Kosmik
Honorary Master
NV-series
N-series virtual machines are ideal for compute and graphics-intensive workloads, helping customers to fuel innovation through scenarios like high-end remote visualization, deep learning, and predictive analytics. N-series virtual machines feature the NVIDIA Tesla accelerated platform and NVIDIA RTX Enterprise technology, providing the highest-end graphics support available in the cloud today.| Instance | vCPU(s) | RAM | Temporary storage | GPU | Pay as you go | 1 year savings plan | 3 year savings plan | Spot | Add to estimate |
|---|---|---|---|---|---|---|---|---|---|
| NV6 | 6 | 56 GiB | 340 GiB | 1X M60 | $996.4500/month | $829.4479/month ~16% savings | $614.7111/month ~38% savings | $99.6450/month ~89% savings | |
| NV12 | 12 | 112 GiB | 680 GiB | 2X M60 | $1,992.9000/month | $1,658.8885/month ~16% savings | $1,229.4222/month ~38% savings | $199.2900/month ~89% savings | |
| NV24 | 24 | 224 GiB | 1,440 GiB | 4X M60 | $3,985.8000/month | $3,317.7770/month ~16% savings | $2,458.8371/month ~38% savings | $398.5800/month ~89% savings |
The Azure NV6, NV6_Promo, NV12, NV12_Promo, NV24, NV24_Promo Virtual Machine sizes will be retired on 31 August 2023. Refer to our migration guide for help on next steps.
NVv3-series
NVv3 series virtual machines are powered by NVIDIA Tesla M60 GPUs and NVIDIA RTX Enterprise technology with Intel Broadwell CPUs. These virtual machines are targeted for GPU accelerated graphics applications and virtual desktops where customers want to visualize their data, simulate results to view, work on CAD or render and stream content. Additionally, these virtual machines can run single precision workloads such as encoding and rendering. NVv3 virtual machines support Premium Storage and come with twice the system memory (RAM) when compared with its predecessor NV Series. Compared to the NVv2 series the NVv3 configurations are hyperthreaded.| Instance | vCPU(s) | RAM | Temporary storage | GPU | Pay as you go | 1 year savings plan | 3 year savings plan | Spot | Add to estimate |
|---|---|---|---|---|---|---|---|---|---|
| NV12s v3 | 12 | 112 GiB | 736 GiB | 1X M60 | $1,040.9800/month | $813.4244/month ~21% savings | $642.1810/month ~38% savings | $104.0980/month ~89% savings | |
| NV24s v3 | 24 | 224 GiB | 1,474 GiB | 2X M60 | $2,080.5000/month | $1,625.7027/month ~21% savings | $1,283.4568/month ~38% savings | $208.0500/month ~89% savings | |
| NV48s v3 | 48 | 448 GiB | 2,948 GiB | 4X M60 | $4,161.0000/month | $3,251.4054/month ~21% savings | $2,566.9209/month ~38% savings | $416.1000/month ~89% savings |
NVads A10 v5 series
The NVads A10 v5 is based on the Nvidia A10 GPU and optimized for graphics and visualization workloads. The NVads A10 v5 VM-series provides GPU resourcing flexibility and allows you to choose VMs starting with 1/6th of GPU and scale to multi-GPU configurations. The GPU partitioning technology combined with the powerful AMD Milan processors and up to 880GB of memory, provides right sized GPU VMs from entry level VDI workload to high end graphics workload requiring more vCPU, Memory and GPU resources.| Instance | vCPU(s) | RAM | Temporary storage | GPU | Pay as you go | 1 year savings plan | 3 year savings plan | Spot | Add to estimate |
|---|---|---|---|---|---|---|---|---|---|
| NV6adsv5 | 6 | 55 GiB | 180 GiB | 1/6X A10 | $430.7000/month | $358.5176/month ~16% savings | $265.6981/month ~38% savings | $107.6750/month ~75% savings | |
| NV12ads A10 v5 | 12 | 110 GiB | 320 GiB | 1/3X A10 | $861.4000/month | $717.0279/month ~16% savings | $531.3962/month ~38% savings | $215.3500/month ~75% savings | |
| NV18ads A10 v5 | 18 | 220 GiB | 720 GiB | 1/2X A10 | $1,518.4000/month | $1,263.9147/month ~16% savings | $936.6995/month ~38% savings | $379.6000/month ~75% savings | |
| NV36ads A10 v5 | 36 | 440 GiB | 720 GiB | 1X A10 | $3,036.8000/month | $2,527.8294/month ~16% savings | $1,873.3990/month ~38% savings | $759.2000/month ~75% savings | |
| NV36adms A10 v5 | 36 | 880 GiB | 720 GiB | 1X A10 | $4,289.4800/month | $3,570.5614/month ~16% savings | $2,646.1770/month ~38% savings | $1,072.3700/month ~75% savings | |
| NV72ads A10 v5 | 72 | 880 GiB | 1,400 GiB | 2X A10 | $6,187.4800/month | $5,150.4566/month ~16% savings | $3,817.0532/month ~38% savings | $1,546.8700/month ~75% savings |
NDs-series
NDs-series virtual machines is a new addition to the GPU family specifically designed for the needs of the AI and deep learning workloads offering excellent performance for training and inference. It’s powered by NVIDIA Tesla P40 GPUs based on the new Pascal Architecture. These instances provide excellent performance for FP32 (single precision floating point operations), for AI workloads utilizing CNTK, TensorFlow, Caffe, and other frameworks. The NDs-series also offers a much larger GPU memory size (24 GB), enabling it to fit much larger neural net models. Like our NC-series, the NDs-series offers a configuration with a secondary low-latency, high-throughput network through RDMA, and InfiniBand connectivity so you can run large-scale training jobs spanning many GPUs. In addition, N-series offers an NDs24r configuration that provides a low latency, high-throughput network interface optimized for tightly coupled parallel computing workloads.| Instance | Core(s) | RAM | Temporary storage | GPU | Pay as you go | 1 year savings plan | 3 year savings plan | Spot | Add to estimate |
|---|---|---|---|---|---|---|---|---|---|
| ND6s | 6 | 112 GiB | 736 GiB | 1X P40 | $1,957.8600/month | $1,629.7250/month ~16% savings | $1,207.8069/month ~38% savings | $223.3917/month ~88% savings | |
| ND12s | 12 | 224 GiB | 1,474 GiB | 2X P40 | $3,914.9900/month | $3,258.8368/month ~16% savings | $2,415.1539/month ~38% savings | $446.7001/month ~88% savings | |
| ND24rs | 24 | 448 GiB | 2,948 GiB | 4X P40 | $8,616.9200/month | $7,172.7245/month ~16% savings | $5,315.7797/month ~38% savings | $983.1903/month ~88% savings | |
| ND24s | 24 | 448 GiB | 2,948 GiB | 4X P40 | $7,830.7100/month | $6,518.2795/month ~16% savings | $4,830.7677/month ~38% savings | $893.4842/month ~88% savings |
NDv2 series
NDv2-series virtual machine is the latest, fastest and most powerful addition to the GPU family specifically designed for the cutting edge demands of distributed HPC, AI and machine learning workloads. The VMs feature 8 NVIDIA V100 Tensor Core GPUs with NVIDIA NVLink interconnected GPUs with 32 GB of memory each, 40 non-hyperthreaded Intel Xeon Platinum 8168 processor cores, and 672 GiB of system memory. The NDv2-series VMs also feature 100 Gb/sec EDR InfiniBand with support for standard Mellanox OFED drivers and all MPI types and versions. With total of 256 GB of GPU memory and 100 Gb/sec InfiniBand interconnect NDv2-series VMs allowing for easy scale-up to supercomputer-sized clusters that can run the most demanding ML models and distributed AI training workloads utilizing Cuda, TensorFlow, Pytorch, Caffe, and other Frameworks.| Instance | Core(s) | RAM | Temporary storage | GPU | Pay as you go | 1 year savings plan | 3 year savings plan | Spot | Add to estimate |
|---|---|---|---|---|---|---|---|---|---|
| ND40rs v2 | 40 | 672 GiB | 2,900 GiB | 8X V100 (NVlink) | $20,093.9800/month | $16,726.2272/month ~16% savings | $12,395.9767/month ~38% savings | $4,541.2395/month ~77% savings |
