Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
S
Slurm
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Package Registry
Model registry
Operate
Environments
Terraform modules
Monitor
Incidents
Service Desk
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Terms and privacy
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
tud-zih-energy
Slurm
Commits
b9c4edda
Commit
b9c4edda
authored
13 years ago
by
Danny Auble
Browse files
Options
Downloads
Patches
Plain Diff
BLUEGENE - breakout setting up wires more like 2.2 so we don't have to run
through that extra logic when starting up from smap.
parent
f2d05f81
No related branches found
No related tags found
No related merge requests found
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
src/plugins/select/bluegene/ba_common.c
+28
-82
28 additions, 82 deletions
src/plugins/select/bluegene/ba_common.c
src/plugins/select/bluegene/ba_common.h
+3
-0
3 additions, 0 deletions
src/plugins/select/bluegene/ba_common.h
with
31 additions
and
82 deletions
src/plugins/select/bluegene/ba_common.c
+
28
−
82
View file @
b9c4edda
...
...
@@ -560,58 +560,6 @@ static ba_mp_t *_internal_loc2ba_mp(int level, uint16_t *coords,
}
#endif
/*
* set values of every grid point (used in smap)
*/
static
void
_init_grid
(
node_info_msg_t
*
node_info_ptr
)
{
int
i
=
0
,
j
,
x
;
if
(
!
node_info_ptr
)
return
;
for
(
j
=
0
;
j
<
(
int
)
node_info_ptr
->
record_count
;
j
++
)
{
uint16_t
coord
[
cluster_dims
];
node_info_t
*
node_ptr
=
&
node_info_ptr
->
node_array
[
j
];
select_nodeinfo_t
*
nodeinfo
=
NULL
;
if
(
!
node_ptr
->
name
)
continue
;
memset
(
coord
,
0
,
sizeof
(
coord
));
if
(
cluster_dims
==
1
)
{
coord
[
0
]
=
j
;
}
else
{
if
((
i
=
strlen
(
node_ptr
->
name
))
<
cluster_dims
)
continue
;
for
(
x
=
0
;
x
<
cluster_dims
;
x
++
)
coord
[
x
]
=
select_char2coord
(
node_ptr
->
name
[
i
-
(
cluster_dims
+
x
)]);
}
for
(
x
=
0
;
x
<
cluster_dims
;
x
++
)
if
((
int16_t
)
coord
[
x
]
<
0
)
break
;
if
(
x
<
cluster_dims
)
continue
;
xassert
(
node_ptr
->
select_nodeinfo
);
nodeinfo
=
node_ptr
->
select_nodeinfo
->
data
;
xassert
(
nodeinfo
);
nodeinfo
->
ba_mp
=
coord2ba_mp
(
coord
);
if
(
!
nodeinfo
->
ba_mp
)
continue
;
nodeinfo
->
ba_mp
->
index
=
j
;
if
(
IS_NODE_DOWN
(
node_ptr
)
||
IS_NODE_DRAIN
(
node_ptr
))
{
ba_update_mp_state
(
nodeinfo
->
ba_mp
,
node_ptr
->
node_state
);
}
nodeinfo
->
ba_mp
->
state
=
node_ptr
->
node_state
;
//nodeinfo->ba_mp = ba_mp;
}
}
/**
* Initialize internal structures by either reading previous block
* configurations from a file or by running the graph solver.
...
...
@@ -630,7 +578,6 @@ extern void ba_init(node_info_msg_t *node_info_ptr, bool sanity_check)
slurm_conf_node_t
**
ptr_array
;
int
coords
[
HIGHEST_DIMENSIONS
];
char
*
p
=
'\0'
;
int
num_mps
=
0
;
int
real_dims
[
HIGHEST_DIMENSIONS
];
char
dim_str
[
HIGHEST_DIMENSIONS
+
1
];
...
...
@@ -690,7 +637,6 @@ extern void ba_init(node_info_msg_t *node_info_ptr, bool sanity_check)
}
hostlist_parse_int_to_array
(
number
,
coords
,
cluster_dims
,
cluster_base
);
memcpy
(
DIM_SIZE
,
coords
,
sizeof
(
DIM_SIZE
));
}
for
(
j
=
0
;
j
<
cluster_dims
;
j
++
)
{
...
...
@@ -698,7 +644,6 @@ extern void ba_init(node_info_msg_t *node_info_ptr, bool sanity_check)
/* this will probably be reset below */
real_dims
[
j
]
=
DIM_SIZE
[
j
];
}
num_mps
=
node_info_ptr
->
record_count
;
}
node_info_error:
for
(
j
=
0
;
j
<
cluster_dims
;
j
++
)
...
...
@@ -754,12 +699,10 @@ node_info_error:
info
(
"are you sure you only have 1 midplane? %s"
,
ptr_array
[
0
]
->
nodenames
);
num_mps
=
1
;
for
(
j
=
0
;
j
<
cluster_dims
;
j
++
)
{
DIM_SIZE
[
j
]
++
;
/* this will probably be reset below */
real_dims
[
j
]
=
DIM_SIZE
[
j
];
num_mps
*=
DIM_SIZE
[
j
];
}
}
...
...
@@ -800,25 +743,10 @@ setup_done:
debug
(
"We are using %s of the system."
,
dim_str
);
}
if
(
bg_recover
!=
NOT_FROM_CONTROLLER
)
{
if
(
!
num_mps
)
{
num_mps
=
1
;
for
(
i
=
0
;
i
<
cluster_dims
;
i
++
)
num_mps
*=
DIM_SIZE
[
i
];
}
ba_main_mp_bitmap
=
bit_alloc
(
num_mps
);
ba_create_system
();
bridge_setup_system
();
_init_grid
(
node_info_ptr
);
for
(
i
=
1
;
i
<=
LONGEST_BGQ_DIM_LEN
;
i
++
)
_build_geo_bitmap_arrays
(
i
);
}
ba_initialized
=
true
;
if
(
bg_recover
!=
NOT_FROM_CONTROLLER
)
ba_setup_wires
();
}
...
...
@@ -844,6 +772,29 @@ extern void ba_fini(void)
// debug3("pa system destroyed");
}
extern
void
ba_setup_wires
(
void
)
{
int
num_mps
,
i
;
static
bool
wires_setup
=
0
;
if
(
!
ba_initialized
||
wires_setup
)
return
;
wires_setup
=
1
;
num_mps
=
1
;
for
(
i
=
0
;
i
<
cluster_dims
;
i
++
)
num_mps
*=
DIM_SIZE
[
i
];
ba_main_mp_bitmap
=
bit_alloc
(
num_mps
);
ba_create_system
();
bridge_setup_system
();
for
(
i
=
1
;
i
<=
LONGEST_BGQ_DIM_LEN
;
i
++
)
_build_geo_bitmap_arrays
(
i
);
}
extern
void
destroy_ba_mp
(
void
*
ptr
)
{
ba_mp_t
*
ba_mp
=
(
ba_mp_t
*
)
ptr
;
...
...
@@ -1522,18 +1473,13 @@ extern int validate_coord(uint16_t *coord)
int
dim
,
i
;
char
coord_str
[
cluster_dims
+
1
];
char
dim_str
[
cluster_dims
+
1
];
static
int
*
dims
=
NULL
;
if
(
!
dims
)
dims
=
select_g_ba_get_dims
();
xassert
(
!
dims
);
for
(
dim
=
0
;
dim
<
cluster_dims
;
dim
++
)
{
if
(
coord
[
dim
]
>=
dims
[
dim
])
{
if
(
coord
[
dim
]
>=
DIM_SIZE
[
dim
])
{
if
(
ba_debug_flags
&
DEBUG_FLAG_BG_ALGO_DEEP
)
{
for
(
i
=
0
;
i
<
cluster_dims
;
i
++
)
{
coord_str
[
i
]
=
alpha_num
[
coord
[
i
]];
dim_str
[
i
]
=
alpha_num
[
dims
[
i
]];
dim_str
[
i
]
=
alpha_num
[
DIM_SIZE
[
i
]];
}
coord_str
[
i
]
=
'\0'
;
dim_str
[
i
]
=
'\0'
;
...
...
This diff is collapsed.
Click to expand it.
src/plugins/select/bluegene/ba_common.h
+
3
−
0
View file @
b9c4edda
...
...
@@ -203,6 +203,9 @@ extern void ba_init(node_info_msg_t *node_info_ptr, bool load_bridge);
*/
extern
void
ba_fini
(
void
);
/* setup the wires for the system */
extern
void
ba_setup_wires
(
void
);
extern
void
destroy_ba_mp
(
void
*
ptr
);
extern
void
pack_ba_mp
(
ba_mp_t
*
ba_mp
,
Buf
buffer
,
uint16_t
protocol_version
);
extern
int
unpack_ba_mp
(
ba_mp_t
**
ba_mp_pptr
,
Buf
buffer
,
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment