github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-02-07 13:15:01 -05:00
Code Issues Packages Projects Releases Wiki Activity

linearizer! (#714)

Browse Source 
* linearizer outputs something

* working ish

* cstyle codegen

* clang mostly works

* fix load valid

* fix numberless loop

* fancy gen

* working

* fix enet compiler

* cleanups

* float4 upcasting

* less lines

* supports_float4

* constant folding

* mulacc

* internet tests flaky in CI

* 90% image support

* fix image generic

* bugs exposed with shapetracker and single view

* new llvm

* use vload, remove OLD

* that's really poorly done

* ending up being more lines
This commit is contained in:
George Hotz
2023-03-19 23:43:49 -07:00
committed by GitHub
parent b629fd4cd8
commit 5495c7d64e
20 changed files with 792 additions and 776 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
test/test_dtype.py
View File

@@ -4,9 +4,9 @@ from tinygrad.helpers import getenv
from tinygrad.lazy import Device
from tinygrad.tensor import Tensor, dtypes
# for GPU, cl_khr_fp16 isn't supported
# for GPU, cl_khr_fp16 isn't supported (except now we don't need it!)
# for LLVM, it segfaults because it can't link to the casting function
@unittest.skipIf(getenv("CI", "") != "" and Device.DEFAULT in ["GPU", "LLVM"], "float16 broken in some CI backends")
@unittest.skipIf(getenv("CI", "") != "" and Device.DEFAULT in ["LLVM"], "float16 broken in some CI backends")
class TestDtype(unittest.TestCase):
def test_half_to_np(self):
a = Tensor([1,2,3,4], dtype=dtypes.float16)