We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent dd71541 commit 3ed4a70Copy full SHA for 3ed4a70
test/integration/triton/softmax.jl
@@ -19,6 +19,8 @@ function softmax_triton(x::AbstractMatrix{T}) where {T}
19
out = similar(x_transposed)
20
n_rows, n_cols = size(x_transposed)
21
22
+ BLOCK_SIZE = nextpow(2, n_cols)
23
+
24
function grid_fn(metadata)
25
occupancy = (
26
metadata.device_properties.regs_per_block ÷
@@ -43,7 +45,7 @@ function softmax_triton(x::AbstractMatrix{T}) where {T}
43
45
n_rows,
44
46
n_cols,
47
BLOCK_SIZE,
- num_stages;
48
+ num_stages=3;
49
grid=grid_fn,
50
)
51
0 commit comments