lyken
/
nac3
forked from M-Labs/nac3
1
0
Fork 0
Code Pull Requests Activity
676 Commits 24 Branches 0 Tags 4.9 MiB
Commit Graph

676 Commits

Author SHA1 Message Date
Sebastien Bourdeauducq bc0f82cad8 Revert "nac3artiq/codegen: fixed smax problem" 
We have LLVM 12 now and can use the intrinsic.

This reverts commit 98d9f73afb.
 2021-11-02 14:00:28 +08:00
Sebastien Bourdeauducq 624dfe8cd1 upgrade to LLVM 12 2021-11-02 14:00:20 +08:00
pca006132 e47597bb8a Merge branch 'context-manager' 2021-11-02 11:17:00 +08:00
pca006132 98d9f73afb nac3artiq/codegen: fixed smax problem 
It turns out the smax intrinsic I use is a new one that is not supported
in LLVM11. Now implemented with signed integer compare and select.
 2021-11-02 11:10:21 +08:00
Sebastien Bourdeauducq da2886565b runkernel: sort out cargo rigmarole with rustc link arg 2021-11-01 00:28:27 +08:00
Sebastien Bourdeauducq b37cf6de08 nac3artiq: share isa->time_fns map 2021-11-01 00:03:15 +08:00
pca006132 083eacc268 with parallel/sequential support 
Behavior of parallel and sequential:
Each function call (indirectly, can be inside a sequential block) within a parallel
block will update the end variable to the maximum now_mu in the block.
Each function call directly inside a parallel block will reset the timeline after
execution. A parallel block within a sequential block (or not within any block) will
set the timeline to the max now_mu within the block (and the outer max now_mu will also
be updated).

Implementation: We track the start and end separately.
- If there is a start variable, it indicates that we are directly inside a
parallel block and we have to reset the timeline after every function call.
- If there is a end variable, it indicates that we are (indirectly) inside a
parallel block, and we should update the max end value.

Note: requires testing, it is difficult to inspect the output IR
 2021-10-31 23:54:37 +08:00
Sebastien Bourdeauducq 137efebb33 runkernel: add simple host kernel runner 2021-10-31 23:52:43 +08:00
Sebastien Bourdeauducq 443b95d909 nac3artiq: do not use custom linker script when targeting host 2021-10-31 23:51:50 +08:00
Sebastien Bourdeauducq 8b73a123cc nac3artiq: support compiling for the host 2021-10-31 23:02:21 +08:00
pca006132 84c5201243 with parallel/sequential support 
Behavior of parallel and sequential:
Each function call (indirectly, can be inside a sequential block) within a parallel
block will update the end variable to the maximum now_mu in the block.
Each function call directly inside a parallel block will reset the timeline after
execution. A parallel block within a sequential block (or not within any block) will
set the timeline to the max now_mu within the block (and the outer max now_mu will also
be updated).

Implementation: We track the start and end separately.
- If there is a start variable, it indicates that we are directly inside a
parallel block and we have to reset the timeline after every function call.
- If there is a end variable, it indicates that we are (indirectly) inside a
parallel block, and we should update the max end value.

Note: requires testing, it is difficult to inspect the output IR
 2021-10-31 17:16:21 +08:00
pca006132 558c3f03ef nac3core/codegen: list comprehension support 2021-10-24 16:53:43 +08:00
pca006132 45673b0ecc nac3core/codegen: cleanup 2021-10-24 16:53:43 +08:00
pca006132 181607008d nac3core/codegen: supports list iter 2021-10-24 14:39:50 +08:00
pca006132 fb92b6d364 nac3core: supports range iterator 2021-10-23 23:53:36 +08:00
pca006132 2f6ba69770 nac3core/typecheck: check if value is none 2021-10-23 21:31:14 +08:00
pca006132 cc83bbc63a nac3core/codegen: fix broken test 2021-10-17 13:07:45 +08:00
pca006132 279f47f633 nac3core/codegen: avoid sending unifiers 
Previously, we have to copy types from one unification table to another,
and make the table sendable. This requires cloning (processing) the
whole table 3 times per function call which is not efficient and uses
more memory than required when the unification table is large.

We now use a concrete type table to only copy the type we need. This
reduces the overhead as we only need to process the unification table
for once (when we do the function codegen), and reduces memory usage by
a bit (but not noticeable when the unification table is small, i.e. the
types are simple).
 2021-10-17 13:02:18 +08:00
pca006132 9850cbe313 nac3core/codegen: optimize for every function 
This speeds up compilation and reduces memory usage.
 2021-10-17 12:56:11 +08:00
pca006132 1f5bea2448 nac3core/codegen: refactor according to #23 2021-10-16 22:17:36 +08:00
pca006132 c4259d14d1 fixed some clippy warnings 2021-10-16 18:08:13 +08:00
pca006132 26076c37ba nac3core/typecheck: supports recursive type inference 2021-10-16 15:56:49 +08:00
Sebastien Bourdeauducq fd0b11087e nac3core: use round instead of rint. Closes #61 2021-10-11 08:18:52 +08:00
Sebastien Bourdeauducq 3a1dd893a1 nac3artiq/demo: get closer to regular ARTIQ 2021-10-10 17:45:38 +08:00
pca006132 a4ccac2329 nac3artiq: implements #55, #56 2021-10-10 16:26:01 +08:00
pca006132 77542170fd nac3core: fixes #60 2021-10-10 15:01:06 +08:00
pca006132 a3ce5be10b nac3core: fixes #32 and #57 2021-10-09 16:20:49 +08:00
Sebastien Bourdeauducq a22552a012 nac3artiq: work around #56 2021-10-09 15:52:45 +08:00
Sebastien Bourdeauducq 6ba74ed9f6 nac3artiq: allow creating drivers on device 2021-10-09 15:51:47 +08:00
Sebastien Bourdeauducq 8b32c8270d nac3artiq: explain delayed registration 2021-10-09 15:21:41 +08:00
Sebastien Bourdeauducq 5749141efb nac3artiq: add simple KernelInvariant CPython wrapper 2021-10-08 23:46:46 +08:00
Sebastien Bourdeauducq 3b10172810 nac3artiq: get closer to original ARTIQ semantics in demo 
Currently crashes the compiler with:
thread '<unnamed>' panicked at 'called `Option::unwrap()` on a `None` value', nac3core/src/codegen/expr.rs:395:58
 2021-10-08 23:41:41 +08:00
Sebastien Bourdeauducq 82efb0e720 missing file from previous commit 2021-10-08 23:21:36 +08:00
Sebastien Bourdeauducq d3a21d75fa handle time cursor functions on the Rust side entirely 
This is preparation for with sequential/with parallel support.
 2021-10-08 23:14:22 +08:00
pca006132 a07674a042 nac3artiq: host object supports typevar 2021-10-08 22:45:08 +08:00
Sebastien Bourdeauducq c5bcd352a5 flake: add hydraJobs 2021-10-08 08:18:42 +08:00
Sebastien Bourdeauducq 79d3c5caae switch to Nix flakes 2021-10-08 00:04:22 +08:00
pca006132 c697e522d3 nac3artiq: #33 demo 
The python API is changed a bit to allow running constructor with
@kernel annotation.
 2021-10-07 15:58:19 +08:00
pca006132 08947d20c2 nac3artiq: implements #33 2021-10-07 15:57:45 +08:00
pca006132 62673cf608 nac3artiq: add back builtins (fix #53) 2021-10-06 16:46:41 +08:00
pca006132 11144301ca nac3artiq: added simple host value support 2021-10-06 16:07:42 +08:00
ychenfo 4fcb54e463 nac3core: fix #46, better toplevel return type error msg 2021-10-03 18:07:45 +08:00
ychenfo 24b2111c64 nac3core: fix #45 toplevel better error msg for methods/functions 2021-10-03 17:25:28 +08:00
ychenfo f5ce1afe0b fix tests and switch to insta 
Use a library called 'insta' to better organize those longer correct test outputs in toplevel tests. 'insta' creates `.snap` files as snapshots of the test output, and will automatically do the diff if the output is different. This makes maintaining test cases with larger outputs a lot easier.

Reviewed-on: M-Labs/nac3#42
Co-authored-by: ychenfo <yc@m-labs.hk>
Co-committed-by: ychenfo <yc@m-labs.hk>
 2021-10-03 16:39:12 +08:00
Sebastien Bourdeauducq 915460ecb7 nac3artiq: fix typo 2021-10-03 16:18:15 +08:00
Sebastien Bourdeauducq b2c7f51d57 nac3artiq: guarantee ordering of pinned now stores 2021-10-03 10:00:42 +08:00
Sebastien Bourdeauducq 248d8cbece nac3artiq: handle now-pinning depending on target 2021-10-02 23:40:06 +08:00
Sebastien Bourdeauducq c429a86586 nac3artiq: refactor timeline functions 2021-10-02 23:35:28 +08:00
Sebastien Bourdeauducq c5e731f16d nac3artiq: implement timeline functions 2021-10-02 23:22:46 +08:00
Sebastien Bourdeauducq 0cbe4778d2 nac3standalone: demonstrate scalar conversion functions in mandelbrot 2021-10-02 19:21:59 +08:00