Software Engineering/Hardware & Acceleration 2022.02.22

[CUDA] CUDA C/C++ Basics Supercomputing 2011 Tutorial

참고자료 : https://www.nvidia.com/docs/IO/116711/sc11-cuda-c-basics.pdf

Software Engineering/Hardware & Acceleration 2021.11.18

[GPU] NVIDIA GPU 적정 온도 확인하기

NVIDIA 계열의 GPU의 온도를 확인하는 명령어는 다음과 같다. $ nvidia-smi -q -d temperature 아래는 현재 GPU의 Temperature 상태들을 나타낸다. ==============NVSMI LOG============== Timestamp : Thu Nov 18 15:52:50 2021 Driver Version : 460.73.01 CUDA Version : 11.2 Attached GPUs : 4 GPU 00000000:1A:00.0 Temperature GPU Current Temp : 64 C GPU Shutdown Temp : 90 C GPU Slowdown Temp : 87 C GPU Max Operating Temp : 83 C GPU Target Temper..

Software Engineering/Hardware & Acceleration 2021.02.05

[GPU] GeForce 20 series

en.wikipedia.org/wiki/GeForce_20_series GeForce 20 series - Wikipedia Series of GPUs by Nvidia The GeForce 20 series is a family of graphics processing units developed by Nvidia.[4] Serving as the successor to the GeForce 10 series,[5] the line started shipping on September 20, 2018,[6] and after several editions, on July 2, en.wikipedia.org

Software Engineering/Hardware & Acceleration 2021.01.26

[CUDA] 기존 CUDA 버전 삭제하기

다음과 같이 간단하게 CUDA 관련 파일들을 제거할 수 있다. 이는 기존에 설치된 CUDA 를 제거하고 새로운 버전(다운 그레이드 또는 업그레이드)을 설치하기 위함이다. 1. CUDA 제거 sudo apt-get --purge remove 'cuda*' sudo apt-get autoremove --purge 'cuda*' 2. CUDA 관련 폴더 제거 sudo rm -rf /usr/local/cuda-11.0 sudo rm -rf /usr/local/cuda

Software Engineering/Hardware & Acceleration 2021.01.11

[CUDA] Supported CUDA level of GPU and card

보다 더 자세한 사항은 맨 아래 링크를 참고하길 바란다. CUDA SDK 1.0 support for compute capability 1.0 – 1.1 (Tesla)[25] CUDA SDK 1.1 support for compute capability 1.0 – 1.1+x (Tesla) CUDA SDK 2.0 support for compute capability 1.0 – 1.1+x (Tesla) CUDA SDK 2.1 – 2.3.1 support for compute capability 1.0 – 1.3 (Tesla)[26][27][28][29] CUDA SDK 3.0 – 3.1 support for compute capability 1.0 – 2.0 (Tesla, Fermi)[30][31] CUD..

Software Engineering/Hardware & Acceleration 2020.07.13

[PyCUDA] PyCUDA 2019.1.2 소스 빌드하여 설치하기

CUDA 10.0 에서 CUDA 10.2 로 업데이트 하였을 때, 이미 설치가 되어있었던 pyCUDA 를 import 하였더니 아래와 같은 에러가 떴었다. ImportError: libcurand.so.10.0: cannot open shared object file: No such file or directory 이는 pyCUDA 가 자꾸 삭제된 CUDA 10.0 을 찾는 메세지 였고, linux 명령어인 find / -name "libcurand.so.10.0*" 으로 관련된 파일을 모두 삭제했는데도 불구하고 CUDA 10.0 을 찾는 문제가 발생하였다. 그래서 PyCUDA 를 다시 pip uninstall pycuda 하고 다시 설치하였는데도 불구하고 계속 같은 문제가 발생되어 cuda 10.2 을..

Software Engineering/Hardware & Acceleration 2020.07.06

[CUDA] CUDA Driver Version

CUDA Driver Running a CUDA application requires the system with at least one CUDA capable GPU and a driver that is compatible with the CUDA Toolkit. See Table 2. For more information various GPU products that are CUDA capable, visit https://developer.nvidia.com/cuda-gpus. Each release of the CUDA Toolkit requires a minimum version of the CUDA driver. The CUDA driver is backward compatible, meani..

Software Engineering/Hardware & Acceleration 2020.05.14

[PyCUDA] 정리중

https://documen.tician.de/pycuda/driver.html Device Interface — PyCUDA 2019.1.2 documentation Pagelocked Allocation pycuda.driver.pagelocked_empty(shape, dtype, order='C', mem_flags=0) Allocate a pagelocked numpy.ndarray of shape, dtype and order. mem_flags may be one of the values in host_alloc_flags. It may only be non-zero on CUDA 2.2 and newer. documen.tician.de https://documen.tician.de/pyc..

Software Engineering/Hardware & Acceleration 2020.03.04

[GPU] GPU Performance 및 Titan V, RTX 2080 Ti Benchmark

TPU V100 TiTan V Titan RTX RTX 2080 Ti RTX 2080 Titan Xp Titan X GTX 1080 Ti RTX 2070 RTX 2060 GTX 1080 GTX 1070 Ti GTX 1070 GTX 1660 Ti GTX 1060 Titan V Benchmark FP32 TensorFlow Training Performance FP16 TensorFlow Training Performance RTX 2080 Ti Benchmark RTX 2080 Ti - FP32 TensorFlow Performance (1 GPU) For FP32 training of neural networks, the RTX 2080 Ti is... 37% faster than RTX 2080 35%..

Software Engineering/Hardware & Acceleration 2019.08.04

[CUDA] PyCUDA documentation

간단한 예제 import pycuda.autoinit import pycuda.driver as drv import numpy from pycuda.compiler import SourceModule mod = SourceModule(""" __global__ void multiply_them(float *dest, float *a, float *b) { const int i = threadIdx.x; dest[i] = a[i] * b[i]; } """) multiply_them = mod.get_function("multiply_them") a = numpy.random.randn(400).astype(numpy.float32) b = numpy.random.randn(400).astype(numpy...