Closed cgyurgyik closed 3 years ago
Since this repo doesn't have any CI, did you run
runt
to check if the results work? If so, merge it.
Running say,
runt -i polybench/linear-algebra-2mm.fuse -d
I'm getting different outputs:
...
2│+ "A0_0": [
3│+ 10,
4│+ 98,
5│+ 38,
6│+ 18,
7│+ 34,
8│+ 97,
9│+ 61,
10│+ 71,
11│+ 42,
12│+ 99,
13│+ 90,
14│+ 76,
15│+ 93,
16│+ 89,
17│+ 64,
18│+ 4,
19│+ 17,
20│+ 79,
21│+ 0,
22│+ 24,
23│+ 78,
24│+ 47,
25│+ 2,
26│+ 98,
27│+ 49,
28│+ 18,
29│+ 76,
30│+ 82,
31│+ 12,
32│+ 92,
33│+ 80,
34│+ 36,
35│+ 37,
36│+ 34,
37│+ 60,
38│+ 9,
39│+ 73,
40│+ 12,
41│+ 87,
42│+ 7,
43│+ 69,
44│+ 46,
45│+ 80,
46│+ 77,
47│+ 69,
48│+ 50,
49│+ 82,
50│+ 15,
51│+ 89,
52│+ 2,
53│+ 52,
54│+ 13,
55│+ 11,
56│+ 4,
57│+ 65,
58│+ 49,
59│+ 14,
60│+ 80,
61│+ 81,
62│+ 61,
63│+ 32,
64│+ 81,
65│+ 32,
66│+ 85
67│+ ],
2 68│ "A_int0_0": [
3 69│ [
~
82 148│ ]
83 149│ ],
150│+ "B0_0": [
151│+ 39,
152│+ 50,
153│+ 44,
154│+ 33,
155│+ 9,
156│+ 13,
157│+ 49,
158│+ 54,
159│+ 77,
160│+ 21,
161│+ 35,
162│+ 70,
163│+ 24,
164│+ 99,
165│+ 51,
166│+ 65,
167│+ 69,
168│+ 91,
169│+ 33,
170│+ 61,
171│+ 65,
172│+ 75,
173│+ 0,
174│+ 36,
175│+ 67,
176│+ 14,
177│+ 92,
178│+ 92,
179│+ 47,
180│+ 6,
181│+ 0,
182│+ 63,
183│+ 60,
184│+ 73,
185│+ 86,
186│+ 23,
187│+ 12,
188│+ 45,
189│+ 47,
190│+ 76,
191│+ 63,
192│+ 14,
193│+ 51,
194│+ 60,
195│+ 74,
196│+ 16,
197│+ 15,
198│+ 87,
199│+ 89,
200│+ 73,
201│+ 32,
202│+ 63,
203│+ 46,
204│+ 52,
205│+ 96,
206│+ 45,
207│+ 67,
208│+ 42,
209│+ 83,
210│+ 8,
211│+ 40,
212│+ 43,
213│+ 81,
214│+ 73
215│+ ],
...
So it looks like A0_0
is included in mine but not in the benchmark's expected
?
Odd, maybe worth checking what happens on an older commit of the Calyx compiler.
Hey @cgyurgyik, this needs to be merged sooner rather than later since other PRs are blocked on it (specifically, the unify std_div and std_mod PR needs to run these tests)
Hey @cgyurgyik, this needs to be merged sooner rather than later since other PRs are blocked on it (specifically, the unify std_div and std_mod PR needs to run these tests)
With exams I won't have time to touch this until likely the weekend. Given that this is unsigned bitnum, I'm pretty confident that all my PR changes is the JSON format. It's hard to say whether the .expect
file diffs is from my PR or an earlier one without integrated tests, and I haven't been able to follow your recommendation above yet.
Since this repo doesn't have any CI, did you run
runt
to check if the results work? If so, merge it.