Instances›AWS EC2›p4d›p4d.24xlarge
Compute
vCPUs96
Memory1152 GiB
Mem / vCPU12.00 GiB
ProcessorXeon Platinum 8275L
Clock speed3 GHz
Architecturex86_64
Storage
Type8 × 1,000 GB NVMe SSD
Bandwidth19,000 Mbps
Throughput2375 MB/s
IOPS80,000
Networking
Performance4x 100 Gigabit
Max ENIs60
IPv6Yes
GPUNVIDIA A100
Family sizes
p4d.24xlarge96 vCPU · 1152 GiB
p4d.24xlarge

The p4d.24xlarge is a GPU AWS EC2 instance: 96 vCPU and 1152 GiB of memory, 12.00 GiB per vCPU. In the name, "p" is accelerated computing, "4" is the generation, "d" marks local NVMe storage and "24xlarge" is the size within the p4d family. It runs on Intel Xeon Platinum 8275L at 3 GHz, and AWS introduced it in November 2020. AWS has since shipped p5. Rackspace Spot has no equivalent to offer: 8 × NVIDIA A100 wired together behind a high-speed interconnect is one tightly coupled machine, not 8 independent nodes — and that interconnect is usually what the instance was bought for. Spot's largest VM tops out at 16 vCPU and 128 GiB, and even that is not the missing piece here. The table below prices what AWS charges; the honest answer on the Spot side is that this workload stays where it is.

Match qualityNo equivalent

p4d.24xlarge is 8 Ampere, Ampere, Ampere, Ampere, Ampere, Ampere, Ampere, Ampere GPUs (320 GiB VRAM) wired together inside one chassis. The interconnect is what you bought — that is what makes it one accelerator with pooled memory rather than several machines — and Rackspace Spot has nothing that reproduces it. The Spot GPU class is built on the NVIDIA A30 (Ampere, 24 GB VRAM), and separate nodes of it on ordinary datacentre networking are not a substitute for a single interconnect-joined node.

Split the job across independent single-GPU nodes only if it is data-parallel with no per-step synchronisation and the model fits one card. Model-parallel work, or anything synchronising gradients every step, does not survive the move. Rackspace Spot sells x86_64 Kubernetes nodes up to 16 vCPU, plus one NVIDIA A30 GPU class.

Pricing by region
1 hr
RegionAWS on-demandAWS Spot1-yr Reserved3-yr ReservedRackspace SpotYou save
US East, N. Virginia$21.958$17.084$13.921$9.374
US Central, Chicagono AWS region
US Central, Dallasno AWS region
Hong Kong
Sydney, Australia$28.545$10.711$18.097$12.187
London, UK$28.545$10.052
US West, San Jose
US Central, Dallas 2no AWS region
US East, N. Virginia 2$21.958$17.084$13.921$9.374
When to use it
Good for
  • ●Remote graphics workstations and rendering
  • ●CUDA-accelerated batch jobs
  • ●Model inference and embedding generation
  • ●Workloads that fit inside 1152 GiB of RAM
  • ●Jobs that want 8 × 1,000 GB of local NVMe scratch space
  • ●Network-heavy services — AWS rates this size at 4x 100 Gigabit
Avoid if
  • ●The working set needs more than 1152 GiB of RAM
  • ●The work is single-threaded. Most of these 96 vCPU would sit idle.
  • ●Anything the CPU can do fast enough. GPU hours are the most expensive on the price list.
  • ●More than 12.00 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