Instances›AWS EC2›g5›g5.12xlarge
Compute
vCPUs48
Memory192 GiB
Mem / vCPU4.00 GiB
ProcessorEPYC 7R32
Clock speed2.8 GHz
Architecturex86_64
Storage
Type1 × 3,800 GB NVMe SSD
Bandwidth16,000 Mbps
Throughput2000 MB/s
IOPS65,000
Networking
Performance40 Gigabit
Max ENIs15
IPv6Yes
GPUNVIDIA A10G
Family sizes
g5.xlarge4 vCPU · 16 GiB
g5.2xlarge8 vCPU · 32 GiB
g5.4xlarge16 vCPU · 64 GiB
g5.8xlarge32 vCPU · 128 GiB
g5.12xlarge48 vCPU · 192 GiB
g5.16xlarge64 vCPU · 256 GiB
g5.24xlarge96 vCPU · 384 GiB
g5.48xlarge192 vCPU · 768 GiB
g5.12xlarge

The g5.12xlarge pairs 48 vCPU with 192 GiB of RAM, the 4.00 GiB-per-vCPU ratio that marks a GPU instance. In the name, "g" is graphics, "5" is the generation and "12xlarge" is the size within the g5 family. It runs on AMD EPYC 7R32 at 2.8 GHz, and AWS introduced it in November 2021. AWS has since shipped g6 and g7; g6.12xlarge is the direct replacement at this size, scoring 1,056,861 CoreMark/s against this instance's 895,147 — roughly 18% more throughput for the same vCPU count. Rackspace Spot runs one GPU class, gpu.vs2.xlargeplusplus — a single NVIDIA A30 with 24 vCPU and 64 GiB, in San Jose only. Against this instance's 4 × NVIDIA A10G that is a downgrade per card, honest for inference and not for the training this instance was bought for.

Match qualityWeaker GPU

g5.12xlarge puts 4 Ampere, Ampere, Ampere, Ampere GPUs (96 GiB VRAM) in one machine. Spot has a single GPU class and no multi-GPU node, so there is no shape here that places 4 cards behind one host. Per card the NVIDIA A30 (Ampere, 24 GB VRAM) is a step down from what g5.12xlarge carries, and the aggregate is not reachable at all.

That makes this honest for data-parallel inference and embarrassingly parallel batch work, where each card is independent, and dishonest for anything that shards a model across cards or synchronises every step. The catch is location rather than hardware: US West, San Jose is the only Spot region that offers it, so latency to your users and egress to your data are the things to check.

No local diskg5.12xlarge comes with 1 × 3,800 GB NVMe SSD attached to the machine. Spot nodes use network storage only, so that disk doesn't come with it. If you were using it as fast scratch space, you'd move to a persistent volume or object storage, which are slower to read.
Pricing by region
1 hr
RegionAWS on-demandAWS Spot1-yr Reserved3-yr ReservedRackspace SpotYou save
US West, San Jose$0.191
When to use it
Good for
  • ●Model inference and embedding generation
  • ●Video transcoding and image processing
  • ●Machine learning experimentation and fine-tuning
  • ●Workloads that fit inside 192 GiB of RAM
  • ●Jobs that want 1 × 3,800 GB of local NVMe scratch space
  • ●Network-heavy services — AWS rates this size at 40 Gigabit
Avoid if
  • ●Starting a new deployment. g6.12xlarge is the newer generation at this size.
  • ●The working set needs more than 192 GiB of RAM
  • ●The work is single-threaded. Most of these 48 vCPU would sit idle.
  • ●Anything the CPU can do fast enough. GPU hours are the most expensive on the price list.
  • ●More than 4.00 GiB of RAM per vCPU is needed — that ratio is fixed for the whole family
Rackspace Spot sells spare capacity at market prices. You pay what the market decides, not a list price.
  • Spot sells spare capacity at market prices. It is the cheapest compute on the internet.
  • Prices are set by supply and demand. You pay what the market decides.
  • Compute is delivered as fully managed Kubernetes. All infrastructure is taken care of.
  • Egress costs $0.01/GB after the free allowance. AWS charges $0.09/GB.
Build your Cloudspace
Frequently asked questions