tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
docs: mic: convert docs to ReST and rename to *.rst
Convert Intel Many Integrated Core architecture docs to ReST.
The conversion is trivial: just add title and literal block
markups, and adjust some identation.
At its new index.rst, let's add a :orphan: while this is not linked to
the main index.rst file, in order to avoid build warnings.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
authored by
Mauro Carvalho Chehab
and committed by
Jonathan Corbet
09bbf055
d67297ad
+57
-25
+18
Documentation/mic/index.rst
+18
Documentation/mic/index.rst
···
1
+
:orphan:
2
+
3
+
=============================================
4
+
Intel Many Integrated Core (MIC) architecture
5
+
=============================================
6
+
7
+
.. toctree::
8
+
:maxdepth: 1
9
+
10
+
mic_overview
11
+
scif_overview
12
+
13
+
.. only:: subproject and html
14
+
15
+
Indices
16
+
=======
17
+
18
+
* :ref:`genindex`
+5
-1
Documentation/mic/mic_overview.txt
Documentation/mic/mic_overview.rst
+5
-1
Documentation/mic/mic_overview.txt
Documentation/mic/mic_overview.rst
···
1
+
======================================================
2
+
Intel Many Integrated Core (MIC) architecture overview
3
+
======================================================
4
+
1
5
An Intel MIC X100 device is a PCIe form factor add-in coprocessor
2
6
card based on the Intel Many Integrated Core (MIC) architecture
3
7
that runs a Linux OS. It is a PCIe endpoint in a platform and therefore
···
49
45
virtio backends are situated on the host rather than the card given better
50
46
single threaded performance for the host compared to MIC, the ability of
51
47
the host to initiate DMA's to/from the card using the MIC DMA engine and
52
-
the fact that the virtio block storage backend can only be on the host.
48
+
the fact that the virtio block storage backend can only be on the host::
53
49
54
50
+----------+ | +----------+
55
51
| Card OS | | | Host OS |
+34
-24
Documentation/mic/scif_overview.txt
Documentation/mic/scif_overview.rst
+34
-24
Documentation/mic/scif_overview.txt
Documentation/mic/scif_overview.rst
···
1
+
========================================
2
+
Symmetric Communication Interface (SCIF)
3
+
========================================
4
+
1
5
The Symmetric Communication Interface (SCIF (pronounced as skiff)) is a low
2
6
level communications API across PCIe currently implemented for MIC. Currently
3
7
SCIF provides inter-node communication within a single host platform, where a
···
12
8
abilities of the hardware. SCIF has been used to implement an offload compiler
13
9
runtime and OFED support for MPI implementations for MIC coprocessors.
14
10
15
-
==== SCIF API Components ====
11
+
SCIF API Components
12
+
===================
13
+
16
14
The SCIF API has the following parts:
15
+
17
16
1. Connection establishment using a client server model
18
17
2. Byte stream messaging intended for short messages
19
18
3. Node enumeration to determine online nodes
···
35
28
DMA, CPU copies or remote memory mapping via mmap. SCIF supports both user and
36
29
kernel mode clients which are functionally equivalent.
37
30
38
-
==== SCIF Performance for MIC ====
31
+
SCIF Performance for MIC
32
+
========================
33
+
39
34
DMA bandwidth comparison between the TCP (over ethernet over PCIe) stack versus
40
-
SCIF shows the performance advantages of SCIF for HPC applications and runtimes.
35
+
SCIF shows the performance advantages of SCIF for HPC applications and
36
+
runtimes::
41
37
42
38
Comparison of TCP and SCIF based BW
43
39
···
76
66
is distributed @ https://software.intel.com/en-us/mic-developer
77
67
78
68
Here is some pseudo code for an example of how two applications on two PCIe
79
-
nodes would typically use the SCIF API:
69
+
nodes would typically use the SCIF API::
80
70
81
-
Process A (on node A) Process B (on node B)
71
+
Process A (on node A) Process B (on node B)
82
72
83
-
/* get online node information */
84
-
scif_get_node_ids(..) scif_get_node_ids(..)
85
-
scif_open(..) scif_open(..)
86
-
scif_bind(..) scif_bind(..)
87
-
scif_listen(..)
88
-
scif_accept(..) scif_connect(..)
89
-
/* SCIF connection established */
73
+
/* get online node information */
74
+
scif_get_node_ids(..) scif_get_node_ids(..)
75
+
scif_open(..) scif_open(..)
76
+
scif_bind(..) scif_bind(..)
77
+
scif_listen(..)
78
+
scif_accept(..) scif_connect(..)
79
+
/* SCIF connection established */
90
80
91
-
/* Send and receive short messages */
92
-
scif_send(..)/scif_recv(..) scif_send(..)/scif_recv(..)
81
+
/* Send and receive short messages */
82
+
scif_send(..)/scif_recv(..) scif_send(..)/scif_recv(..)
93
83
94
-
/* Register memory */
95
-
scif_register(..) scif_register(..)
84
+
/* Register memory */
85
+
scif_register(..) scif_register(..)
96
86
97
-
/* RDMA */
98
-
scif_readfrom(..)/scif_writeto(..) scif_readfrom(..)/scif_writeto(..)
87
+
/* RDMA */
88
+
scif_readfrom(..)/scif_writeto(..) scif_readfrom(..)/scif_writeto(..)
99
89
100
-
/* Fence DMAs */
101
-
scif_fence_signal(..) scif_fence_signal(..)
90
+
/* Fence DMAs */
91
+
scif_fence_signal(..) scif_fence_signal(..)
102
92
103
-
mmap(..) mmap(..)
93
+
mmap(..) mmap(..)
104
94
105
-
/* Access remote registered memory */
95
+
/* Access remote registered memory */
106
96
107
-
/* Close the endpoints */
108
-
scif_close(..) scif_close(..)
97
+
/* Close the endpoints */
98
+
scif_close(..) scif_close(..)