mirror of
https://github.com/tinygrad/tinygrad.git
synced 2026-04-29 03:00:14 -04:00
bfcec234a2
* ugh worst branch name * compiler refactor continues * scc -> cloc * buf -> _buf * finish _buf, and program -> runtime * gpu is still working, clang isn't * clang in new style * ops_metal * something broke it * improve metal * clean up tons of cl crap * hack fix sync * cleaner gpu * gpu metal clang * cleanups * minor refactor * GPUCodegen * fix up LLVM * blind CUDA refactor * codegen / runtime * keep ops naming * linter passes * woah, llvm was allocing 4x what it needed to * bugfixes * fix openpilot compiler * fix compile_efficientnet * method cache should fix tests * deal with duped functions
25 lines
775 B
Python
25 lines
775 B
Python
#!/usr/bin/env python
|
|
import unittest
|
|
from tinygrad.tensor import Tensor, Device
|
|
from tinygrad.nn import Conv2d
|
|
from tinygrad.ops import GlobalCounters
|
|
|
|
@unittest.skipUnless(Device.DEFAULT == "GPU", "Only GPU supports cache")
|
|
class TestConvShapetracker(unittest.TestCase):
|
|
def test_conv_3x3_one_view(self):
|
|
inp = Tensor.randn(1,16,10,10).realize()
|
|
conv = Conv2d(16, 32, (3,3))
|
|
conv(inp).realize()
|
|
GlobalCounters.cache = []
|
|
conv(inp).realize()
|
|
test = GlobalCounters.cache
|
|
GlobalCounters.cache = None
|
|
assert len(test) == 1, f"conv should only have one kernel {[x[0].name for x in test]}"
|
|
print(test[0][0].prg)
|
|
for arg in test[0][1]:
|
|
print(arg.st)
|
|
assert len(arg.st.views) == 1
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|