• Joined on 2019-03-28
sb10q commented on pull request M-Labs/artiq-zynq#309 2024-07-31 12:05:15 +08:00
kernel: add linalg functions

Remove all those "safety" headers, C calls are never "safe" as per Rust's definition.

sb10q commented on pull request M-Labs/artiq-zynq#309 2024-07-31 12:03:33 +08:00
kernel: add linalg functions

As far as I can tell, these two are trivial and do not need firmware calls. Suggest moving to IRRT.

sb10q commented on pull request M-Labs/artiq-zynq#309 2024-07-31 12:02:43 +08:00
kernel: add linalg functions

Needs update in ARTIQ and PR there.

sb10q commented on pull request M-Labs/artiq-zynq#309 2024-07-31 12:01:45 +08:00
kernel: add linalg functions

I don't think this ValueError addition belongs in this PR?

sb10q commented on pull request M-Labs/artiq-zynq#309 2024-07-29 18:44:34 +08:00
kernel: add linalg functions

Do we really need to call the firmware for something as simple as transpose? Should probably go to IRRT instead.

sb10q commented on pull request M-Labs/it-infra#40 2024-07-29 15:33:50 +08:00
Deploy web2019 to the intl domain

cp -a or similar?

sb10q deleted branch fix/32bit-codegen-issues from M-Labs/nac3 2024-07-29 15:05:07 +08:00
sb10q pushed to master at M-Labs/nac3 2024-07-29 15:05:07 +08:00
318a675ea6 standalone: Rename -m32 to -i386
32e52ce198 standalone: Revert using uint32_t as slice length
Compare 2 commits »
sb10q merged pull request M-Labs/nac3#481 2024-07-29 15:05:05 +08:00
Updates to 32-bit codegen
sb10q commented on pull request M-Labs/nac3#481 2024-07-29 15:02:29 +08:00
Updates to 32-bit codegen

Half that memory may be uninitialized and also this will break on any big-endian architecture (not that there are many left out there).

sb10q opened issue M-Labs/web2019#144 2024-07-29 14:42:42 +08:00
search function should display "no results" when appropriate
sb10q commented on pull request M-Labs/nac3#481 2024-07-29 13:50:05 +08:00
Updates to 32-bit codegen

Thought the compiler had to be changed as well? Or does it work right now just because x86 is little endian?

sb10q commented on pull request M-Labs/nac3#481 2024-07-29 13:28:42 +08:00
Updates to 32-bit codegen

These changes from #477 would need to be reverted as well image

sb10q commented on pull request M-Labs/nac3#477 2024-07-29 13:15:23 +08:00
Fix codegen issues for 32-bit targets

Yes, just like they are defined in any other serious compiler.

sb10q commented on pull request M-Labs/nac3#477 2024-07-29 13:11:55 +08:00
Fix codegen issues for 32-bit targets

Is there a reason not to follow the Rust cslice?

The firmware uses Rust cslice and this will become a problem when we start using it on 64-bit CPUs like Zynq MPSoC.

sb10q pushed to master at M-Labs/web2019 2024-07-29 12:55:24 +08:00
580dd1964c advertise AOM integration
sb10q pushed to master at M-Labs/nix-scripts 2024-07-28 11:54:31 +08:00
e9432e25cb gluelogic: update vivado
sb10q pushed to master at M-Labs/artiq-extrapkg 2024-07-27 23:24:04 +08:00
25ee3c68f1 wand: remove (no qt6)
sb10q pushed to master at M-Labs/artiq-extrapkg 2024-07-27 22:44:54 +08:00
7f00ab0cb6 update conda, py3.12
sb10q pushed to master at M-Labs/artiq-extrapkg 2024-07-27 22:42:19 +08:00
d5c99474ed linien: py3.12