1 |
Hi all, |
2 |
|
3 |
*Summary:* |
4 |
|
5 |
Last week's plan: |
6 |
1) Start working towards multiplexed & multilib installs for mpi-select. |
7 |
2) Start testing new mpi-select against mpich and openmpi implementations, |
8 |
seeing if both implementations can be installed against something like hpl. |
9 |
3) Refine wiki/documentation for the project |
10 |
|
11 |
mpi-select, for multilib-like multiplexing, should be able to be wrapped up |
12 |
in the next few weeks. The eclass will make it easier for users to pick |
13 |
multiple implementations & their versions to build against other software. |
14 |
|
15 |
Next Week's Plan: |
16 |
1) Loop over commands in the build process for multiple implementations. |
17 |
2) Add more mpi implementation's versions to the overlay, testing with |
18 |
mpi-providers |
19 |
3) Adjust libraries/binaries installations globally. This will later be for |
20 |
mpi-dependent things only |
21 |
|
22 |
*Day-by-day-breakdown:* |
23 |
*June 19th:* |
24 |
*Today's Plans:* |
25 |
|
26 |
Clean up documentation (README, metadata.xml's) and migrate wiki page. |
27 |
Work in progress. Progress has been made, but I will need to continue |
28 |
improving this. |
29 |
|
30 |
Test and add more implementation versions to the overlay. |
31 |
Completed. Still working on the latest openmpi version. |
32 |
|
33 |
Fix find statement in mpi-providers. |
34 |
Work in progress. I'm testing this using pipes/xargs with cp, will continue |
35 |
to work on this. |
36 |
|
37 |
Start committing code to emulate multilib multiplexing with MPI installs. |
38 |
|
39 |
Incomplete. This will be my primary objective Tuesday. |
40 |
|
41 |
*Tuesday's Plans:* |
42 |
|
43 |
Implement multilib-like exported functions (configure, compile, install). |
44 |
|
45 |
|
46 |
Continue work on mpi-providers find statement. |
47 |
|
48 |
|
49 |
Continue building openmpi-2.1.1, add to overlay when tested. |
50 |
|
51 |
|
52 |
*June 20th:* |
53 |
*Tuesday's Plans* |
54 |
|
55 |
Implement multilib-like exported functions (configure, compile, install). |
56 |
|
57 |
Work in progress. I would like to get this to the point where I can write a |
58 |
detailed email tonight. |
59 |
|
60 |
Continue work on mpi-providers find statement. |
61 |
Work in progress. Along with pipe/xargs testing, I'm looking into command |
62 |
substitution with proper error handling. |
63 |
|
64 |
|
65 |
Continue building openmpi-2.1.1, add to overlay when tested. |
66 |
|
67 |
Work in progress. I'm running into Fortran build issues in the |
68 |
configuration phase. If this issue persists, I will be including details in |
69 |
my email tonight. |
70 |
|
71 |
|
72 |
*Wednesday's Plans:* |
73 |
|
74 |
Integrate more multilib-like properties in mpi-select to discuss further |
75 |
with mentors. |
76 |
|
77 |
Resolve Fortran issues in openmpi-2.1.1 build. If not, include in email. |
78 |
|
79 |
Email mentors regarding details, progress, and help needed. |
80 |
|
81 |
*June 21st:* |
82 |
*Wednesday's Progress:* |
83 |
|
84 |
With homework, class, and some power outages, I essentially took the day |
85 |
off other than some housekeeping in the repository. I will look to make up |
86 |
my hours Thursday, Friday, and Saturday, as I do not have class any of |
87 |
these days. |
88 |
|
89 |
|
90 |
*Thursday's Goals:* |
91 |
|
92 |
Integrate more multilib-like properties in mpi-select to discuss further |
93 |
with mentors. |
94 |
|
95 |
Resolve Fortran issues in openmpi-2.1.1 build. If not, include in email. |
96 |
|
97 |
Email mentors regarding details, progress, and help needed. |
98 |
|
99 |
*June 22nd:* |
100 |
|
101 |
*Thursday's Goals:* |
102 |
|
103 |
Integrate more multilib-like properties in mpi-select to discuss further |
104 |
with mentors. |
105 |
Work in progress. I currently have a general framework setup, and am |
106 |
working towards a proper implementation. |
107 |
|
108 |
Resolve Fortran issues in openmpi-2.1.1 build. If not, include in email. |
109 |
Awaiting response from an email I very recently sent. |
110 |
|
111 |
|
112 |
Email mentors regarding details, progress, and help needed. |
113 |
Completed. This, along with our weekly meeting, will likely continue to |
114 |
clarify anything that I still have questions about. |
115 |
|
116 |
|
117 |
*Friday's Goals:* |
118 |
|
119 |
Resolve issues detailed in email correspondence, potentially through weekly |
120 |
meeting. |
121 |
|
122 |
Continue mpi-select src_* implementation, get mpi_foreach_implementation to |
123 |
properly loop identical commands over different implementations. |
124 |
|
125 |
*June 23rd:* |
126 |
|
127 |
*Friday's Goals:* |
128 |
|
129 |
Resolve issues detailed in email correspondence, potentially through weekly |
130 |
meeting. |
131 |
Completed. A lot of my questions were answered via email, and resolutions |
132 |
were created. |
133 |
|
134 |
Continue mpi-select src_* implementation, get mpi_foreach_implementation to |
135 |
properly loop identical commands over different implementations. |
136 |
|
137 |
Work in progress. This will be a big part of next week's work. |
138 |
|
139 |
|
140 |
*Monday's Goals:* |
141 |
|
142 |
Complete loop to repeat commands per selected implementation. |
143 |
|
144 |
Implement src_* per abi, per mpi implementation. |
145 |
|
146 |
|
147 |
|
148 |
|
149 |
Thanks, |
150 |
Michael |