Kugadzira uye kuongorora CUDA maapplication paNVIDIA Tesla P40 (24 GB VRAM, Pascal architecture). Unogona kuisa CUDA Toolkit vhezheni yaunoda. GPU servers hazvigone kupihwa; taura nesu kuti uzive kana zvavepo.
$ apt install nvidia-cuda-toolkit && nvcc hello_cuda.cu -o hello_cuda && ./hello_cuda
# Mufananidzo wekuisa command
CUDA i NVIDIA's parallel computing platform. A GPU server inokupa CUDA nharaunda yekunyora, kuyedza uye benchmarking GPU code pa Pascal (compute capacity 6.1) GPU.
Install nvcc, CUDA mabhuku uye profiling zvinhu mushanduro yako code zvinangwa.
Kuisa cuDNN ye deep learning primitives.
Profile uye debug GPU kodhi neNVIDIA Nsight.
Install chero mushandisi, CUDA vhezheni kana kuvandudza chirongwa.
CUDA Development inogona kushandisa NVIDIA GPU kuti iite basa rayo. VPS.org GPU servers hazvina kuwanikwa kuti utore; taura nesu kuti uzive kana vachiri. CUDA Development inowanzo GPU-yakakwira basa. Ita shuwa kuti software yako ine CUDA rutsigiro uye kuti yako driver / runtime versions inosangana nezvinodiwa zvebasa re CUDA Development.
GPU servers hadzina kuwanikwa kuti utore. Nhasi unogona kuisa CUDA Development pane yakajairika CPU package, uko inoita basa rakakura kupfuura pane GPU: SSH in uye shandisa apt install nvidia-cuda-toolkit && nvcc hello_cuda.cu -o hello_cuda && ./hello_cuda.
VRAM zvinodiwa CUDA Development kuenderana nemhando uye batch size; tarisa chirongwa ruzivo rwezvinoreva.
GPU mutengo achaburitswa kana GPU servers anogona kupihwa. Tibate kana iwe uchida kuziva kana ichiitika.
Yega. Every VPS.org server inouya nefull root access, saka unogona kutamba zvimwe zvinhu pamwe CUDA Development mukati server's RAM uye kuchengetera.
Every VPS.org server inouya nepfupi root SSH kuwanikwa, saka unogona kuisa madhiraivha, mabhuku uye CUDA Development asi iwe unofanirwa.
Hapana server inouya neCUDA runtime kana NVIDIA mutyairi akaiswa, uye GPU servers hazvina kuwanikwa kuti aite kusvika parizvino. Unogona kuisa zvinyorwa CUDA Development zvinoda iwe pachako.
VPS.org servers ndezvemapurojekiti akaomarara, kwete ephemeral instances: CUDA Development models, configs uye data inogara paSSD storage yeserver kusvika wabvisa.
Tenga data rekushandisa pane SSD ye server yekuwedzera kuwanikwa kwenguva pfupi panguva ye CUDA Development runs, uye kopa zvakapera zviwanikwa (zviyero, zvizvarwa, embeddings) kune yekuchengetera kunze kwe VPS.org; snapshots uye backups vanogara pamwechete kuchengetera cluster.
Unogona kusimudzira chero VPS.org server kune yakakura chirongwa kubva kudzora panel; resize rinotangazve uye yako CUDA Development kuiswa inotora pamusoro. GPU servers hazvina kuwanikwa kusvika parizvino.
Yeah. Backups ndivo akabhadhara add-on (zuva nezuva + 30% yemari chirongwa, akachengetwa 7 mazuva; zuva nezuva + 20%, akachengetwa 28 mazuva) uye snapshots vakasununguka, izvo zvinokosha pamberi CUDA Development mazuva. Both stay on the same storage cluster as the server.
Contact support within 30 days of your first purchase to request a refund. GPU servers are not sold yet, so no GPU order is involved.
GPU kugona kuri kuwanikwa chete. Tiudzai nezvebasa rako uye tichakuudza kuti ripi rinowanikwa uye ripi risiri.