C4 VMs now GA: Unmatched performance and control for your enterprise workloads

Today, we're excited to announce the general availability of the C4 machine series, the most performant general-purpose VM for Compute Engine and Google Kubernetes Engine (GKE) customers. C4 VMs are engineered from the ground up and fine-tuned to deliver industry-leading performance, with up to 20% better price-performance for general-purpose workloads, and 45% better price-performance for CPU-based inference versus comparable generally available VMs from other leading hyperscalers. C4 also provides enterprise-grade reliability and granular control to support customers' most business-critical workloads.

Together with the N4 machine series, C4 VMs provide the performance and flexibility you need to handle a majority of workloads, all powered by Google’s Titanium. With Titanium offload technology, C4 provides high performance connectivity with up to 200 Gbps of networking bandwidth and scalable storage with up to 500k IOPS and 10 GB/s throughput on Hyperdisk Extreme. C4 instances scale up to 192 vCPUs and 1.5TB of DDR5 memory and feature the latest generation performance with Intel’s 5th generation Xeon processors (code-named Emerald Rapids) offering predefined shapes in high-cpu, standard, and high-mem configurations.

Whether you're running powerful databases, supporting high-traffic game servers or media streaming, running CPU-based inference, or handling massive data processing and analytics, C4 VMs provide the performance, features, and stability to meet your business needs.

“Google Cloud's C4 VMs deliver a significant 61% increase in single-thread performance versus N2. For our workloads, this translates to faster build times and more efficient deployments, empowering tens of thousands of developers to innovate and iterate at an unprecedented pace.” - Will Stewart, CEO & Co-founder, Northflank

Industry-leading Performance

C4 VMs outperform comparable GA offerings from other cloud providers across a wide range of workloads, delivering up to 20% better price-performance on the estimated SPECrate®2017_int_base benchmark enabling you to optimize your demanding workloads to get the most out of your cloud resources.

For database workloads, C4 VMs benefit from the latest innovations with Google Cloud’s high-performance Hyperdisk storage and Titanium networking, demonstrating up to 2x more transactions per minute on MySQL and 2x better performance for throughput-focused Redis workloads so you can generate significantly more queries per second with each VM. If you’re doing data analytics, C4's performance improvements translate to faster time-to-insights, enabling you to process and analyze your data more efficiently. For gaming and streaming, the improvements in compute and network latency lead to smoother gameplay, reduced buffering, and an overall enhanced user experience.

Responsiveness, predictability, and control

C4 delivers notable improvements in responsiveness, predictability, and consistency. This makes C4 a great fit for applications where every millisecond matters, from high-frequency trading and gaming, to latency-sensitive mission-critical enterprise systems that require consistent performance and uptime.

With Titanium, C4 offers improved infrastructure performance, lifecycle management, reliability, and security. Storage is offloaded to the Titanium adapter, reserving the host resources for running your workloads. And by leveraging hitless upgrades and live migration, the vast majority of infrastructure maintenance updates are performed with near-zero impact to your workloads, minimizing disruptions and providing predictable performance. For real-time workloads, C4 offers up to 80% better CPU responsiveness compared to previous generations, resulting in faster trades and a smoother gaming and streaming experience. 

C4’s single-core max turbo frequency is the highest of any prior-generation VM, shooting up to 4.0 GHz to accelerate single-threaded workloads like financial and Monte Carlo simulations, and ML tasks like recommendation systems, image classification, and SLM training, tuning, and inferencing. C4 VMs are also the first to support Compute Engine’s new frequency control feature, the Turbo API, currently in preview. Turbo API allows you to turn frequency boost on or off, giving you fine-grained control over performance, and enabling you to balance between top performance and consistency. Workloads like batch processing or scientific simulations can benefit from being able to boost to the highest possible frequency, while web servers and databases may benefit from being capped at all-core turbo frequency (3.1 GHz) to prioritize consistent performance.

“Our real-time bidding platform relies on split-second decision-making and high throughput request handling. We tested C4 VMs with real production traffic and observed up to 2.5x improvement over N2 instances for our real-time bidding platform, helping us ensure faster bid processing, minimize latency, and maximize the effectiveness of customer ad campaigns.” - Alexander Zeikidze, Senior DevOps Engineer, Digital Turbine

Accelerating CPU-based Inference

C4 VMs represent a major leap forward in performance and cost-efficiency for machine learning inference compared to previous generations. C4 supports Intel's latest built-in accelerator, Advanced Matrix Extensions (AMX), which enables more throughput to significantly speed up CPU-based machine learning inference. For recommendation inference, C4 VMs demonstrate up to 40% price-performance uplift compared to C3, and up to 45% better price-performance versus the closest competitive offering. Compared to GPUs, C4 VMs offer greater deployment flexibility across workload types and provide a familiar operational experience.

This translates to more powerful and affordable inference capabilities, enabling customers to leverage C4 VMs to optimize their small-to-medium machine learning deployments without having to compromise on efficiency.

Next-level storage with Hyperdisk

C4 VMs come standard with the latest Hyperdisk storage technology, now available in Hyperdisk Balanced and Hyperdisk Extreme options. Hyperdisk Balanced delivers a blend of performance and cost-efficiency for a wide range of workloads. Hyperdisk Extreme pushes the boundaries of storage performance, delivering ultra-low latency and extreme throughput for the most demanding applications. With Titanium, C4 VMs unlock the full potential of Hyperdisk, achieving scalable, high-performance I/O up to 500k IOPS and 10 GB/s throughput on Hyperdisk Extreme.

Hyperdisk's benefits extend beyond raw performance. With Hyperdisk you can dynamically tune IOPS and bandwidth in real time, so your workloads have the resources they need. Hyperdisk storage pools (for Hyperdisk Balanced volumes) enable flexible management and allocation of storage resources, optimizing utilization and cost-efficiency.

Hyperdisk not only provides superior performance and flexibility but also significantly reduces your total cost of ownership (TCO) compared to traditional storage solutions. The combination of Hyperdisk's advanced features and Titanium's storage acceleration delivers an unparalleled storage experience for your C4 VMs.

Powering the future together

Selecting the right VM is crucial for your success, and can be complex. Together, the C4 and the previously announced N4 VMs cover the full spectrum of your general-purpose workloads, allowing you to lower costs while meeting your performance and workload requirements. For workloads that need top-tier performance, precise control and predictability, or advanced features, C4 VMs are the ideal choice. For everything else, N4 delivers exceptional value and an excellent balance of performance, cost-efficiency, and flexibility.

“C4 and N4 VMs have demonstrated significant performance improvements, with an increase of over 15% compared to the previous generation of VMs. We’re excited to use them both as the foundation of our computing infrastructure. The combination of C4's performance and N4's flexibility enables us to scale our systems while meeting customer demands for reliability and stability.” - Christian Tüffers, Director – IT Architect, Deutsche Börse

“We are excited about Google Cloud’s new C4 and N4 machines. The new C4 VMs show 35% performance boost for real-time ML threat detection and N4 VMs provide 20% higher throughput for Dynamic Analysis of Malware samples at a lower cost vs. the previous generation.” - Xu Zou, VP of Product, CDSS, Palo Alto Networks

Get started with C4 VMs today

C4 VMs are available now in us-central1 (Iowa), us-east1 (S. Carolina), us-east4 (North Virginia), europe-west4 (Netherlands), europe-west1 (Belgium), and asia-southeast1 (Singapore). We invite you to explore C4’s unparalleled performance, reliability, and control, and experience the dynamite future of cloud computing.

“Google again propels the industry forward with its C4 and N4 VM innovations, featuring Intel’s latest 5th Gen Intel Xeon for customers. C4 currently offers the fastest multi-purpose Intel instance generally available from leading hyperscalers. These VMs offer customers unmatched performance, improved TCO, and enhanced features to fuel their most demanding workloads. Additionally, thanks to advancements from built-in AI optimizations like Intel AMX and other unique Intel Xeon processor capabilities, customers are experiencing remarkable acceleration for their inference workloads on C4.” - Jason Kennedy, GM & Sr. Director of Xeon Product Management, Intel DCAI

To start using C4 instances, select C4 under the General Purpose machine family when creating a new VM or GKE node pool in the Cloud console. Stay up-to-date on regional availability by visiting our regions and zones page or contact your Google Cloud sales representative for more information.