gpu Questions

1

PyTorch 1.12 changed the default fp32 math to be "highest precision", and introduced the torch.set_float32_matmul_precision API, allowing users to specify which precision out of medium, h...
Contention asked 30/7, 2023 at 9:39

1

Solved

UDPATE I found an error in my code. I was running my render function in sub-blocks from years ago and forgot I had set it as such. So it was calling the GPU read function far more often than I thou...
Presentational asked 17/8, 2024 at 19:21

2

Solved

Background I'm trying to understand whether a GPU's Last-Level Cache is invalidated or preserved across multiple kernel launches, so that the effective memory bandwidth can be increased. I'm aware ...
Spellbound asked 2/9, 2023 at 8:26

1

I am attempting to launch thrust::fill on two different device vectors in parallel on different CUDA streams. However, when I look at the kernel launches in NSight Systems, they appear to be serial...
Aldarcy asked 22/7, 2024 at 18:28

4

Solved

I use this notebook from Kaggle to run LSTM neural network. I had started training of neural network and I saw that it is too slow. It is almost three times slower than CPU training. CPU perfoma...
Valentine asked 24/9, 2018 at 13:56

10

Solved

tf-nightly version = 2.12.0-dev2023203 Python version = 3.10.6 CUDA drivers version = 525.85.12 CUDA version = 12.0 Cudnn version = 8.5.0 I am using Linux (x86_64, Ubuntu 22.04) I am coding in Vis...
Johnsonjohnsonese asked 2/3, 2023 at 11:19

4

Solved

As most of you know CPUs are not well designed to do floating point calculation in contrast to GPUs. I am wondering how to use GPU's power without any abstraction layer or driver. Can I program for...
Competent asked 22/7, 2010 at 7:49

9

Solved

I am working on a object detection project and wanting to process the project with my GPU. I have completed the NVIDIA setup tutorial and everything works fine. My object detection code originally ...
Tinny asked 24/5, 2022 at 2:25

3

Solved

CUDA document does not specific how many CUDA process can share one GPU. For example, if I launch more than one CUDA programs by the same user with only one GPU card installed in the system, what i...
Columbarium asked 27/7, 2015 at 0:55

3

Solved

I want to know how much RAM my visitors have and all the information available about their GPU. Is there any way to achieve this via JavaScript or maybe ActionScript (Flash)?
Unesco asked 15/4, 2013 at 22:2

2

I am frequently rerunning the same mxnet script while I try to iron out some bugs in a new script (and I am new to mxnet). Pretty often when I try to run my script I get an error that the GPU is ou...
Straticulate asked 5/12, 2018 at 20:45

3

Solved

Right now, I'm working on my master's thesis and I need to train a huge Transformer model on GCP. And the fastest way to train deep learning models is to use GPU. So, I was wondering which GPU shou...
Kaminsky asked 22/10, 2021 at 9:35

2

When I run "torch.rand(10).to("cuda")", I face "error : Torch not compiled with CUDA enabled" GPU : Nvidia RTX 3080 Ti $nvidia-smi +----------------------------------...
Galata asked 16/3, 2023 at 5:37

5

Solved

The error: nvidia-container-cli: initialization error: load library failed: libnvidia-ml.so.1: cannot open shared object file: no such file or directory: unknown I am trying to use nvidia/cuda i...
Chapland asked 14/10, 2022 at 4:17

2

Solved

LLVM has a back end for both AMD and NVIDIA GPUS. Is it currently possible to compile c++ (or a subset) to GPU code with clang and run it? Obviously things like the standard library would be unavai...
Millymilman asked 7/3, 2017 at 5:36

2

Solved

I save the last hidden layer of Bert for my following process using pickle on GPU. # output is the last hidden layer of bert, transformed on GPU with open(filename, 'wb') as f: pk.dump(output, f) ...
Woodenhead asked 21/7, 2020 at 6:47

3

I have some PyTorch code in one Jupyter Notebook which needs to run on one specified GPU (that is, not 'GPU 0') since others already work on 'GPU 0'. I have tried this: import os os.environ['CUDA_V...
Propagable asked 27/7, 2022 at 14:35

2

Solved

I have been trying to use a G5g EC2 instance with PyTorch 2.0 but I have been struggling to get it working. I want this specific instance because the arm processor makes it significantly cheaper, a...
Godchild asked 7/9, 2023 at 19:50

2

Solved

I'm trying to run stylegan2 on Google Colab but with all the files on my Drive and avoiding using !git clone from the github of stylegan2. Here is my code on the specific cell: %tensorflow_version ...
Hirschfeld asked 2/7, 2020 at 13:20

2

Solved

If we use a combination of the Dataset and Dataloader classes (as shown below), I have to explicitly load the data onto the GPU using .to() or .cuda(). Is there a way to instruct the dataloader to ...
Gluteal asked 28/1, 2021 at 6:56

5

I have tried downloading small files from google Colaboratory. They are easily downloaded but whenever I try to download files which have a large sizes it shows an error? What is the way to downloa...
Godsey asked 22/3, 2018 at 12:8

4

Solved

I am trying to use GPU with Tensorflow. My Tensorflow version is 2.4.1 and I am using Cuda version 11.2. Here is the output of nvidia-smi. +---------------------------------------------------------...
Glenoid asked 11/4, 2021 at 13:37

2

Solved

I am one of those miserable creatures who own a AMD GPU (RX 5700, Navi10). I want to use up-to-date PyTorch libraries to do some Deep Learning on my local machine and stop using cloud instances. I ...
Review asked 4/8, 2022 at 0:32

1

How can I train an XGBoost model on a GPU but run predictions on CPU without allocating any GPU RAM? My situation: I create an XGBoot model (tree_method='gpu_hist') in Python with predictor='cpu_pr...
Inmate asked 3/12, 2020 at 16:45

4

Solved

In training loop, I load a batch of data into CPU and then transfer it to GPU: import torch.utils as utils train_loader = utils.data.DataLoader(train_dataset, batch_size=128, shuffle=True, num_wo...
Arris asked 31/5, 2020 at 4:30

© 2022 - 2025 — McMap. All rights reserved.