(no commit message)
[libreriscv.git] / meetings / sync_up / sync_up_2023-11-21.mdwn
1 # Tuesday 21st November 17:00 UTC
2
3 * Previous week's notes: [[meetings/sync_up/sync_up_2023-11-14]]
4 * Next day's notes: [[meetings/sync_up/sync_up_2023-11-22]]
5 * Next week's notes: [[meetings/sync_up/sync_up_2023-11-28]]
6
7 # Sadoon
8
9 - Learning SVP64 from Luke, setvl (Horizontal-First), svindex.
10 - Luke suggested to look at Chacha20 example for Vertical-First
11 (but not required).
12 - Look at Luke's examples on Poly1305 bugreport
13
14
15 # Jacob
16
17 - bug #1169
18
19 - Helped Andrey with bug #1210 (see info in Andrey's section).
20 - Implemented a solution,
21 [commit diff](https://git.libre-soc.org/?p=openpower-isa.git;a=commitdiff;h=d9544764b1710f3807a9c0685d150a665f70b9a2)
22 - Solution checked by Luke and Andrey, pushed to master.
23
24 ## Luke
25
26
27 ## David
28
29 - Prioritise keeping to Nlnet grant deadlines, **absolute priority**.
30
31 # Andrey
32
33 - Took Monday off as a break.
34
35 - Luke and I in conversation with Michiel to finalise the ongoing grant MoU.
36 *(The MoU was signed the following week.)*
37
38 - Raised a bug report relating to ISACaller not taking a sv.bc branch in
39 Vertical-First mode,
40 [bug #1210](https://bugs.libre-soc.org/show_bug.cgi?id=1210).
41 - After checking with Python pdb, it was spotted that ISACaller correctly
42 computes NIA (Next Instruction Address) based on branch condition,
43 but this correct NIA is overwritten by extraneous NIA update (which defaults
44 to to the next instruction address).
45 - After Jacob pushed a fix in a branch, I ran the test and the sv.bc branch
46 was taken correctly.
47
48 - Continued to work on Libre-SOC documentation included in [[HDL_workflow]]
49 and [[HDL_workflow/libresoc_bug_process]]. See
50 [bug #1126](https://bugs.libre-soc.org/show_bug.cgi?id=1126) for updates.
51
52 # Luke
53
54 * cookbook example pospopcount with shriya
55 * load/store shifted ls004
56 * cookbook example maxloc, but stalled due to needing
57 ddffirst
58 * Working on Maxloc assembler example, run into same issue as Jacob,
59 where scalar src/dest but need loop to continue.
60 * So this feature is pretty important and needs to be added
61 to the SVP64 spec
62
63 # Shriya
64
65 * Load/store-shifted corrections, jacob to review (thank you)
66 * maxloc research
67 * pospopcount images
68
69 [[!tag meeting2023]]
70 [[!tag meeting_sync_up]]
71