On 11/30/2015 02:42 PM, appar...@raf.org wrote:
> appar...@raf.org wrote:
>
>> Seth Arnold wrote:
>>
>>> On Sat, Nov 28, 2015 at 09:02:18PM -0800, John Johansen wrote:
This check is well above the range of values I would recommend (some
where between 1-2x the number of cpus. More jobs ca
On 11/29/2015 03:11 PM, Christian Boltz wrote:
> Hello,
>
> Am Samstag, 28. November 2015 schrieb John Johansen:
>> On 11/28/2015 01:54 PM, Christian Boltz wrote:
>>> Am Samstag, 28. November 2015 schrieb John Johansen:
> ...
>>> So the parser will error out if a too big job number is given _and_
appar...@raf.org wrote:
> Seth Arnold wrote:
>
> > On Sat, Nov 28, 2015 at 09:02:18PM -0800, John Johansen wrote:
> > > This check is well above the range of values I would recommend (some
> > > where between 1-2x the number of cpus. More jobs can help with smaller
> >
> > Two times makes more s
On 11/30/2015 12:48 PM, Seth Arnold wrote:
> On Sat, Nov 28, 2015 at 09:02:18PM -0800, John Johansen wrote:
>> This check is well above the range of values I would recommend (some
>> where between 1-2x the number of cpus. More jobs can help with smaller
>
> Two times makes more sense for most CPUs
Seth Arnold wrote:
> On Sat, Nov 28, 2015 at 09:02:18PM -0800, John Johansen wrote:
> > This check is well above the range of values I would recommend (some
> > where between 1-2x the number of cpus. More jobs can help with smaller
>
> Two times makes more sense for most CPUs but eight may be mor
On 11/30/2015 12:41 PM, Seth Arnold wrote:
> On Sat, Nov 28, 2015 at 11:08:05AM -0800, John Johansen wrote:
>> +static void setup_parallel_compile(void)
>> +{
>> +/* jobs_count and paralell_max set by default, config or args */
>> +long n = sysconf(_SC_NPROCESSORS_ONLN);
>
> We should hand
On Sat, Nov 28, 2015 at 09:02:18PM -0800, John Johansen wrote:
> This check is well above the range of values I would recommend (some
> where between 1-2x the number of cpus. More jobs can help with smaller
Two times makes more sense for most CPUs but eight may be more appropriate
for e.g. POWER8
On Sat, Nov 28, 2015 at 11:08:05AM -0800, John Johansen wrote:
> +static void setup_parallel_compile(void)
> +{
> + /* jobs_count and paralell_max set by default, config or args */
> + long n = sysconf(_SC_NPROCESSORS_ONLN);
We should handle an error return here ..
> + if (jobs_count
Hello,
Am Samstag, 28. November 2015 schrieb John Johansen:
> On 11/28/2015 01:54 PM, Christian Boltz wrote:
> > Am Samstag, 28. November 2015 schrieb John Johansen:
...
> > So the parser will error out if a too big job number is given _and_
> > if there are enough profiles to load (otherwise the
On 11/28/2015 01:54 PM, Christian Boltz wrote:
> Hello,
>
> Am Samstag, 28. November 2015 schrieb John Johansen:
>> v2
>> switch to default --jobs=auto
>> check and bail on bad debug argument
>> update comments on work_sync_one()
>> check for an invalid number of maximum jobs()
>> put an upper lim
Hello,
Am Samstag, 28. November 2015 schrieb John Johansen:
> v2
> switch to default --jobs=auto
> check and bail on bad debug argument
> update comments on work_sync_one()
> check for an invalid number of maximum jobs()
> put an upper limit on the maximum number of jobs to 8*# of cpus
> +static
v2
switch to default --jobs=auto
check and bail on bad debug argument
update comments on work_sync_one()
check for an invalid number of maximum jobs()
put an upper limit on the maximum number of jobs to 8*# of cpus
---
commit ce5d771201ca36530072e328ea2ed6824e7dc859
Author: John Johansen
Date:
On 11/20/2015 05:19 PM, Seth Arnold wrote:
> On Fri, Nov 20, 2015 at 04:35:20PM -0800, John Johansen wrote:
>> Do you really think, this 1 little condition tucked away is that much more
>> maintenance? The way it is set up, it is completely transparent to the
>> rest of the code. That is we can de
On Fri, Nov 20, 2015 at 04:35:20PM -0800, John Johansen wrote:
> Do you really think, this 1 little condition tucked away is that much more
> maintenance? The way it is set up, it is completely transparent to the
> rest of the code. That is we can delete it and the rest of the code doesn't
> need
On 11/20/2015 04:01 PM, Seth Arnold wrote:
> On Sat, Oct 17, 2015 at 03:49:17PM -0700, John Johansen wrote:
>> This adds a basic support for parallel compiles. It uses a fork()/wait
>> @@ -286,8 +324,13 @@ static int process_arg(int c, char *optarg)
>> option = OPTION_ADD;
>>
On Sat, Oct 17, 2015 at 03:49:17PM -0700, John Johansen wrote:
> This adds a basic support for parallel compiles. It uses a fork()/wait
> @@ -286,8 +324,13 @@ static int process_arg(int c, char *optarg)
> option = OPTION_ADD;
> break;
> case 'd':
> - de
On Fri, Oct 23, 2015 at 12:47:03PM -0700, John Johansen wrote:
> So one thing I wanted to ask here was whether or not we should default this
> to --jobs=auto instead of 1
We may need some tweaking to the algorithm to get good results on phones,
desktops, and huge servers, but --jobs=auto is defini
On 10/24/2015 06:05 AM, Christian Boltz wrote:
> Hello,
>
> Am Freitag, 23. Oktober 2015 schrieb John Johansen:
>> So one thing I wanted to ask here was whether or not we should default
>> this to --jobs=auto instead of 1
>>
>> 1 is safe in that it is the current behavior, but I think we want this
Hello,
Am Freitag, 23. Oktober 2015 schrieb John Johansen:
> So one thing I wanted to ask here was whether or not we should default
> this to --jobs=auto instead of 1
>
> 1 is safe in that it is the current behavior, but I think we want this
> on by default
So we can choose between telling our u
On 10/17/2015 03:49 PM, John Johansen wrote:
> This adds a basic support for parallel compiles. It uses a fork()/wait
> model due to the parsers current dependence on global variables and
> structures. It has been setup in a similar manner to how cilk handles
> multithreading to make it easy to por
On 10/23/2015 05:06 AM, intrigeri wrote:
> John Johansen wrote (17 Oct 2015 22:49:17 GMT) :
>> -j or -jauto or --jobs=auto sets the max jobs to the # of cpus
>
> Thanks! I can't wait to see this merged upstream (sorry I have no time
> to test it here). This will be immensely helpful for r
John Johansen wrote (17 Oct 2015 22:49:17 GMT) :
> -j or -jauto or --jobs=auto sets the max jobs to the # of cpus
Thanks! I can't wait to see this merged upstream (sorry I have no time
to test it here). This will be immensely helpful for replacing shell
code when migrating stuff to syste
This adds a basic support for parallel compiles. It uses a fork()/wait
model due to the parsers current dependence on global variables and
structures. It has been setup in a similar manner to how cilk handles
multithreading to make it easy to port to a managed thread model once
the parser removes t
23 matches
Mail list logo