd


___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.



___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


Re: [boinc_dev] CPU throttling and GPU apps

2013-07-08 Thread McLeod, John
How about changing it to not throttle apps that use less than the current 
throttling value?  E.g. if throttling is set at .9, don't throttle a task that 
uses .8.

-Original Message-
From: boinc_dev [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf Of David 
Anderson
Sent: Friday, July 05, 2013 11:50 PM
To: boinc_dev@ssl.berkeley.edu
Subject: Re: [boinc_dev] CPU throttling and GPU apps

I changed it not throttle apps that use  .5 CPUs
-- David

On 04-Jul-2013 2:38 PM, Eric J Korpela wrote:
 The only pro I can think of would be to reduce GPU use to keep
 temperature or power use down, but that would be better implemented as
 GPU throttling.

 On Thu, Jul 4, 2013 at 5:52 AM, Bernd Machenschalk
 bernd.machensch...@aei.mpg.de wrote:
 On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:

 I guess there are several pros and cons, e.g.:

 cons:
 - one one hand, GPU apps (depending on the CPU share?) get a higher OS
 prio (in terms of niceness) to prevent the GPU being starved. Throttling
 the CPU might very well cause this starvation
 - if a GPU app has a rather low CPU runtime share in the first place,
 further CPU throttling does not seem too useful.
 - in order to avoid GPU load to interfere with the user doing non-BOINC
 related stuff, there is already the setting Suspend GPU work while
 computer is in use.


 Here's one more:

 When not synchronized with GPU-CPU communication (kernel launches, data
 transfer) throtteling an App can break any running GPU task. I'm not sure
 whether the throtteling implementations of all BOINC Clients that are being
 used properly honor critical sections, nor am I that all GPU apps of all
 projects make proper use of these.


 pros:
 I can't think about many


 Actually I can't think about any.

 Best,
 Bernd


 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.
 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.

___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


Re: [boinc_dev] CPU throttling and GPU apps

2013-07-08 Thread Richard Haselgrove
What do we mean by 'use' a certain fraction of a CPU, anyway?

AFAIK, projects have a rather crude tool by which they declare what proportion 
of an application's fpops are performed on the CPU - cpu_fracx/cpu_frac in 
the xml format of the plan class definition - and nothing else. The *scheduled* 
CPU usage (and I assume this is what David is referring to) is calculated by 
the server from this frac and the relative speeds of the host's CPU and GPU.

These limited tools can lead to assumptions which are widely divergent from 
reality, in either direction.

A concrete example: I run two hosts on GPUGrid (who should know a thing or two 
about GPU programming)

Host 43404 is told to schedule 0.55 CPUs for each GPU task: host 132158 is told 
to schedule 0.667 CPUs. Not a great deal of difference.

But compare the reality.
http://www.gpugrid.net/results.php?hostid=43404state=3 uses typically ~900 CPU 
seconds per task, just 0.06 CPUs
http://www.gpugrid.net/results.php?hostid=132158state=3 uses ~30,000 CPU 
seconds per task, or over 0.99 CPUs


Both hosts are running CUDA 4.2 applications/runtime, but there are many 
differences between them.

GPU type: Fermi/Kepler
GPU driver: 310.70/314.22
Operating system: WinXP-32/Win7-64

BOINC version: v6.12.34/v7.0.64
BOINC mode: Service/User
Application: Short (v6.52)/Long (v6.18)

We could - conceivably - program the BOINC platform to take account of all 
these variables, but I suspect that would be absurdly complicated, and quite 
probably fruitless (from reading the project message boards, I suspect the main 
cause of the difference between 0.06 CPU and 0.99 CPU usage is the current 
handling of the newer Kepler architecture - so an application-level, rather 
than BOINC, issue)

All-in-all, I think it would be better to go back to using the CPU 
throttle/thermal control feature for its original purpose - controlling the 
thermal output of pure-CPU apps (defined as apps which have *no* co-processor 
specified). If we need a thermal control mechanism for GPUs (which was the 
original request by Admin Team St.Petersburg), add one properly tailored to 
the hardware characteristics of co-processors.




 From: McLeod, John john.mcl...@sap.com
To: David Anderson da...@ssl.berkeley.edu; boinc_dev@ssl.berkeley.edu 
boinc_dev@ssl.berkeley.edu 
Sent: Monday, 8 July 2013, 16:21
Subject: Re: [boinc_dev] CPU throttling and GPU apps
 

How about changing it to not throttle apps that use less than the current 
throttling value?  E.g. if throttling is set at .9, don't throttle a task that 
uses .8.

-Original Message-
From: boinc_dev [mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf Of David 
Anderson
Sent: Friday, July 05, 2013 11:50 PM
To: boinc_dev@ssl.berkeley.edu
Subject: Re: [boinc_dev] CPU throttling and GPU apps

I changed it not throttle apps that use  .5 CPUs
-- David

On 04-Jul-2013 2:38 PM, Eric J Korpela wrote:
 The only pro I can think of would be to reduce GPU use to keep
 temperature or power use down, but that would be better implemented as
 GPU throttling.

 On Thu, Jul 4, 2013 at 5:52 AM, Bernd Machenschalk
 bernd.machensch...@aei.mpg.de wrote:
 On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:

 I guess there are several pros and cons, e.g.:

 cons:
     - one one hand, GPU apps (depending on the CPU share?) get a higher OS
 prio (in terms of niceness) to prevent the GPU being starved. Throttling
 the CPU might very well cause this starvation
     - if a GPU app has a rather low CPU runtime share in the first place,
 further CPU throttling does not seem too useful.
     - in order to avoid GPU load to interfere with the user doing non-BOINC
 related stuff, there is already the setting Suspend GPU work while
 computer is in use.


 Here's one more:

 When not synchronized with GPU-CPU communication (kernel launches, data
 transfer) throtteling an App can break any running GPU task. I'm not sure
 whether the throtteling implementations of all BOINC Clients that are being
 used properly honor critical sections, nor am I that all GPU apps of all
 projects make proper use of these.


 pros:
     I can't think about many


 Actually I can't think about any.

 Best,
 Bernd


 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.
 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.

___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address

Re: [boinc_dev] CPU throttling and GPU apps

2013-07-05 Thread David Anderson

I changed it not throttle apps that use  .5 CPUs
-- David

On 04-Jul-2013 2:38 PM, Eric J Korpela wrote:

The only pro I can think of would be to reduce GPU use to keep
temperature or power use down, but that would be better implemented as
GPU throttling.

On Thu, Jul 4, 2013 at 5:52 AM, Bernd Machenschalk
bernd.machensch...@aei.mpg.de wrote:

On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:


I guess there are several pros and cons, e.g.:

cons:
- one one hand, GPU apps (depending on the CPU share?) get a higher OS
prio (in terms of niceness) to prevent the GPU being starved. Throttling
the CPU might very well cause this starvation
- if a GPU app has a rather low CPU runtime share in the first place,
further CPU throttling does not seem too useful.
- in order to avoid GPU load to interfere with the user doing non-BOINC
related stuff, there is already the setting Suspend GPU work while
computer is in use.



Here's one more:

When not synchronized with GPU-CPU communication (kernel launches, data
transfer) throtteling an App can break any running GPU task. I'm not sure
whether the throtteling implementations of all BOINC Clients that are being
used properly honor critical sections, nor am I that all GPU apps of all
projects make proper use of these.



pros:
I can't think about many



Actually I can't think about any.

Best,
Bernd


___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


[boinc_dev] CPU throttling and GPU apps

2013-07-04 Thread Heinz-Bernd Eggenstein
Hi all!

While hunting for a bug recently, we at Einstein@Home came across a 
question that I would like to present for discussion here:

If an app has a non zero GPU share (GPU app for short), should CPU 
throttling (as configured thru the preferences setting Use at most x % of 
CPU time) be applied to it?

I guess there are several pros and cons, e.g.:

cons: 
  - one one hand, GPU apps (depending on the CPU share?) get a higher OS 
prio (in terms of niceness) to prevent the GPU being starved. Throttling 
the CPU might very well cause this starvation
  - if a GPU app has a rather low CPU runtime share in the first place, 
further CPU throttling does not seem too useful.
  - in order to avoid GPU load to interfere with the user doing non-BOINC 
related stuff, there is already the setting Suspend GPU work while 
computer is in use.  

pros:
  I can't think about many, maybe consistency and user expectation?

Volunteer reports at E@H seem to suggest that in the current BOINC client 
version, GPU apps are indeed CPU throttled, right?
Browsing thru the source code, my initial impression is that only NCI 
(non-CPU-intensive) apps are excluded from throttling.

Cheers
HB 

-
Heinz-Bernd Eggenstein
Max Planck Institute for Gravitational Physics
Callinstrasse 38
D-30167 Hannover,  Germany
Tel.: +49-511-762-19466 (Room 037)
___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


Re: [boinc_dev] CPU throttling and GPU apps

2013-07-04 Thread Richard Haselgrove
We had some discussion about this issue back in January.

For most of the lifetime of BOINC on GPUs, CPU throttling has *not* been 
applied to GPU apps:

it was dis-applied on 27 Oct 2008 - 
http://boinc.berkeley.edu/trac/changeset/3719268f1c3807dcac821c455090e2243133bb8d/boinc-old
and re-applied on 12 Jan 2013 - 
http://boinc.berkeley.edu/trac/changeset/04de2936d47256edc77ba032541b096aff69b2d5/boinc-old

The re-application seemed to be at the request of a single user: David wrote

I checked in a change so that CPU throttling applies to GPU apps also.
Can anyone think of a reason not to do this?
-- David

On 11-Jan-2013 11:42 PM, Admin Team St.Petersburg wrote:
 And here's another. In the client, the CPU has the ability to change power
 steering, and the GPU can not. To do.


That message, and the subsequent discussion, can be followed at 
http://lists.ssl.berkeley.edu/pipermail/boinc_dev/2013-January/019305.html

From memory (and a very quick refresh view of that thread) the main response 
was a request for a separate control to throttle GPU apps if desired - we 
concluded with

The request for a separate GPU throttling preference is noted; however we're 
not going to get to it immediately.
-- David




 From: Heinz-Bernd Eggenstein heinz-bernd.eggenst...@aei.mpg.de
To: boinc_dev@ssl.berkeley.edu 
Sent: Thursday, 4 July 2013, 12:15
Subject: [boinc_dev] CPU throttling and GPU apps
 

Hi all!

While hunting for a bug recently, we at Einstein@Home came across a 
question that I would like to present for discussion here:

If an app has a non zero GPU share (GPU app for short), should CPU 
throttling (as configured thru the preferences setting Use at most x % of 
CPU time) be applied to it?

I guess there are several pros and cons, e.g.:

cons: 
  - one one hand, GPU apps (depending on the CPU share?) get a higher OS 
prio (in terms of niceness) to prevent the GPU being starved. Throttling 
the CPU might very well cause this starvation
  - if a GPU app has a rather low CPU runtime share in the first place, 
further CPU throttling does not seem too useful.
  - in order to avoid GPU load to interfere with the user doing non-BOINC 
related stuff, there is already the setting Suspend GPU work while 
computer is in use.  

pros:
  I can't think about many, maybe consistency and user expectation?

Volunteer reports at E@H seem to suggest that in the current BOINC client 
version, GPU apps are indeed CPU throttled, right?
Browsing thru the source code, my initial impression is that only NCI 
(non-CPU-intensive) apps are excluded from throttling.

Cheers
HB 

-
Heinz-Bernd Eggenstein
Max Planck Institute for Gravitational Physics
Callinstrasse 38
D-30167 Hannover,  Germany
Tel.: +49-511-762-19466 (Room 037)
___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.



___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


Re: [boinc_dev] CPU throttling and GPU apps

2013-07-04 Thread Bernd Machenschalk

On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:


I guess there are several pros and cons, e.g.:

cons:
   - one one hand, GPU apps (depending on the CPU share?) get a higher OS
prio (in terms of niceness) to prevent the GPU being starved. Throttling
the CPU might very well cause this starvation
   - if a GPU app has a rather low CPU runtime share in the first place,
further CPU throttling does not seem too useful.
   - in order to avoid GPU load to interfere with the user doing non-BOINC
related stuff, there is already the setting Suspend GPU work while
computer is in use.


Here's one more:

When not synchronized with GPU-CPU communication (kernel launches, data transfer) throtteling an App can break any running GPU task. I'm not sure 
whether the throtteling implementations of all BOINC Clients that are being used properly honor critical sections, nor am I that all GPU apps of all 
projects make proper use of these.



pros:
   I can't think about many


Actually I can't think about any.

Best,
Bernd

___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.


Re: [boinc_dev] CPU throttling and GPU apps

2013-07-04 Thread Eric J Korpela
The only pro I can think of would be to reduce GPU use to keep
temperature or power use down, but that would be better implemented as
GPU throttling.

On Thu, Jul 4, 2013 at 5:52 AM, Bernd Machenschalk
bernd.machensch...@aei.mpg.de wrote:
 On 04.07.13 13:15, Heinz-Bernd Eggenstein wrote:

 I guess there are several pros and cons, e.g.:

 cons:
- one one hand, GPU apps (depending on the CPU share?) get a higher OS
 prio (in terms of niceness) to prevent the GPU being starved. Throttling
 the CPU might very well cause this starvation
- if a GPU app has a rather low CPU runtime share in the first place,
 further CPU throttling does not seem too useful.
- in order to avoid GPU load to interfere with the user doing non-BOINC
 related stuff, there is already the setting Suspend GPU work while
 computer is in use.


 Here's one more:

 When not synchronized with GPU-CPU communication (kernel launches, data
 transfer) throtteling an App can break any running GPU task. I'm not sure
 whether the throtteling implementations of all BOINC Clients that are being
 used properly honor critical sections, nor am I that all GPU apps of all
 projects make proper use of these.


 pros:
I can't think about many


 Actually I can't think about any.

 Best,
 Bernd


 ___
 boinc_dev mailing list
 boinc_dev@ssl.berkeley.edu
 http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
 To unsubscribe, visit the above URL and
 (near bottom of page) enter your email address.
___
boinc_dev mailing list
boinc_dev@ssl.berkeley.edu
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.