whitequark
a2a00e8b35
Set LLVM inlining threshold explicitly.
...
llvmlite's pass manager builder does not use a default when it
is not specified explicitly, unlike the C++ one.
2016-01-10 16:15:23 +00:00
whitequark
f92ce291ef
transforms.llvm_ir_generator: call devirtualized functions directly.
2016-01-10 15:54:46 +00:00
whitequark
bbd0ca7827
Commit missing parts of 9366a29.
2016-01-10 15:45:56 +00:00
whitequark
c46fccdf4b
transforms.llvm_ir_generator: annotate invariant loads as such.
2016-01-10 15:45:34 +00:00
whitequark
b298170973
transforms.artiq_ir_generator: annotate embedded functions as internal.
2016-01-10 15:45:20 +00:00
whitequark
4198601abb
coredevice.exceptions: add CacheError.
2016-01-10 14:43:30 +00:00
whitequark
211af77e77
validators.escape: cache_get result lives forever.
2016-01-10 14:43:21 +00:00
whitequark
f8e50f2b0c
Remove redundant ksupport API.
2016-01-10 14:27:46 +00:00
whitequark
72a3a12d52
Fix versioneer ( closes #199 ).
2016-01-10 13:42:04 +00:00
whitequark
cc45694f5a
Commit missing parts of 9366a29.
2016-01-10 13:08:26 +00:00
whitequark
9366a29483
Implement core device storage ( fixes #219 ).
2016-01-10 13:04:55 +00:00
1543141cee
update copyright year
2016-01-08 11:53:04 +08:00
whitequark
f2f1deb9db
transforms.llvm_ir_generator: quote recrusive objects correctly ( fixes #213 ).
2016-01-07 19:52:31 +00:00
whitequark
027d54ca94
Enlarge coredevice buffers to 2.5MiB ( fixes #215 ).
...
This should be enough for sending a 2MiB int32 list.
2016-01-07 18:29:35 +00:00
whitequark
dcc4763914
Remove kernel-related constants from runtime.ld.
...
It's not used anywhere and it's one more place to adjust, which
will someday lead to obscure copy-paste bugs.
2016-01-07 18:27:38 +00:00
whitequark
577108554f
Move kernel CPU address space up to 0x40800000.
2016-01-07 18:26:11 +00:00
8778aef6a1
gui/schedule: hide vertical header
2016-01-07 16:56:48 +08:00
028ed72b72
doc: typo
2016-01-07 10:06:43 +08:00
87dd09a71c
gateware: compress bitstreams
2016-01-06 15:40:28 -07:00
0bbe886669
doc/manual: umask warning and openocd installation refinement
2016-01-06 13:12:42 -07:00
73b4aff3c2
README: drop old travis badge
2016-01-06 11:25:12 -07:00
358ad2e927
artiq_flash: drop redundant instruction, tweak doc
2016-01-06 11:24:05 -07:00
25ce0928bd
build/doc: simplify proxy bitstream locations
2016-01-05 10:23:50 -07:00
whitequark
400999f6ad
Revert "conda: use BUILDNUMBER from environment."
...
This reverts commit ff09a982c5
2016-01-05 16:06:05 +00:00
whitequark
07ef8906f4
conda: fix buildscripts to get bscan bitstreams correctly.
2016-01-05 15:47:41 +00:00
whitequark
ff09a982c5
conda: use BUILDNUMBER from environment.
2016-01-05 15:46:42 +00:00
57ae630c5c
conda: package new, openocd-compatible flash proxy bitstreams
2016-01-05 21:02:13 +08:00
whitequark
407a7d7bf8
conda: remove flterm dependency ( #185 ).
2016-01-05 04:31:02 +00:00
whitequark
e8bba0a7d6
Commit missing parts of 6a6d7d.
2016-01-05 04:21:46 +00:00
whitequark
6a6d7dab19
transforms.artiq_ir_generator: add support for user-defined context managers.
2016-01-05 04:10:40 +00:00
d633c8e1f8
doc: update and cleanup flashing/installing/building
2016-01-05 10:21:34 +08:00
d64d962c3e
remove redundant udev rules, update entry points
2016-01-05 10:20:54 +08:00
9aff99568d
artiq_flash: use openocd, python
2016-01-05 10:20:41 +08:00
b10da212ab
bit2bin: add small bitstream conversion tool
2016-01-05 10:20:24 +08:00
whitequark
2955f9a1e0
Commit missing parts of 5baf18b.
2016-01-05 00:16:58 +08:00
whitequark
dfbf55fed2
transforms/inferencer: add support for user-defined context manager.
2016-01-05 00:11:03 +08:00
whitequark
5baf18ba0d
transforms.inferencer: factor out _unify_attribute.
2016-01-04 22:13:05 +08:00
whitequark
03dd1c3a43
Refactor the logic of printing diagnostics to solely rely on Engine.
2016-01-04 22:11:54 +08:00
whitequark
05fa80818a
transforms.inferencer: make sure parallel/sequential is lone manager.
2016-01-04 21:26:03 +08:00
whitequark
3aa5acbaf3
transforms.llvm_ir_generator: don't crash if embedding is not enabled.
2016-01-03 19:07:11 +08:00
whitequark
81427af506
Remove selective attribute writeback; always synchronize all attributes.
...
The reasons are:
1. Shadow memory manipulation added ~12 instructions to TTLOut.pulse
(without inlining), and it's already barely fast enough.
2. More importantly, code such as self.ts[1] did not trigger
attribute writeback, and there seems to be no easy way to fix
that.
2016-01-03 04:57:58 +08:00
e41b9db787
gui: minor cleanup
2016-01-03 00:47:43 +08:00
6007b64898
applets: add big_number
2016-01-03 00:46:33 +08:00
5ad6c5cbe0
applets: add simple applet library
2016-01-03 00:46:15 +08:00
whitequark
38a99fde52
Implement selective attribute writeback using shadow memory.
2016-01-02 22:51:04 +08:00
whitequark
2e33084a5f
transforms.llvm_ir_generator: implement instrumentation for attribute writeback.
2016-01-02 06:51:30 +08:00
whitequark
5f68cc6a21
transforms.artiq_ir_generator: handle raise in except: with finally:.
2016-01-02 06:51:29 +08:00
whitequark
05bdd5c4a9
Commit missing parts of 8aa34ee9.
2015-12-31 22:42:21 +08:00
whitequark
693a364d85
transforms.artiq_ir_generator: fix typo.
2015-12-31 22:39:33 +08:00
whitequark
ff0ab736e9
Commit missing parts of 8aa34ee9.
2015-12-31 22:36:25 +08:00
