David Mak derppening
  • Joined on 2023-08-31
derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-03-13 00:45:23 +08:00
f663efe1d0 WIP - Refactored NDArrays to split into refcounted and non-refcounted
Compare 71 commits »
derppening commented on pull request M-Labs/nac3#752 2026-03-13 00:44:31 +08:00
Fix incorrect overflow

If this bug is reproducible as a nac3artiq/demo demo, I would put that there as well.

derppening commented on pull request M-Labs/nac3#752 2026-03-11 19:54:44 +08:00
Fix incorrect overflow

Actually, could you explain why the previous approach does not work? Is it solely because it doesn't take the type hint into consideration when parsing the type of the list elements?

derppening commented on pull request M-Labs/nac3#752 2026-03-09 15:08:45 +08:00
Fix incorrect overflow

FYI it works for KernelInvariant[np.uint32] now but not for KernelInvariant[list[np.uint32]]. I'm still trying to figure out why. Maybe I need more time to understand how the whole thing…

derppening approved M-Labs/nac3#752 2026-03-08 16:26:14 +08:00
Fix incorrect overflow

LGTM.

derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-03-05 16:01:47 +08:00
Compare 2 commits »
derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-03-04 17:53:12 +08:00
c57d44fec0 WIP - Refactor NDArrayValue methods into
e6b6573200 WIP; DO NOT MERGE - More fixes towards IRRT
Compare 10 commits »
derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-03-03 16:55:18 +08:00
2c37a8af79 WIP - Refactor NDArrayValue methods into
f3b80edd2d WIP; DO NOT MERGE - More fixes towards IRRT
20c15208a5 WIP - Add ndarray.{base,offset}
Compare 47 commits »
derppening opened issue M-Labs/nac3#751 2026-03-02 14:06:41 +08:00
Mismatching behavior for numpy.array(list, copy=False)
derppening commented on pull request M-Labs/nac3#750 2026-03-01 09:14:52 +08:00
Add documentation for NAC3

I do agree though that this is the wrong format

That is my major concern about the documentation in its current form. Moving a lot of these back into the Rustdoc and perhaps compacting it…

derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-02-28 21:16:03 +08:00
4736d720d8 WIP; DO NOT MERGE - More fixes towards IRRT
eb20d02be1 WIP - Add ndarray.{base,offset}
27a206f6b8 WIP: Add standard support library to IRRT
Compare 3 commits »
derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-02-27 17:03:02 +08:00
9aef78f7fd [core] More fixes towards refcount-aware lists
d54151b01c [core] Fix segfault due to uninitialized list data
Compare 3 commits »
derppening commented on pull request M-Labs/nac3#750 2026-02-25 09:22:55 +08:00
Add documentation for NAC3

Several general comments:

derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-02-24 18:33:52 +08:00
c6fa263791 [core] More fixes towards stmt.rs
295bc3f466 [core] Use AllocationScope::Default by default
cc037dd745 [core] codegen: Add RefType::refcounted_type, relax
423ad6f085 [core] Fix len(list) to be refcount-aware
Compare 41 commits »
derppening pushed to wip/feature/ctrc at M-Labs/nac3 2026-02-24 16:58:38 +08:00
fc7c69868d WIP Stash
5a235e6248 [core] Use refcounted arrays for lists
bc89f92859 [standalone] Update flags
7fd45dde85 [core] Keep more LLVM entries in debug mode
Compare 61 commits »
derppening pushed to refactor/anyhow at M-Labs/nac3 2026-02-23 15:43:08 +08:00
6e30620117 [artiq] Collapse PyResult<anyhow::Result<_>> into anyhow::Result<_>
derppening pushed to refactor/anyhow at M-Labs/nac3 2026-02-22 00:14:52 +08:00
dc670f3df3 [meta] Refactor Result to anyhow::Result
derppening pushed to refactor/anyhow at M-Labs/nac3 2026-02-21 17:40:28 +08:00
9d2d597b2c [meta] Refactor Result to anyhow::Result
bc3c7f0e21 [core] Refactor ModuleContext and Builder in CodeGenContext
Compare 2 commits »
derppening created pull request M-Labs/nac3#747 2026-02-21 16:57:42 +08:00
WIP: Refactor to use anyhow
derppening commented on pull request M-Labs/nac3#731 2026-02-21 16:50:17 +08:00
Unify primitive type constructors with their class definitions

After looking at this again, I don't think this "cleanup" makes sense.

Treating all primitive-named functions as constructors breaks the fundamental assumption that a constructor creates…