github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-09 06:58:11 -05:00
Code Issues Packages Projects Releases Wiki Activity

unify cpu and llvm (#11982)

Browse Source 
* try unify cpu and llvm

* fixes

* fix

* ops

* no llvm

* fix

* rm

* lvmm is ot

* oops

* override

* no llvm

* ignore

* skip llvm

* ooops
This commit is contained in:
nimlgen
2025-09-09 13:54:44 +03:00
committed by GitHub
parent 50cc7175cb
commit 1c6c42715f
25 changed files with 104 additions and 111 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
test/test_ops.py
View File

@@ -2,7 +2,7 @@ import time, math, unittest, functools, platform, warnings
import numpy as np
from typing import List, Callable
import torch
from tinygrad.helpers import getenv, IMAGE, DEBUG, CI, Context, TRANSCENDENTAL, AMD_LLVM
from tinygrad.helpers import getenv, IMAGE, DEBUG, CI, Context, TRANSCENDENTAL, CPU_LLVM, AMD_LLVM
from tinygrad import Tensor, Device, dtypes
from tinygrad.tensor import _to_np_dtype
from tinygrad.device import is_dtype_supported
@@ -1304,7 +1304,7 @@ class TestOps(unittest.TestCase):
np.arange(64,128,dtype=np.float32).reshape(8,8)])
def test_small_gemm_eye(self):
helper_test_op(None, lambda x,y: x.matmul(y), lambda x,y: x@y, vals=[np.eye(8).astype(np.float32), np.eye(8).astype(np.float32)])
@unittest.skipIf(CI and Device.DEFAULT in ["NV", "LLVM", "GPU", "CUDA"] or IMAGE
@unittest.skipIf(CI and Device.DEFAULT in ["NV", "GPU", "CUDA"] or (Device.DEFAULT == "CPU" and CPU_LLVM) or IMAGE
or (Device.DEFAULT == "WEBGPU" and platform.system() == "Windows"), "not supported on these in CI/IMAGE")
def test_gemm_fp16(self):
helper_test_op([(64,64), (64,64)], lambda x,y: x.half().matmul(y.half()), atol=5e-3, rtol=5e-3)
@@ -2262,7 +2262,7 @@ class TestOps(unittest.TestCase):
lambda x,w: torch.nn.functional.conv2d(x,w,stride=2),
lambda x,w: Tensor.conv2d(x,w,stride=2))
@unittest.skipIf(Device.DEFAULT != "LLVM", "DEVECTORIZE=0 only for LLVM")
@unittest.skipUnless(Device.DEFAULT == "CPU" and CPU_LLVM, "DEVECTORIZE=0 only for LLVM")
def test_strided_conv2d_simple_vec(self):
with Context(DEVECTORIZE=0): self.test_strided_conv2d_simple()