Expanding our GPU portfolio with NVIDIA Tesla V100

Cloud-based hardware accelerators like Graphic Processing Units, or GPUs, are a great choice for computationally demanding workloads such as machine learning and high-performance computing (HPC). We strive to provide the widest selection of popular accelerators on Google Cloud to meet your needs for flexibility and cost. To that end, we’re excited to announce that NVIDIA Tesla V100 GPUs are now publicly available in beta on Google Compute Engine and Google Kubernetes Engine (GKE), and that NVIDIA Tesla P100 GPUs are now generally available.

Today’s most demanding workloads and industries require the fastest hardware accelerators. You can now select as many as eight NVIDIA Tesla V100 GPUs, 96 vCPU and 624GB of system memory in a single VM, receiving up to 1 petaflop of mixed precision hardware acceleration performance. The next generation of NVLink interconnects deliver up to 300GB/s of GPU-to-GPU bandwidth, 9X over PCIe, boosting performance on deep learning and HPC workloads by up to 40%. NVIDIA V100s are available immediately in the following regions: us-west1, us-central1 and europe-west4. Each V100 GPU is priced as low as $2.48 per hour for on-demand VMs and $1.24 per hour for Preemptible VMs. Like our other GPUs, the V100 is also billed by the second and Sustained Use Discounts apply.

Our customers often ask which GPU is the best for their CUDA-enabled computational workload. If you’re seeking a balance between price and performance, the NVIDIA Tesla P100 GPU is a good fit. You can select up to four P100 GPUs, 96 vCPUs and 624GB of memory per virtual machine. Further, the P100 is also now available in europe-west4 (Netherlands) in addition to us-west1, us-central1, us-east1, europe-west1 and asia-east1.

Our GPU portfolio offers a wide selection of performance and price options to help meet your needs. Rather than selecting a one-size-fits-all VM, you can attach our GPUs to custom VM shapes and take advantage of a wide selection of storage options, paying for only the resources you need.

* Maximum vCPU count and system memory limit on the instance might be smaller depending on the zone or the number of GPUs selected.

** GPU prices listed as hourly rate, per GPU attached to a VM that are billed by the second. Pricing for attaching GPUs to preemptible VMs is different from pricing for attaching GPUs to non-preemptible VMs. Prices listed are for US regions. Prices for other regions may be different. Additional Sustained Use Discounts of up to 30% apply to GPU on-demand usage only.

Google Cloud makes managing GPU workloads easy for both VMs and containers. On Google Compute Engine, customers can use instance templates and managed instance groups to easily create and scale GPU infrastructure. You can also use NVIDIA V100s and our other GPU offerings in Kubernetes Engine, where Cluster Autoscaler helps provide flexibility by automatically creating nodes with GPUs, and scaling them down to zero when they are no longer in use. Together with Preemptible GPUs, both Compute Engine managed instance groups and Kubernetes Engine’s Autoscaler let you optimize your costs while simplifying infrastructure operations.

LeadStage, a marketing automation provider, is impressed with the value and scale of GPUs on Google Cloud.

Chaos Group provides rendering solutions for visual effects, film, architectural, automotive design and media and entertainment, and is impressed with the speed of NVIDIA V100s on Google Cloud.

If you have computationally demanding workloads, GPUs can be a real game-changer. Check our GPU page to learn more about how you can benefit from P100, V100 and other Google Cloud GPUs!