github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-09 15:08:02 -05:00
Code Issues Packages Projects Releases Wiki Activity

from_blob for cuda (#9223)

Browse Source 
* from_blob for cuda

* maybe docs?

* minor docs

* example

* waiting 9224

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
This commit is contained in:
nimlgen
2025-02-24 14:02:06 +03:00
committed by GitHub
parent fc32ff80d6
commit 1d06d61b16
3 changed files with 55 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

45
docs/runtime.md
View File

@@ -13,3 +13,48 @@ tinygrad supports various runtimes, enabling your code to scale across a wide ra
| [CPU (C Code)](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_cpu.py) | Runs on CPU using the clang compiler | `clang` compiler in system `PATH` |
| [LLVM (LLVM IR)](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_llvm.py) | Runs on CPU using the LLVM compiler infrastructure | llvm libraries installed and findable |
| [WEBGPU](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_webgpu.py) | Runs on GPU using the Dawn WebGPU engine (used in Google Chrome) | Dawn library installed and findable. Download binaries [here](https://github.com/wpmed92/pydawn/releases/tag/v0.1.6). |
## Interoperability
tinygrad provides interoperability with OpenCL and PyTorch, allowing efficient tensor data sharing between frameworks through the `Tensor.from_blob` API. This enables zero-copy operations by working directly with external memory pointers.
**Important**: When using external memory pointers with tinygrad tensors, you must ensure these pointers remain valid throughout the entire lifetime of the tinygrad tensor to prevent memory corruption.
### `CUDA` PyTorch Interoperability
You can seamlessly work with CUDA tensors between PyTorch and tinygrad without data copying:
```python
from tinygrad.dtype import _from_torch_dtype
tensor1 = torch.tensor([1.0, 2.0, 3.0], device=torch.device("cuda"))
tiny_tensor1 = Tensor.from_blob(tensor1.data_ptr(), tensor1.shape, dtype=_from_torch_dtype(tensor1.dtype), device='CUDA')
```
### `QCOM` OpenCL Interoperability
tinygrad supports OpenCL interoperability on `QCOM` backend.
Buffer interop allows direct access to OpenCL memory buffers:
```python
# create raw opencl buffer.
cl_buf = cl.clCreateBuffer(cl_context, cl.CL_MEM_READ_WRITE, 0x100, None, status := ctypes.c_int32())
# extract pointers
cl_buf_desc_ptr = to_mv(ctypes.addressof(cl_buf), 8).cast('Q')[0]
rawbuf_ptr = to_mv(cl_buf_desc_ptr, 0x100).cast('Q')[20] # offset 0xA0 is a raw gpu pointer.
# create tiny tensor
tiny = Tensor.from_blob(rawbuf_ptr, (8, 8), dtype=dtypes.int, device='QCOM')
```
And the same for the images:
```python
# create cl image.
cl_img = cl.clCreateImage2D(cl_context, cl.CL_MEM_READ_WRITE, cl.cl_image_format(cl.CL_RGBA, cl.CL_FLOAT), w, h, 0, None, status := ctypes.c_int32())
# extract pointers
cl_buf_desc_ptr = to_mv(ctypes.addressof(cl_img), 8).cast('Q')[0]
rawbuf_ptr = to_mv(cl_buf_desc_ptr, 0x100).cast('Q')[20] # offset 0xA0 is a raw gpu pointer.
# create tiny tensor
tiny = Tensor.from_blob(rawbuf_ptr, (h*w*4,), dtype=dtypes.imagef((h,w)), device='QCOM')
```