Compare GCP Instances

Instance Typea3-megagpu-8ga4-highgpu-8g
DetailsOptimized for accelerated high performance computing workloads.

a – Accelerator optimized
3 – Generation
megagpu – Large amound of GPU memory - 80x RAM/GPU ratio
8g – Number of GPUs
Optimized for accelerated high performance computing workloads.

a – Accelerator optimized
4 – Generation
highgpu – High amound of GPU memory - 40x RAM/GPU ratio
8g – Number of GPUs
Instance FamilyAccelerator optimizedAccelerator optimized
DescriptionAccelerator Optimized: 8 NVIDIA H100 GPU, 208 vCPUs, 1872GB RAM, 16 local SSDAccelerator Optimized: 8 NVIDIA B200 GPU, 224 vCPUs, 3968GB RAM
vCPUs208224
Memory1872 GiB3968 GiB
CPU Architecturex86_64x86_64
ProcessorIntel Sapphire RapidsIntel Emerald Rapids
Shared CPUnono
GPUyesyes
GPUs88
GPU Namenvidia-h100-mega-80gbnvidia-b200
GPU Total RAM (GiB)6401440
Sustained Discounts (SUDs)nono
Sole Tenancyyesno
Nested Virtualizationnono
Custom Shapesnono
Confidential Computenono
Maximum Persistent Disks128128
Maximum Persistent Disks Size524288 GB524288 GB
Disk Interface TypeNVMeNVMe
Local SSDyesyes
Max Local SSD6000 GB12000 GB
Extreme Persistent Disksnono
Hyperdisk Extremeyesyes
Hyperdisk Throughputyesno
Network InterfacesgVNICgVNIC, MRDMA
Avg Price Diff1.00x1.59x

Regional Prices

US Dollar ($)
Per Hour
On Demand
Geography
Region
a3-megagpu-8g
a4-highgpu-8g
EuropeBelgium (europe-west1)n/an/a
EuropeFinland (europe-north1)n/an/a
HiddenHidden (hidden-1)101.5371n/a
HiddenHidden (hidden-2)103.8609n/a
HiddenHidden (hidden-3)108.8129n/a
HiddenHidden (hidden-4)115.2678n/a
HiddenHidden (hidden-5)117.5135168.5136
HiddenHidden (hidden-6)118.4216n/a
HiddenHidden (hidden-7)119.2514n/a
APACJapan (asia-northeast1)n/an/a
North AmericaOhio (us-east5)93.4091n/a
North AmericaOregon (us-west1)92.2143n/a
North AmericaVirginia (us-east4)92.2143n/a
Activate subscription to see all regions and unlock all features

Related Comparisons

Other generations of these machines worth weighing up.