-
Latest GPUCompiler enables IR verification on CI, and it immediately spotted the following bug:
```
device/intrinsics/wmma: Error During Test at /var/lib/buildkite-agent/builds/gpuci-8/julialang/c…
-
Seems like you wanna to do an implementation of int8 gemm to do NLP inference? I wonder why you choose to hands on wmma api to do int8 gemm instead of opensource library like cutlass or vendor library…
-
**Describe the bug**
I encountered an error while testing CUDA. This is my first time trying to use CUDA. The culprit appears to be cudadrv (drivers?). I already tried updating my Nvidia drivers.
…
-
The CI is ready, working on understanding the results - quite different from what we get from `torchbench.py`
-
Using the current CUDA.jl master (`146ad00c0`) the following kernel works with Julia 1.7.2 but fails with Julia 1.8.0-beta1.
```julia
function kernel_wmma_int8_lowlevel(a_dev, b_dev, c_dev, d_dev)…
-
### 请提出你的问题 Please ask your question
- PaddlePaddle 2.4.0rc0
- Ubuntu 18.04
参考的文档：https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/performance_improving/amp_cn.html
使用的模型包括deepsp…
-
**Platform**
Android
**Version**
4.4.11 ver 808 from Google play store.
**Describe the bug**
Air sweep = crash page
**To Reproduce**
Basically choose air sweep and it'll crash.
-
**Please leave any comments or edit this issue directly to adjust the release notes! Also see the rc0 vote thread in #12103**.
# Introduction
The TVM community has worked since the v0.8 release …
-
Hi,
I would like to execute a CUTLASS GEMM (A*B+C) that uses the Tensor Cores on my Volta architecture with :
- matrix A size = 6x123
- matrix B size = 64x6
- matrix C size = 64x123
So, it …
-
While running `test_integration_cuda_tensorcore.py`, I got the following error.
```
@tvm.script.tir
class Module:
def main(var_A: ty.handle, var_B: ty.handle, var_C: ty.handle) -> None:
…