compiler: Uncache data carriers #1913
Conversation
Codecov Report
@@ Coverage Diff @@
## master #1913 +/- ##
=======================================
Coverage 89.64% 89.65%
=======================================
Files 211 211
Lines 35634 35639 +5
Branches 5369 5369
=======================================
+ Hits 31945 31951 +6
Misses 3189 3189
+ Partials 500 499 -1
Continue to review full report at Codecov.
|
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
FabioLuporini
force-pushed
the
uncache-data-carriers
branch
from
7b61c6b to
5b9adbe
Compare
| @@ -303,7 +303,7 @@ | |||
| " for (int y = y_m; y <= y_M; y += 1)\n", | |||
| " {\n", | |||
| " float r3 = -2.0F*u[time][x + 2][y + 2];\n", | |||
| " u[time + 1][x + 2][y + 2] = dt*(r0*u[time][x + 2][y + 2] + c*(r1*r3 + r1*u[time][x + 1][y + 2] + r1*u[time][x + 3][y + 2] + r2*r3 + r2*u[time][x + 2][y + 1] + r2*u[time][x + 2][y + 3]));\n", | |||
| " u[time + 1][x + 2][y + 2] = dt*(c*(r1*r3 + r1*u[time][x + 1][y + 2] + r1*u[time][x + 3][y + 2] + r2*r3 + r2*u[time][x + 2][y + 1] + r2*u[time][x + 2][y + 3]) + r0*u[time][x + 2][y + 2]);\n", | |||
There was a problem hiding this comment.
Is this change of order deterministic?
| # Obviously: | ||
| assert foo0 is not foo1 | ||
| assert foo0 is not foo2 | ||
| assert foo1 is not foo2 |
There was a problem hiding this comment.
Should the dtype and value be checked too for safety?
There was a problem hiding this comment.
"checked" how? like foo0.dtype is foo1.dtype ? that's obvious imho
Oh, it boils down to how SymPy internally orders objects... it uses the |
Scalar data carriers are unique, so why attempt to cache them?