GPU Cloud
GPU cloud: ku kiro seerfarada NVIDIA GPU saacadle
Kiro seerfarada NVIDIA GPU ee LLM inference, fine-tuning iyo samaynta sawirrada. Qiimo toos ah, biil saacadle ah, templates-ka PyTorch, vLLM iyo ComfyUI, iyo helitaanka SSH.
- Kayd toos ah, qiimaha oo la xaqiijiyo intaadan dalban
- Templates diyaarsan oo loogu talagalay PyTorch, vLLM, Ollama iyo ComfyUI
- Seerfarada saacadlaha ah ee la damiyo waxaa laga qaadaa lacagta disk-ga oo keliya
Adigoo adeegsanaya Cloudj GPU Cloud waxaad server GPU ah ka kireysan kartaa kaydka tooska ah ee diyaar u ah: dooro nooca GPU-ga, VRAM iyo disk-ga, kaga bilow template sida PyTorch, vLLM, Ollama ama ComfyUI, kadibna kula xiriir SSH. Server-yada saacadlaha ah lacagta waxaa looga gooyaa hadhaaga akoonkaaga, server-ka saacadlaha ah ee la damiyeyna waxa kaliya oo biil looga qaadaa disk-giisa.
Qiimaha kirada GPU iyo helitaanka tooska ah
Kaydka tooska ah: helitaanka iyo qiimaha waxay isbeddelaan marka loo eego suuqa. Koonsolka ayaa xaqiijinaya qiimaha kama dambaysta ah ka hor inta aadan dalban.
| GPU | VRAM GPU kasta | Qaab-dhismeedka | Xisaabinta FP32 | Xadka awoodda korontada | Laga bilaabo | |
|---|---|---|---|---|---|---|
| NVIDIA RTX PRO 6000 Max-Q | 96 GB | Blackwell | ≈ 119 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA RTX PRO 6000 WS | 96 GB | Blackwell | ≈ 119 TFLOPS | 600 W | – | Faahfaahin → |
| NVIDIA RTX 5090 | 32 GB | Blackwell | ≈ 107.6 TFLOPS | 575 W | – | Faahfaahin → |
| NVIDIA RTX 5090D | 32 GB | Blackwell | ≈ 107.6 TFLOPS | 575 W | – | Faahfaahin → |
| NVIDIA RTX PRO 6000 S | 96 GB | Blackwell | ≈ 93.6 TFLOPS | 600 W | – | Faahfaahin → |
| NVIDIA RTX 4090 | 24 GB | Ada Lovelace | ≈ 81.4 TFLOPS | 450 W | – | Faahfaahin → |
| NVIDIA RTX 6000Ada | 48 GB | Ada Lovelace | ≈ 81.4 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA RTX 6000D | 84 GB | Blackwell | ≈ 77.6 TFLOPS | 600 W | – | Faahfaahin → |
| NVIDIA L40S datacenter | 45 GB | Ada Lovelace | ≈ 73.3 TFLOPS | 350 W | – | Faahfaahin → |
| NVIDIA RTX PRO 5000 | 48 GB | Blackwell | ≈ 69.6 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA H100 SXM datacenter | 80 GB | Hopper | ≈ 53.5 TFLOPS | 700 W | – | Faahfaahin → |
| NVIDIA RTX 5080 | 16 GB | Blackwell | ≈ 53.4 TFLOPS | 360 W | – | Faahfaahin → |
| NVIDIA RTX 4080S | 16 GB | Ada Lovelace | ≈ 50.9 TFLOPS | 320 W | – | Faahfaahin → |
| NVIDIA RTX 4080 | 16 GB | Ada Lovelace | ≈ 48.6 TFLOPS | 285 W | – | Faahfaahin → |
| NVIDIA H100 NVL datacenter | 94 GB | Hopper | ≈ 48.3 TFLOPS | 400 W | – | Faahfaahin → |
| NVIDIA H200 NVL datacenter | 140 GB | Hopper | ≈ 48.3 TFLOPS | 600 W | – | Faahfaahin → |
| NVIDIA RTX 5070 Ti | 16 GB | Blackwell | ≈ 44.3 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA RTX PRO 4000 | 24 GB | Blackwell | ≈ 44.3 TFLOPS | 145 W | – | Faahfaahin → |
| NVIDIA RTX 4070S Ti | 16 GB | Ada Lovelace | ≈ 42.6 TFLOPS | 285 W | – | Faahfaahin → |
| NVIDIA H100 PCIE datacenter | 80 GB | Hopper | ≈ 41 TFLOPS | 350 W | – | Faahfaahin → |
| NVIDIA RTX 4500Ada | 24 GB | Ada Lovelace | ≈ 38.2 TFLOPS | 210 W | – | Faahfaahin → |
| NVIDIA RTX 3090 Ti | 24 GB | Ampere | ≈ 36.4 TFLOPS | 450 W | – | Faahfaahin → |
| NVIDIA RTX A6000 | 48 GB | Ampere | ≈ 36.1 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA RTX 4070S | 12 GB | Ada Lovelace | ≈ 35.8 TFLOPS | 220 W | – | Faahfaahin → |
| NVIDIA RTX 3090 | 24 GB | Ampere | ≈ 35.4 TFLOPS | 350 W | – | Faahfaahin → |
| NVIDIA RTX 3080 Ti | 12 GB | Ampere | ≈ 35.1 TFLOPS | 350 W | – | Faahfaahin → |
| NVIDIA A40 datacenter | 45 GB | Ampere | ≈ 29.9 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA RTX 3080 | 10 GB | Ampere | ≈ 29.5 TFLOPS | 380 W | – | Faahfaahin → |
| NVIDIA RTX 4070 | 12 GB | Ada Lovelace | ≈ 29.3 TFLOPS | 200 W | – | Faahfaahin → |
| NVIDIA L4 datacenter | 22 GB | Ada Lovelace | ≈ 24.2 TFLOPS | 72 W | – | Faahfaahin → |
| NVIDIA RTX 5060 Ti | 16 GB | Blackwell | ≈ 22.8 TFLOPS | 180 W | – | Faahfaahin → |
| NVIDIA RTX 4060 Ti | 16 GB | Ada Lovelace | ≈ 21.6 TFLOPS | 165 W | – | Faahfaahin → |
| NVIDIA RTX A4000 | 16 GB | Ampere | ≈ 20.6 TFLOPS | 140 W | – | Faahfaahin → |
| NVIDIA RTX 3070 | 8 GB | Ampere | ≈ 19.8 TFLOPS | 220 W | – | Faahfaahin → |
| NVIDIA RTX 3060 Ti | 8 GB | Ampere | ≈ 15.7 TFLOPS | 200 W | – | Faahfaahin → |
| NVIDIA A100 PCIE datacenter | 40 GB | Ampere | ≈ 15.6 TFLOPS | 250 W | – | Faahfaahin → |
| NVIDIA A100 SXM4 datacenter | 40 GB | Ampere | ≈ 15.6 TFLOPS | 400 W | – | Faahfaahin → |
| NVIDIA A800 PCIE datacenter | 80 GB | Ampere | ≈ 15.6 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA RTX 2080 Ti | 11 GB | Turing | ≈ 14.6 TFLOPS | 250 W | – | Faahfaahin → |
| NVIDIA RTX 3060 laptop | 12 GB | Ampere | ≈ 12.9 TFLOPS | 80 W | – | Faahfaahin → |
| NVIDIA Tesla V100 datacenter | 16 GB | Volta | ≈ 12.5 TFLOPS | 300 W | – | Faahfaahin → |
| NVIDIA CMP 170HX datacenter | 64 GB | Ampere | ≈ 12.1 TFLOPS | 250 W | – | Faahfaahin → |
| NVIDIA RTX 3060 | 12 GB | Ampere | ≈ 12 TFLOPS | 170 W | – | Faahfaahin → |
| NVIDIA Titan Xp | 12 GB | Pascal | ≈ 11.7 TFLOPS | 160 W | – | Faahfaahin → |
| NVIDIA RTX A2000 | 6 GB | Ampere | ≈ 11.2 TFLOPS | 70 W | – | Faahfaahin → |
| NVIDIA GTX 1080 Ti | 11 GB | Pascal | ≈ 11 TFLOPS | 250 W | – | Faahfaahin → |
| NVIDIA GTX 1080 | 8 GB | Pascal | ≈ 7.8 TFLOPS | 130 W | – | Faahfaahin → |
| NVIDIA GTX 1070 Ti | 8 GB | Pascal | ≈ 7.4 TFLOPS | 180 W | – | Faahfaahin → |
| NVIDIA GTX 1650 | 4 GB | ≈ undefined TFLOPS | W | – | Faahfaahin → | |
| NVIDIA Q RTX 8000 | 45 GB | Turing | ≈ 11.9 TFLOPS | 250 W | – | Faahfaahin → |
| NVIDIA Tesla P40 datacenter | 24 GB | Pascal | ≈ 9.4 TFLOPS | 250 W | – | Faahfaahin → |
| NVIDIA RTX 2070S | 8 GB | Turing | ≈ 8.8 TFLOPS | 210 W | – | Faahfaahin → |
| NVIDIA Q RTX 4000 | 8 GB | Turing | ≈ 7.7 TFLOPS | 125 W | – | Faahfaahin → |
| NVIDIA RTX 2070 | 8 GB | Turing | ≈ 7.7 TFLOPS | 150 W | – | Faahfaahin → |
| NVIDIA RTX 2060S | 8 GB | Turing | ≈ 7.5 TFLOPS | 175 W | – | Faahfaahin → |
| NVIDIA GTX TITAN X | 12 GB | Maxwell | ≈ 6.8 TFLOPS | 150 W | – | Faahfaahin → |
| NVIDIA RTX 2060 | 12 GB | Turing | ≈ 6.5 TFLOPS | 130 W | – | Faahfaahin → |
| NVIDIA GTX 1660 Ti | 6 GB | Turing | ≈ 5.3 TFLOPS | 120 W | – | Faahfaahin → |
| NVIDIA GTX 1660 S | 6 GB | Turing | ≈ 3.6 TFLOPS | 125 W | – | Faahfaahin → |
GPU templates loogu talagalay howlaha AI, LLM iyo rendering
Adeegsiga LLM (inference)
vLLM, SGLang ama Ollama oo leh Open WebUI si loo adeegsado loona tijaabiyo noocyada luuqadda (language models).
Dhisidda iyo hagaajinta (Fine-tuning)
Unsloth Studio si aad ugu hagaajiso noocyada luuqadda xogtaada gaarka ah.
Soo saarista sawirrada iyo muuqaallada
ComfyUI ee loogu talagalay Flux, SDXL iyo socodka shaqada ee muuqaallada (video workflows).
Horumarinta PyTorch iyo CUDA
Sawirrada Ubuntu 24.04 iyo PyTorch oo wata NVIDIA CUDA, lagana geli karo SSH.
Desktop ama VM buuxa
Ubuntu desktop toos uga dhex shaqeynaya biraawsarka, ama mashiin dalwaddaysan (VM) oo buuxa oo KVM ah markaad u baahato.
Sida loo kireysto seerfar GPU 4 tallaabo gudahood
- 01
Dooro template iyo cabbirka disk-ga
- 02
Ka dooro GPU ku habboon kaydka tooska ah ee diyaar u ah
- 03
Xaqiiji qiimaha oo ka sameyso server-ka hadhaaga akoonkaaga
- 04
Kula xiriir SSH ama terminal-ka biraawsarka
Qiimaha iyo qaansheegadka GPU cloud
- Saacadle: GPU-ga waxaa lacag laga qaadaa saacaddii inta server-ku shaqeynayo. Server-ka saacadlaha ah ee la damiyey waxaa kaliya oo lacag laga qaadaa disk-giisa.
- Wareegyada biil bixinta: bishii, saddexdii biloodba, lixdii biloodba ama sanadkii, iyadoo qiimo dhimis kasta la muujinayo ka hor inta aadan bixin.
- Kaydinta disk-ga waxaa lacag looga qaadaa halkii GB iyadoo la socota qiimaha GPU-ga.
- Qiimuhu waa USD; waxaad ku muujin kartaa VND marka loo eego sarifka lacagaha ee hadda jira.
Su'aalaha Badanaa La Isweydiiyo ee GPU cloud
Hawl kasta oo ku filan xusuusta GPU-ga: LLM inference iyo fine-tuning, samaynta sawirrada iyo fiidiyowyada, rendering iyo horumarinta CUDA. Templates-ka PyTorch, vLLM, SGLang, Ollama, Unsloth iyo ComfyUI ayaa softiweerka kuu rakibaya.
Shaxda sare waxay muujinaysaa qaabeynta hadda la heli karo, oo ay ku jirto NVIDIA RTX 5090 oo wadata 32 GB oo VRAM ah marka ay kaydka ku jirto. Helitaanku wuu isbeddelaa maalinta gudaheeda.
Qiimuhu wuxuu ka yimaadaa kaydka tooska ah ee GPU waxaana laga yaabaa inuu isbeddelo. Console-ku wuxuu muujinayaa qiimaha la xaqiijiyay intaadan dalban, waxaana seerfarkaaga lagu dalacayaa qiimahaas la xaqiijiyay.
Haa. Ka sokow biilka saacadlaha ah, waxaad dooran kartaa wareeg bille, saddex-biloodle, lix-biloodle ama sannadle ah, iyadoo qiimo-dhimis kasta la muujinayo intaadan bixin. Kaydka disk-ga waxaa loo xisaabiyaa halkii GB oo lagu daray qiimaha GPU-ga.
Biilka saacadlaha ah, seerfarka dansan waxaa laga qaadaa lacagta disk-giisa oo keliya. Tirtir seerfarka marka aadan mar dambe u baahnayn xogtiisa.
Adigoo u maraya SSH iyo furahaaga, ama terminal-ka biraawsarkaaga ee toos uga dhex shaqeeya console-ka. Desktop iyo VM templates waxay sidoo kale bixiyaan galaangal muuqaal ah (graphical access).