We launched the first Graviton-powered EC2 instances in late 2018 and announced the follow-on Graviton2 processor just a year later. The dual SIMD units, support for int8 and fp16 instructions, and other architectural improvements between generations combine to make the Graviton2 a highly cost-effective workhorse processor. Today, you can choose between General Purpose (M6g and […]
We launched the first Graviton-powered EC2 instances in late 2018 and announced the follow-on Graviton2 processor just a year later. The dual SIMD units, support for int8 and fp16 instructions, and other architectural improvements between generations combine to make the Graviton2 a highly cost-effective workhorse processor.
Today, you can choose between General Purpose (M6g and M6gd), Compute-Optimized (C6g, C6gn, and C6gd), Memory-Optimized (R6g and R6gd), and Burstable (T4g) instances, all powered by fast, efficient Graviton2 processors. Our customers use these instances to run application servers, gaming servers, HPC workloads, video encoding, ad servers, and more. Multiple benchmarks (including this one and this one) have shown that these Graviton2-based instances deliver better price-performance than existing EC2 instances.
New X2gd Instances
I’m happy to announce the latest in our ever-growing roster of Graviton2-powered instances! The new X2gd instances have twice as much memory per vCPU as the memory-optimized R6g instances, and are designed for your memory-hunger workloads. This includes in-memory databases (Redis and Memcached), open source relational databases, Electronic Design Automation design & verification, real-time analytics, caching services, and containers.
X2gd instances are available in eight sizes, and also in bare metal form. Here are the specs:
|Local NVMe Storage||Network Bandwidth
|x2gd.medium||1||16||1 x 59 GiB||Up to 10||Up to 4.750|
|x2gd.large||2||32||1 x 118 GiB||Up to 10||Up to 4.750|
|x2gd.xlarge||4||64||1 x 237 GiB||Up to 10||Up to 4.750|
|x2gd.2xlarge||8||128||1 x 475 GiB||Up to 10||Up to 4.750|
|x2gd.4xlarge||16||256||1 x 950 GiB||Up to 10||4.750|
|x2gd.8xlarge||32||512||1 x 1900 GiB||12||9.500|
|x2gd.12xlarge||48||768||2 x 1425 GiB||20||14.250|
|x2gd.16xlarge||64||1024||2 x 1900 GiB||25||19.000|
|x2gd.metal||64||1024||2 x 1900 GiB||25||19.000|
When compared to the existing X1 instances, the new X2gd instances offer 55% better price/perfomance. The X2gd instances also offer the lowest price per GiB of memory of any current EC2 instance.
On the compute side, the X2gd instances provide the same amount of CPU power as the other Graviton2-powered instances (M6g, C6g, R6g, and T4g). Each vCPU is an entire physical core, which speeds compute-heavy EDA and financial service workloads, and also encourages denser packing of containers onto an instance of a particular size.
In addition to the fast SSD-based local NVMe storage, X2gd instances support Elastic Network Adapter (ENA), and can deliver up to 25 Gbps of network bandwidth between instances when launched within a Placement Group.
X2gd instances are built on AWS Nitro System, and you can use your existing Arm-compatible EC2 AMIs.
X2gd instances are available now in the US East (N. Virginia), US East (Ohio), US West (Oregon), and Europe (Ireland) Regions, with more regions to come. You can launch them today in On-Demand and Spot form, and you can purchase Savings Plans or Reserved Instances.