g7e.12xlarge by Amazon Web Services

g7e.12xlarge is a Graphics intensive [Extra storage or memory] Gen7 12xlarge server offered by Amazon Web Services with 48 vCPUs, 512 GiB of memory and 3.8 TB of storage. The pricing starts at 2.367 USD per hour.
48 vCPU
512 GiB Memory
3.8 TB Storage
2 GPU
Spare SCore
51,809
(All-cores)
2,108
(Single-core)

Specifications

Server Metadata

Vendor ID
aws
Name
g7e.12xlarge
Description
Graphics intensive [Extra storage or memory] Gen7 12xlarge
Family
g7e
Hw Virt
Status
active
Observed At
2026-11-04T00:05:29.042959

Availability

REGION / IDSPOTONDEMAND
Northern Virgina (US) / us-east-1
4.921833 USD/h 8.2861 USD/h
Ohio (US) / us-east-2
4.159933 USD/h 8.2861 USD/h
Oregon (US) / us-west-2
7.978425 USD/h 8.2861 USD/h
Aragón (ES) / eu-south-2
2.40175 USD/h 8.7313 USD/h
Seoul (KR) / ap-northeast-2
9.1686 USD/h 10.1873 USD/h
Stockholm (SE) / eu-north-1
5.0156 USD/h 11.9533 USD/h
Tokyo (JP) / ap-northeast-1
5.7021 USD/h 12.0173 USD/h
Mumbai (IN) / ap-south-1
13.5338 USD/h 13.5338 USD/h
Frankfurt (DE) / eu-central-1
10.5524 USD/h 14.0909 USD/h
(MY) / ap-southeast-5
- 14.1956 USD/h
(TH) / ap-southeast-7
- 14.1956 USD/h
London (GB) / eu-west-2
4.1874 USD/h 14.4017 USD/h
Jakarta (ID) / ap-southeast-3
7.3252 USD/h 15.7743 USD/h

Processor

vCPUs
48
Hypervisor
nitro
CPU Allocation
Dedicated
CPU Cores
24
CPU Speed
2.4 GHz
CPU Architecture
x86_64
CPU Manufacturer
Intel
CPU Family
Xeon
CPU Model
8559C
CPU L1D Cache
48 KiB
CPU L1D Cache Total
1 MiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
768 KiB
CPU L2 Cache
2 MiB
CPU L2 Cache Total
48 MiB
CPU L3 Cache
320 MiB
CPU L3 Cache Total
320 MiB
CPU Flags
fpu, vme, de, pse, tsc, msr, pae, mce, cx8, apic, sep, mtrr, pge, mca, cmov, pat, pse36, clflush, mmx, fxsr, sse, sse2, ss, ht, syscall, nx, pdpe1gb, rdtscp, lm, constant_tsc, arch_perfmon, rep_good, nopl, xtopology, nonstop_tsc, cpuid, aperfmperf, tsc_known_freq, pni, pclmulqdq, monitor, ssse3, fma, cx16, pdcm, pcid, sse4_1, sse4_2, x2apic, movbe, popcnt, tsc_deadline_timer, aes, xsave, avx, f16c, rdrand, hypervisor, lahf_lm, abm, 3dnowprefetch, cpuid_fault, ssbd, ibrs, ibpb, stibp, ibrs_enhanced, fsgsbase, tsc_adjust, bmi1, avx2, smep, bmi2, erms, invpcid, avx512f, avx512dq, rdseed, adx, smap, avx512ifma, clflushopt, clwb, avx512cd, sha_ni, avx512bw, avx512vl, xsaveopt, xsavec, xgetbv1, xsaves, avx_vnni, avx512_bf16, wbnoinvd, ida, arat, avx512vbmi, umip, pku, ospke, waitpkg, avx512_vbmi2, gfni, vaes, vpclmulqdq, avx512_vnni, avx512_bitalg, tme, avx512_vpopcntdq, rdpid, cldemote, movdiri, movdir64b, md_clear, serialize, amx_bf16, avx512_fp16, amx_tile, amx_int8, flush_l1d, arch_capabilities
Ecpus
24.6
Scalability
51.25

System Resources and Accelerators

MEMORY
Memory Amount
512 GiB
Memory Amount Actual
512 GiB
Memory Generation
DDR5
Memory Speed
5600 Mhz
ACCELERATORS
Accelerator Count
2
Accelerator Memory Min
96 GiB
Accelerator Memory Total
192 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Blackwell
Accelerator Model
RTX PRO 6000
Accelerators
  • NVIDIA Blackwell NVIDIA RTX PRO 6000 Blackwell Server Edition (Memory amount: 97887, Firmware version: 590.48.01, BIOS version: 98.02.9E.00.03, Clock rate: 2430)
  • NVIDIA Blackwell NVIDIA RTX PRO 6000 Blackwell Server Edition (Memory amount: 97887, Firmware version: 590.48.01, BIOS version: 98.02.9E.00.03, Clock rate: 2430)
STORAGE
Storage Size
3800 GB
Storage Type
nvme ssd
Storages
  • 3800 GB nvme ssd
NETWORK
Network Speed Baseline
400 Gbps
Network Speed Max
400 Gbps
Network Storage Speed Baseline
25 Gbps
Network Storage Speed Max
25 Gbps
Inbound Traffic
0 GB/month
Outbound Traffic
0 GB/month
IPv4
0

CPU and System Topology

Server Description

High-density acceleration platform featuring dual Blackwell GPUs with 192 GB VRAM, 512 GB DDR5 memory, and high-speed 400 Gbps networking.

GPU AcceleratedMemory OptimizedStorage & Database

Amazon Web Services g7e.12xlarge is a graphics-intensive x86_64 instance built on the AWS Nitro Hypervisor. Powered by a 24-core Intel Xeon 8559C processor running at 2.4 GHz, it offers 48 vCPUs with AVX-512 and Intel AMX matrix extensions. The system integrates 512 GB of high-speed DDR5 memory, dual NVIDIA Blackwell RTX PRO 6000 GPUs providing 192 GB of total VRAM, a 3800 GB NVMe SSD, and 400 Gbps network bandwidth. Benchmarks highlight top-tier LLM inference speeds across 7B and 70B parameter models, combined with strong memory bandwidth reaching 225.52 GB/s for uncached reads. While multi-core CPU benchmarks reflect average raw compute performance, the server excels in accelerated workloads. The g7e.12xlarge provides high resource density, making it well-suited for large language model serving, advanced AI inference, image and ray-tracing rendering, high-concurrency Redis instances, and memory-intensive data analytics.

Economics

Average Price per Region

Prices per Zone

Lowest Prices

Performance

Workload Profiles

3.4Score
Precomputed compound score for Cache Intensive workloads. A weighted average (geometric mean) of benchmark scores compared to their medians: score = ∏ (x_i / m_i)^(w_i / Σw). The score of 1.0 represents a synthetic baseline server with the median performance of each component benchmark; 0.5 means roughly half the performance; and 2.0 means twice the performance of that reference profile. Component weights: 50% Redis RPS (pipeline=1, SET), 20% Redis RPS (pipeline=16, SET), 10% PassMark Memory Mark (composite), 10% Memory bandwidth (read, 16 MB ~ L3), 10% PassMark single-thread CPU. Rationale for component selection: In-memory key-value store workload, mixing direct Redis performance metrics with memory speed and latency benchmarks, and single-core CPU performance profiles.
Component Score Weight Impact
Raw Reference Normalized Target Actual
Redis RPS (pipeline=1, SET) 9,521,312 ops/sec 1,930,206 ops/sec 5 50.00% 50.00% +120.00%
Redis RPS (pipeline=16, SET) 48,643,051 ops/sec 11,792,140 ops/sec 4.1 20.00% 20.00% +33.00%
PassMark Memory Mark (composite) 2,822 2,412 1.2 10.00% 10.00% +1.80%
Memory bandwidth (read, 16 MB ~ L3) 319,721 MB/sec 108,490 MB/sec 2.9 10.00% 10.00% +11.00%
PassMark single-thread CPU 3,151 Mops/s 2,365 Mops/s 1.3 10.00% 10.00% +2.70%

GPU Bandwidth and Latency

Node (all GPUs)
Node baseline directional bandwidth
50.6 GB/s
Node GPU-to-host bandwidth
57.2 GB/s
Node host-to-GPU bandwidth
55.7 GB/s
Peer-to-peer
GPU-to-GPU P2P duplex bandwidth
52.6 GB/s
GPU-to-GPU P2P gather bandwidth
54.2 GB/s
GPU-to-GPU P2P latency
1,107 ns
GPU-to-GPU P2P bandwidth
54.0 GB/s
Single GPU slot
Per-GPU GPU-to-host bandwidth
57.2 GB/s
Per-GPU host-to-GPU bandwidth
55.6 GB/s
Per-GPU host-device latency
1,031 ns
Per-GPU device memory copy bandwidth
728.8 GB/s
Benchmark Environment
Driver Version
 590.48.01
CUDA Version (driver)
 13.1
nvbandwidth Version
 v0.8
GPU Count
 2
P2P Supported
Yes
Efficiency
SM/CE copy efficiency
0.93

Memory Bandwidth

Compression

OpenSSL

Geekbench Single-Core

Score: 2,272

Geekbench Multi-Core

Score: 20,216

Passmark CPU Scores

BENCHMARKSCORE
Mark
 68,416
Compression
 887,447
Encryption
 54,587
Extended Instructions
 71,925
Floating Point Maths
 175,602
Integer Maths
 247,960
Physics
 10,978
Prime Numbers
 496
Single Threaded
 3,151
String Sorting
 121,492

Passmark Memory Scores

BENCHMARKSCORE
Memory Mark
 2,822
Database Operations
 26,840
Memory Latency
 45
Memory Read Cached
 25,924
Memory Read Uncached
 9,346
Memory Write
 9,068

Stress-ng Raw Scores

Stress-ng Relative Multicore Performance

LLM Inference Speed for Prompt Processing

LLM Inference Speed for Text Generation

Static Web Server

Redis

Alternatives

Servers of the Same Family

INSTANCEvCPUsMEMORYAccelerators
g7e.2xlarge 864 GiB1
g7e.4xlarge 16128 GiB1
g7e.8xlarge 32256 GiB1
g7e.24xlarge 961024 GiB4
g7e.48xlarge 1922048 GiB8

Similar Servers

INSTANCEVENDORvCPUsMEMORYAccelerators

g7e.12xlarge FAQs