1 |
Hi, |
2 |
|
3 |
*Summary:* |
4 |
1) Fix bugs from build.log output- general eclass housekeeping. |
5 |
2) Expand upon implementation-specifc src_* phases while using multilib |
6 |
functions to iterate through implementations. |
7 |
3) Discuss further design & next steps (installing mpi-dependent libs to |
8 |
certain location, etc) |
9 |
|
10 |
A lot of this week entailed fixing broken bits of code to avoid any |
11 |
errors/warnings. There is still some work to be done in this regard, |
12 |
particularly with the exported functions. |
13 |
|
14 |
Next week's plans: |
15 |
1) Continue work on exported functions and respective environment variables. |
16 |
2) Start to objdump (among other things) to try and link against only |
17 |
MPI-dependent libs. |
18 |
3) Continue ironing out bugs and errors from build.log. |
19 |
|
20 |
*Day-by-day breakdown:* |
21 |
|
22 |
*Monday Plans:* |
23 |
|
24 |
Sort out issues I've been having with has_version in mpi_dependencies and |
25 |
function returns in get_mpicc. Via email if not available by IRC. |
26 |
|
27 |
Work in progress- though I am feeling a bit stuck. Even calling the most |
28 |
basic of functions (such as get_mpicc) fails to return a simple string. |
29 |
Perhaps there is something large that I am missing. |
30 |
|
31 |
|
32 |
Revisit mpi_src_* to start tailoring builds to their respective |
33 |
implementations |
34 |
|
35 |
As per our emails, building hpl directly from source seems like a helpful |
36 |
task to get a better understanding of what is needed at each phase. I will |
37 |
be doing this first thing on Tuesday. |
38 |
|
39 |
|
40 |
*Tuesday's Plans:* |
41 |
|
42 |
Build hpl directly from source, document needed environment variable |
43 |
changes. |
44 |
I need to continue tinkering with this, particularly with the TUNING |
45 |
options (using hpl-2.1.tar.gz from ebuild) |
46 |
|
47 |
Continue trying to debug current issues with eclass function calls. |
48 |
Work in progress. The feedback I've gotten has been helpful, though I may |
49 |
come with more questions tomorrow. |
50 |
|
51 |
|
52 |
*Wednesday's Plans:* |
53 |
|
54 |
Continue work on building hpl from source & tuning to specifics of my |
55 |
machine, reference eselect-mpi for env variables. |
56 |
|
57 |
Work in progress. I believe I have a much better understanding of the build |
58 |
process, and will work on proper environment variable setting tomorrow in |
59 |
the eclass. |
60 |
|
61 |
|
62 |
Re-add mpi_foreach_implementation from before and test until it properly |
63 |
iterates over MPI_TARGETS. |
64 |
|
65 |
Work in progress. This will need much more testing as it is a bit of a |
66 |
transition going from build-in multilib functions to something a bit more |
67 |
custom. |
68 |
|
69 |
|
70 |
*Thursday's Plans:* |
71 |
|
72 |
Set proper environment variables in functions like src_prepare and |
73 |
mpi_wrapper. |
74 |
Incomplete. I've been working on my final project for my computer science |
75 |
class, and did not have time to work on this today. The project is due at |
76 |
noon on Friday and is my last assignment for that class. |
77 |
|
78 |
Continue development on mpi_foreach_implementation, add loop for all "${@}" |
79 |
as well as for all MPI_TARGETS. |
80 |
Needs more testing, but I believe the loop implemented will account for the |
81 |
implementations as well as the multilib-ABI information. |
82 |
|
83 |
|
84 |
*Friday's Plans:* |
85 |
|
86 |
Test the mpi_src_* functions with new mpi_foreach loop. |
87 |
|
88 |
This is still broken and needs revising. The new custom function |
89 |
(mpi_foreach) will need some updating, particularly in the inner function |
90 |
where commands are executed. |
91 |
|
92 |
|
93 |
Set proper environment variables in functions like src_prepare and |
94 |
mpi_wrapper. |
95 |
|
96 |
Work in progress. This work will continue on Monday. |
97 |
|
98 |
|
99 |
Thanks, |
100 |
Michael |