-
I tried to run your tutorial from bootstrap/mnist/train.ipynb, but it crashes because of runtime errors related to ninja_build (see details below). When I switch to "device = torch.device('cpu')" it…
-
| | |
| --- | --- |
| Bugzilla Link | [32304](https://llvm.org/bz32304) |
| Version | 4.0 |
| OS | All |
| Attachments | [reproducer](https://user-images.githubusercontent.com/92601847/143755243-9ae…
-
MkFit as it is builds only on Intel architectures, yet at CMS we also have non production ARM and PowerPC builds. To bring MkFit also into this non prod builds I introduced some hacks here:
https://g…
-
## ❓ Questions and Help
Hello,
Currently, PyTorch has less AVX2 support for `float16` in [`aten/src/ATen/cpu/vec256`](https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/cpu/vec256/), …
-
### Background and motivation
#24328 and subsequent changes have added support for CRC-32, CRC-64, XxHash32, XxHash64, and XxHash128 non-cryptographic hash algorithms to the System.IO.Hashing library…
-
- shifting by more bits than the data type size(it's defined in some intel-intrinsic operations, but not sure if legitimate)
- cast out-of-range float to an integer type that cannot hold it
=> doc…
-
As noted in https://github.com/dotnet/runtime/issues/86168, the `float` and `double` overloads of the 128, 256, and 512 bit `GetMantissa()` methods differ from the Intel intrinsics' signatures of the …
-
I appreciate that this is on the face of it an unhelpful issue to report however
It really does feel very slow processing Eiger data on my M2 Pro, and I suspect a chunk of that is that the decompre…
-
With Intel 18 or newer we can use `CHPL_ATOMICS=cstdlib`. The main concern for
us is that gcc >=5 has to be in the `PATH` and that's not something we can
easily guarantee on a Cray.
For the modul…
-
Hi razayunus,
First of all congratulations on your great work.
I have successfully migrated the project to Visual Studio 2019. The examples (TUM RGB-D, ICL-NUI; and TAMU RGB-D) work perfectly and in…