Nvme

2 posts

cloudflare3 min readCurated summary

Inside Gen 13- how we built our most powerful server yet

Cloudflare’s Gen 13 server is a hardware redesign aligned with its Rust-based FL2 request-processing stack. By choosing a 192-core AMD EPYC Turin 9965, doubling memory, expanding storage and networking, and adding stronger security and accelerator support, Cloudflare targets up to twice the throughput of Gen 12 while remaining within latency limits. The design also improves efficiency, rack density, and operational simplicity. ## Gen 13 at a Glance - Uses a 2U, single-socket design. - Key specifications: - 192-core AMD EPYC 9965 processor - 768 GB of DDR5-6400 memory - Three 7.68 TB E1.S PCIe 5.0 NVMe drives - Dual 100 GbE OCP 3.0 networking - 1,300W Titanium-grade power supply - ASPEED AST2600 BMC and AST1060 hardware root of trust - Compared with Gen 12, Gen 13 provides: - Up to 2× throughput - Up to 50% better performance per watt - Up to 60% more throughput per rack at the same power budget - Twice the memory capacity, 1.5× the storage, and 4× the network bandwidth - PCIe encryption in addition to memory encryption - Better support for high-heat PCIe accelerators ## Choosing the CPU - Gen 12 used a 96-core AMD EPYC Genoa-X 9684X with: - 400W TDP - 1,152 MB of L3 cache - Cloudflare evaluated three Turin processors: - 9755: strongest per-core performance - 9845: lower socket power and fewer cores - 9965: highest core count and better efficiency - The Turin 9965 was selected with 192 cores and 384 threads, doubling Gen 12’s hardware threads. - Its L3 cache is much smaller—384 MB total, or 2 MB per core versus Gen 12’s 12 MB per core—but FL2 workloads depend less on large L3 caches than the previous FL1 stack. - FL2 scales nearly linearly with additional cores, allowing the 9965 to deliver up to 100% higher throughput. - Production testing showed the 9965 achieved the best aggregate requests per second and favorable performance per watt at its 500W TDP. - Higher compute density also means fewer servers to provision, patch, monitor, and operate. - Turin’s support for DDR5-6400, PCIe 5.0, and CXL 2.0 provides a longer upgrade and security-support runway. ## Memory Bandwidth and Capacity - Gen 13 doubles memory from 384 GB to 768 GB while retaining 4 GB per core. - All twelve memory channels are populated using one 64 GB DDR5-6400 ECC RDIMM per channel. - This configuration delivers approximately 614 GB/s of peak memory bandwidth per socket, a 33.3% increase over Gen 12. - Using identical DIMMs across all channels enables balanced interleaving, distributing memory accesses across the full memory subsystem. - The design is intended to prevent the 192-core processor from being starved of data during highly parallel workloads. Cloudflare’s central design choice was to match hardware to the characteristics of FL2 rather than preserve Gen 12’s cache-heavy strategy. For workloads that scale well across cores, the Turin 9965 and fully populated memory system offer higher throughput, better rack economics, and simpler fleet operations.

Read original(opens in new tab)
aws2 min readCurated summary

Amazon EC2 C8id, M8id, and R8id instances with up to 22.8 TB local NVMe storage are generally available | Amazon Web Services

Amazon has generally released the EC2 C8id, M8id, and R8id instances, combining custom Intel Xeon 6 processors with up to 22.8 TB of local NVMe SSD storage. Compared with prior sixth-generation instances, they provide up to 43% more compute performance, 3.3× higher memory bandwidth, and significant gains for I/O-intensive databases and analytics. The instances target compute-heavy, balanced, and memory-intensive workloads respectively. ## Instance Families and Workloads - **C8id:** Designed for compute-intensive applications such as video encoding, image processing, and media workloads requiring fast local storage. - **M8id:** Balances compute and memory for data logging, media processing, and medium-sized data stores. - **R8id:** Targets memory-intensive workloads, including large SQL/NoSQL databases, in-memory databases, analytics, and AI inference. ## Capacity and Performance - Scale up to **96xlarge**, compared with 32xlarge in the previous generation. - Offer up to: - **384 vCPUs** - **3 TiB memory** - **22.8 TB local NVMe storage** - Available in **metal-48xl** and **metal-96xl** configurations for workloads needing direct physical-resource access. - Deliver up to: - **43% higher compute performance** - **3.3× greater memory bandwidth** - **46% better I/O-intensive database performance** - **30% faster I/O-intensive real-time analytics queries** ## Networking, Storage, and Compatibility - Support **Instance Bandwidth Configuration**, allowing network or EBS bandwidth to be increased by up to 25% depending on workload needs. - Use sixth-generation AWS Nitro cards to offload virtualization, storage, and networking operations. - Require AMIs with **ENA and NVMe drivers**; current AWS Windows and Linux AMIs include the NVMe driver by default. - Local NVMe devices appear automatically after boot and do not require block device mappings. - Storage is hardware-encrypted with **XTS-AES-256** and unique keys. - Local NVMe storage is temporary: it is destroyed when the instance is stopped or terminated. ## Availability and Purchasing - Available in US East (N. Virginia), US East (Ohio), and US West (Oregon). - R8id instances are also available in Europe (Frankfurt). - Offered as On-Demand, Savings Plans, Spot Instances, Dedicated Instances, and Dedicated Hosts. These instances are best suited to applications that can exploit high-performance, ephemeral local NVMe storage; persistent data should remain on services such as Amazon EBS.

Read original(opens in new tab)