Graphics accelerators (GPUs)
Compute Cloud provides graphics accelerators (GPUs) for different VM configurations. GPUs outperform CPUs in terms of processing certain types of data and are suitable for machine learning (ML), artificial intelligence (AI), and 3D rendering tasks. You can manage GPUs and RAM directly from your VM.
In Compute Cloud, you can use the following GPUs:
GPUs run in TCC mode, which does not use the operating system's graphics drivers.
The NVIDIA A100 GPU based on the Ampere® microarchitecture. It uses third-generation Tensor Cores and delivers 80 GB HBM2 memory with up to 2 TB/s throughput.
VMs with NVIDIA® Tesla® T4 GPUs may only reside in the
il1-c availability zones.
NVIDIA H100 GPU is powered by the NVIDIA Hopper architecture. It features fourth-generation Tensor Cores and has the capability to apply mixed FP8 and FP16 precisions to dramatically accelerate AI calculations for transformers – up to 4x faster training and up to 30x faster inference over the prior generation (NVIDIA A100). NVIDIA H100 offers 80 GB of HBM3 RAM and up to 3.35 TB/s memory bandwidth.
For more information, go to the NVIDIA website:
Configurations of VMs with GPUs
The configurations available for VMs with GPUs are listed below.
AMD EPYC™ with NVIDIA® Ampere® A100 (
Number of GPUs VRAM, GB Number of vCPUs RAM, GB 1 80 28 119 2 160 56 238 4 320 112 476 8 640 224 952
Intel Ice Lake with NVIDIA® Tesla® T4 (
Number of GPUs VRAM, GB Number of vCPUs RAM, GB 1 16 8 32 1 16 16 64 1 16 32 128
Intel Sapphire Rapids with NVIDIA® Hopper® H100 (
Number of GPUs VRAM, GB Number of vCPUs RAM, GB 1 80 20 160 2 160 40 320 4 320 80 640 8 640 160 1280
VM GPUs are provided in full. For example, if a configuration has four GPUs specified, your VM will have four full-featured GPU devices.
For more information about organizational and technical limitations for VMs, see Quotas and limits.
For VMs with GPUs, you can use products with pre-installed NVIDIA drivers available from Cloud Marketplace, install the drivers on another standard image yourself, or create a custom image with pre-installed drivers.