Commit 75448d67 authored by MOMMESSIN Clement's avatar MOMMESSIN Clement Committed by Millian Poquet

[doc.s] remove (msg_' profile type prefix and rename types

parent 4248730b
......@@ -33,6 +33,13 @@ Batsim's public API includes:
``--enable-time-sharing-on-compute`` and
``--disable-time-sharing-on-storage``, as resource roles have been introduced.
### Changed (**breaks workload format**)
- Profile types of parallel jobs has been renamed:
- ``msg_par`` into ``parallel``
- ``msg_par_hg`` into ``parallel_homogeneous``
- ``msg_par_hg_tot`` into ``parallel_homogeneous_total``
- ``msg_par_hg_pfs`` into ``parallel_homogeneous_pfs``
### Added (new command-line options)
- New ``--sg-cfg`` option, that allows to set SimGrid configuration options.
- New ``--dump-execution-context`` option, that dumps the command execution
......
......@@ -67,7 +67,7 @@ strings:
``` json
{
"type": "msg_par",
"type": "parallel",
"cpu": [5e6,5e6,5e6,5e6],
"com": [5e6,5e6,5e6,5e6,
5e6,5e6,5e6,5e6,
......@@ -78,7 +78,7 @@ strings:
``` json
{
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 10e6,
"com": 1e6
}
......
......@@ -28,7 +28,7 @@ Waiting for 20.20 seconds.
}
```
## MSG parallel task
## Parallel task
This profile correspond to a parallel task executed simultaneously on each node
allocated to the job.
......@@ -46,7 +46,7 @@ allocated to the job.
```json
{
"type": "msg_par",
"type": "parallel",
"cpu": [5e6, 0, 0, 0],
"com": [5e6, 0, 0, 0,
5e6,5e6, 0, 0,
......@@ -55,7 +55,7 @@ allocated to the job.
}
```
## MSG parallel homogeneous task
## Parallel homogeneous task
This model is a convenient way to generate homogeneous task computation and
communication. The loopback communication is set to 0.
......@@ -68,13 +68,13 @@ communication. The loopback communication is set to 0.
```json
{
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 10e6,
"com": 1e6
}
```
## MSG parallel homogeneous task with total amount
## Parallel homogeneous task with total amount
This model is a convenient way to generate homogeneous task computation and
communication by giving the total amount work to be done. The loopback
......@@ -92,7 +92,7 @@ any number of resources while conserving the same amount of work to do.
```json
{
"type": "msg_par_hg_tot",
"type": "parallel_homogeneous_total",
"cpu": 10e6,
"com": 1e6
}
......@@ -119,7 +119,7 @@ This job profile is a list of profiles to be executed in a sequence.
```
## MSG homogeneous IO to/from a PFS storage (Parallel File System)
## Homogeneous IO to/from a PFS storage (Parallel File System)
Represents an IO transfer between all the nodes of a job's allocation and a
centralized storage tier. The storage tier is represented by one node.
......@@ -134,14 +134,14 @@ centralized storage tier. The storage tier is represented by one node.
```json
{
"type": "msg_par_hg_pfs",
"type": "parallel_homogeneous_pfs",
"bytes_to_read": 10e5,
"bytes_to_write": 10e5,
"storage": "nfs"
}
```
## MSG IO staging between two storage tiers
## IO staging between two storage tiers
This profile represents an IO transfer between two storage tiers.
......
......@@ -360,7 +360,7 @@ options to the scheduler.
"profiles": {
"26dceb":{
"simple": {
"type": "msg_par",
"type": "parallel",
"cpu": [5e6, 0, 0, 0],
"com": [5e6, 0, 0, 0,
5e6,5e6, 0, 0,
......@@ -368,17 +368,17 @@ options to the scheduler.
5e6,5e6,5e6, 0]
},
"homogeneous": {
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 10e6,
"com": 1e6
},
"homogeneous_no_cpu": {
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 0,
"com": 1e6
},
"homogeneous_no_com": {
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 2e5,
"com": 0
},
......@@ -392,7 +392,7 @@ options to the scheduler.
"delay": 20.20
},
"homogeneous_total": {
"type": "msg_par_hg_tot",
"type": "parallel_homogeneous_total",
"cpu": 10e6,
"com": 1e6
}
......@@ -665,7 +665,7 @@ for resource id 3).
For certain job profiles that involve storage you may need to define a
``storage_mapping`` between the storage label defined in the job profile
definition and the storage resource id on the platform. For example, the job
profile of type ``msg_par_hg_pfs`` contains this field ``"storage": "pfs"``.
profile of type ``parallel_homogeneous_pfs`` contains this field ``"storage": "pfs"``.
In order to select what is the resource that corresponds to the
``"pfs"`` storage, you should provide a mapping for this label:
``"storage_mapping": { "pfs": 2 }``. If no mapping is provided, Batsim will guess
......@@ -674,7 +674,7 @@ the storage mapping only if one storage resource is provided on the platform.
Another optional field is ``additional_io_job`` that permits the scheduler
to add a job, that represents the IO traffic, dynamically at execution
time. This dynamicity is necessary when the IO traffic depends on the job
allocation. It only works for MSG based job profile types for the additional IO
allocation. It only works for parallel task based job profile types for the additional IO
job and the job itself. The given IO job will be merged to the actual job before
its execution. The additional job allocation may be different from the job
allocation itself, for example when some IO nodes are involved.
......@@ -698,7 +698,7 @@ allocation itself, for example when some IO nodes are involved.
"alloc": "2-3 5-6",
"profile_name": "my_io_job",
"profile": {
"type": "msg_par",
"type": "parallel",
"cpu": 0,
"com": [0 ,5e6,5e6,5e6,
5e6,0 ,5e6,0 ,
......
......@@ -88,7 +88,7 @@ resources. The profile of the job called is defined by a Json object,
.. code:: js
"job1_profile": {
"type": "msg_par_hg_tot",
"type": "parallel_homogeneous_total",
"cpu": 1e10,
"com": 1e7,
"io_reads": 8e8,
......@@ -100,7 +100,7 @@ resources. The profile of the job called is defined by a Json object,
that each node will read half of the data from only one centralized
storage (located at the resource id 10), but the nodes of the
application will write on local their local disks (id 2 and 3). A new
profile of type ``msg_par`` is generated an submitted to batsim with
profile of type ``parallel`` is generated an submitted to batsim with
the name ``io_for_this_alloc_on_job1``. The kind job profile is
composed of a computation matrix that may reflect I/O related
computation (here set to 0), and a communication matrix that
......@@ -114,7 +114,7 @@ resources. The profile of the job called is defined by a Json object,
.. code:: js
"io_for_this_alloc_on_job1": {
"type": "msg_par",
"type": "parallel",
"cpu": [0 ,0 ,0 ,0 ,0]
"com": [0 ,0 ,1e8,0 ,0
0 ,0 ,0 ,1e8,0
......
......@@ -44,6 +44,14 @@ Changed (**breaks command-line interface**)
``--enable-time-sharing-on-compute`` and ``--disable-time-sharing-on-storage``,
as resource roles have been introduced.
Changed (**breaks workload format**)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Profile types of parallel jobs has been renamed:
- ``msg_par`` into ``parallel``
- ``msg_par_hg`` into ``parallel_homogeneous``
- ``msg_par_hg_tot`` into ``parallel_homogeneous_total``
- ``msg_par_hg_pfs`` into ``parallel_homogeneous_pfs``
Added (new command-line options)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
......
......@@ -96,7 +96,7 @@ Here is an example of a parallel task that can be used by any job requesting 4 m
.. code:: json
{
"type": "msg_par",
"type": "parallel",
"cpu": [5e6, 0, 0, 0],
"com": [5e6, 0, 0, 0,
5e6,5e6, 0, 0,
......@@ -123,7 +123,7 @@ This profile type is a convenient way to generate an homogeneous `Parallel task`
.. code:: json
{
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 10e6,
"com": 1e6
}
......@@ -147,7 +147,7 @@ It allows such profiles to be used with any number of resources while conserving
.. code:: json
{
"type": "msg_par_hg_tot",
"type": "parallel_homogeneous_total",
"cpu": 10e6,
"com": 1e6
}
......@@ -183,7 +183,7 @@ centralized storage tier. The storage tier is represented by one host of the :re
.. code:: json
{
"type": "msg_par_hg_pfs",
"type": "parallel_homogeneous_pfs",
"bytes_to_read": 10e5,
"bytes_to_write": 10e5,
"storage": "nfs"
......
......@@ -365,7 +365,7 @@ Batsim configuration is sent through the ``config`` object (in ``data``). Custom
"profiles": {
"26dceb":{
"simple": {
"type": "msg_par",
"type": "parallel",
"cpu": [5e6, 0, 0, 0],
"com": [5e6, 0, 0, 0,
5e6,5e6, 0, 0,
......@@ -373,17 +373,17 @@ Batsim configuration is sent through the ``config`` object (in ``data``). Custom
5e6,5e6,5e6, 0]
},
"homogeneous": {
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 10e6,
"com": 1e6
},
"homogeneous_no_cpu": {
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 0,
"com": 1e6
},
"homogeneous_no_com": {
"type": "msg_par_hg",
"type": "parallel_homogeneous",
"cpu": 2e5,
"com": 0
},
......@@ -397,7 +397,7 @@ Batsim configuration is sent through the ``config`` object (in ``data``). Custom
"delay": 20.20
},
"homogeneous_total": {
"type": "msg_par_hg_tot",
"type": "parallel_homogeneous_total",
"cpu": 10e6,
"com": 1e6
}
......@@ -678,7 +678,7 @@ second machine (1, which stands for resource id 3).
For certain job profiles that involve storage you may need to define a
``storage_mapping`` between the storage label defined in the job profile
definition and the storage resource id on the platform. For example, the
job profile of type ``msg_par_hg_pfs_tiers`` contains this field
job profile of type ``parallel_homogeneous_pfs`` contains this field
``"storage": "pfs"``. In order to select what is the resource that
corresponds to the ``"pfs"`` storage, you should provide a mapping for
this label: ``"storage_mapping": { "pfs": 2 }``. If no mapping is
......@@ -688,7 +688,7 @@ resource is provided on the platform.
Another optional field is ``additional_io_job`` that permits the
scheduler to add a job, that represents the IO traffic, dynamically at
execution time. This dynamicity is necessary when the IO traffic depends
on the job allocation. It only works for MSG based job profile types for
on the job allocation. It only works for parallel task based job profile types for
the additional IO job and the job itself. The given IO job will be
merged to the actual job before its execution. The additional job
allocation may be different from the job allocation itself, for example
......@@ -713,7 +713,7 @@ when some IO nodes are involved.
"alloc": "2-3 5-6",
"profile_name": "my_io_job",
"profile": {
"type": "msg_par",
"type": "parallel",
"cpu": 0,
"com": [0 ,5e6,5e6,5e6,
5e6,0 ,5e6,0 ,
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment