g5g.4xlarge by Amazon Web Services

g5g.4xlarge is a Graphics intensive [AWS Graviton processors] Gen5 4xlarge server offered by Amazon Web Services with 16 vCPUs, 32 GiB of memory and 0 GB of storage. The pricing starts at 0.0924 USD per hour.
16 vCPU
32 GiB Memory
1 GPU
Spare SCore
24,287
(All-cores)
1,520
(Single-core)

Specifications

Server Metadata

Vendor ID
aws
Name
g5g.4xlarge
Description
Graphics intensive [AWS Graviton processors] Gen5 4xlarge
Family
g5g
Hw Virt
Status
active
Observed At
2026-10-26T17:35:04.085436

Availability

REGION / IDSPOTONDEMAND
Northern Virgina (US) / us-east-1
0.314675 USD/h 0.828 USD/h
Oregon (US) / us-west-2
0.277433 USD/h 0.828 USD/h
Aragón (ES) / eu-south-2
0.0924 USD/h 0.924 USD/h
Seoul (KR) / ap-northeast-2
0.4583 USD/h 1.0185 USD/h
Frankfurt (DE) / eu-central-1
0.3703 USD/h 1.0358 USD/h
Tokyo (JP) / ap-northeast-1
0.50455 USD/h 1.1176 USD/h
Singapore (SG) / ap-southeast-1
0.45425 USD/h 1.1586 USD/h

Processor

vCPUs
16
Hypervisor
nitro
CPU Allocation
Dedicated
CPU Cores
16
CPU Speed
2.5 GHz
CPU Architecture
arm64
CPU Manufacturer
AWS
CPU Family
ARMv8
CPU Model
Neoverse-N1
CPU L1D Cache
64 KiB
CPU L1D Cache Total
1 MiB
CPU L1I Cache
64 KiB
CPU L1I Cache Total
1 MiB
CPU L2 Cache
1 MiB
CPU L2 Cache Total
16 MiB
CPU L3 Cache
32 MiB
CPU L3 Cache Total
32 MiB
CPU Flags
fp, asimd, evtstrm, aes, pmull, sha1, sha2, crc32, atomics, fphp, asimdhp, cpuid, asimdrdm, lrcpc, dcpop, asimddp, ssbs
Ecpus
16
Scalability
100

System Resources and Accelerators

MEMORY
Memory Amount
32 GiB
Memory Amount Actual
32 GiB
ACCELERATORS
Accelerator Count
1
Accelerator Memory Min
16 GiB
Accelerator Memory Total
16 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Turing
Accelerator Model
T4G
Accelerators
  • NVIDIA Turing NVIDIA T4G (Memory amount: 15360, Firmware version: 590.48.01, BIOS version: 90.04.AF.00.02, Clock rate: 1590)
STORAGE
Storage Size
0 GB
Storages
    NETWORK
    Network Speed Baseline
    5 Gbps
    Network Speed Max
    10 Gbps
    Network Storage Speed Baseline
    4.75 Gbps
    Network Storage Speed Max
    4.75 Gbps
    Inbound Traffic
    0 GB/month
    Outbound Traffic
    0 GB/month
    IPv4
    0

    CPU and System Topology

    Server Description

    An ARM-based accelerated instance featuring dedicated Neoverse cores and an NVIDIA GPU for cost-efficient graphics and lightweight machine learning workloads.

    GPU AcceleratedCompute Optimized

    Amazon Web Services g5g.4xlarge is an ARM64-based virtual machine featuring 16 dedicated AWS Neoverse-N1 physical cores running at 2.5 GHz, paired with 32.0 GB of system memory. Designed for graphics-intensive and accelerated workloads, it includes a single NVIDIA T4G GPU with 16 GB of VRAM. The instance operates on the AWS Nitro hypervisor and provides a baseline network bandwidth of 5 Gbps, though it lacks local storage. Benchmarks show average multi-core CPU and GPU performance, with a stress-ng multi-core score of 24286.82 and a Geekbench ray tracing score of 13953.0, alongside moderate LLM inference capabilities. This server offers a qualitatively cost-efficient profile for ARM-native GPU applications, making it suitable for Android game streaming, lightweight machine learning inference, software compilation, and static web serving.

    Economics

    Average Price per Region

    Prices per Zone

    Lowest Prices

    Performance

    Workload Profiles

    0.87Score
    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) 1,723,779 ops/sec 1,919,289 ops/sec 0.89 50.00% 50.00% -5.70%
    Redis RPS (pipeline=16, SET) 10,739,435 ops/sec 11,728,058 ops/sec 0.92 20.00% 20.00% -1.70%
    PassMark Memory Mark (composite) 1,784 2,416 0.75 10.00% 10.00% -2.80%
    Memory bandwidth (read, 16 MB ~ L3) 163,972 MB/sec 108,466 MB/sec 1.5 10.00% 10.00% +4.10%
    PassMark single-thread CPU 1,109 Mops/s 2,365 Mops/s 0.46 10.00% 10.00% -7.50%

    GPU Bandwidth and Latency

    Node (all GPUs)
    Node baseline directional bandwidth
    		5.7 GB/s
    Node GPU-to-host bandwidth
    		6.6 GB/s
    Node host-to-GPU bandwidth
    		6.6 GB/s
    Single GPU slot
    Per-GPU GPU-to-host bandwidth
    		6.6 GB/s
    Per-GPU host-to-GPU bandwidth
    		6.6 GB/s
    Per-GPU host-device latency
    		780 ns
    Benchmark Environment
    Driver Version
    		 535.183.01
    CUDA Runtime Version
    		 12.6
    nvbandwidth Version
    		 v0.5
    GPU Count
    		 1
    P2P Supported
    No
    Efficiency
    SM/CE copy efficiency
    		1.00

    Memory Bandwidth

    Compression

    OpenSSL

    Geekbench Single-Core

    Score: 948

    Geekbench Multi-Core

    Score: 8,557

    Passmark CPU Scores

    BENCHMARKSCORE
    Mark
    		 9,821
    Compression
    		 76,055
    Encryption
    		 2,188
    Extended Instructions
    		 9,946
    Floating Point Maths
    		 47,967
    Integer Maths
    		 63,206
    Physics
    		 1,959
    Prime Numbers
    		 108
    Single Threaded
    		 1,109
    String Sorting
    		 45,733

    Passmark Memory Scores

    BENCHMARKSCORE
    Memory Mark
    		 1,784
    Database Operations
    		 10,340
    Memory Latency
    		 70
    Memory Read Cached
    		 9,485
    Memory Read Uncached
    		 9,210
    Memory Write
    		 12,567

    Stress-ng Stressors

    CPU integerSingleMultiScaling
    atomic
    		 119.65 103.02 0.86×
    bitops
    		 90,431 1,445,236 15.98×
    branch
    		 183,330 2,931,327 15.99×
    funccall
    		 1,113 17,789 15.98×
    funcret
    		 17,416 277,718 15.95×
    hash
    		 106,580 1,703,589 15.98×
    intmath
    		 26,590 416,306 15.66×
    nop
    		 144,618 2,311,137 15.98×
    prime
    		 284,230 4,550,298 16.01×
    rotate
    		 80,339 1,283,929 15.98×
    str
    		 427,654 6,820,382 15.95×
    wcs
    		 99,246 1,585,635 15.98×
    CryptographySingleMultiScaling
    crypt
    		 2,422 38,700 15.98×
    CompressionSingleMultiScaling
    jpeg
    		 429.76 6,744 15.69×
    zlib
    		 37.26 482 12.94×
    HypervisorSingleMultiScaling
    clock
    		 27,768 309,702 11.15×
    clone
    		 243.55 2,514 10.32×
    context
    		 2,174 34,787 16.00×
    cpu-sched
    		 66.61 1,020 15.32×
    eventfd
    		 54,694 480,271 8.78×
    fault
    		 11,555 142,312 12.32×
    fork
    		 3,875 35,827 9.24×
    getrandom
    		 40,088 640,246 15.97×
    hrtimers
    		 432,475 1,830,991 4.23×
    mprotect
    		 945,534 4,492,972 4.75×
    mtx
    		 8,909,383 575,451,343 64.59×
    munmap
    		 2,669 31,123 11.66×
    null
    		 385,194 6,159,342 15.99×
    pipe
    		 681,328 8,169,168 11.99×
    priv-instr
    		 606,815 9,099,535 15.00×
    resched
    		 4,958,340 2,373,393 0.48×
    switch
    		 91,947 4,604,607 50.08×
    timermix
    		 275,894 4,417,957 16.01×
    tlb-shootdown
    		 4,437 8,538 1.92×
    vdso
    		 42,684,012 682,356,248 15.99×
    vfork
    		 28,454 193,920 6.82×
    yield
    		 12,268,316 13,586,726 1.11×
    zero
    		 1,727,313 27,601,476 15.98×
    CPU float/SIMDSingleMultiScaling
    eigen
    		 1,111 17,758 15.99×
    expmath
    		 199.66 3,191 15.98×
    fma
    		 532,882 8,517,644 15.98×
    fp
    		 86.4 1,379 15.96×
    fractal
    		 5.4 85.81 15.89×
    hyperbolic
    		 203.16 3,248 15.99×
    logmath
    		 244.23 3,904 15.98×
    matrix
    		 4,445 71,061 15.99×
    matrix-3d
    		 523.14 6,411 12.26×
    monte-carlo
    		 160.92 2,575 16.00×
    mpfr
    		 533.73 8,531 15.98×
    powmath
    		 146.75 2,346 15.98×
    trig
    		 274.06 4,381 15.99×
    veccmp
    		 14,276 228,307 15.99×
    vecfp
    		 620.15 9,890 15.95×
    vecmath
    		 2,570 41,099 15.99×
    vecshuf
    		 559.83 8,949 15.99×
    vecwide
    		 15,311 244,640 15.98×
    AlgorithmsSingleMultiScaling
    bitonicsort
    		 4.4 70.59 16.04×
    bsearch
    		 214.74 3,433 15.99×
    bubblesort
    		 0.6 9.4 15.67×
    fibsearch
    		 227.55 3,636 15.98×
    heapsort
    		 6.8 109.59 16.12×
    hsearch
    		 4,084 65,354 16.00×
    insertionsort
    		 3.8 60.8 16.00×
    malloc
    		 2,628,921 31,516,792 11.99×
    mcontend
    		 248.3 26,424 106.42×
    mergesort
    		 23.2 371.17 16.00×
    qsort
    		 16.4 260.58 15.89×
    skiplist
    		 4,461 71,245 15.97×
    tree
    		 1.2 12.79 10.66×
    MemorySingleMultiScaling
    cache
    		 354,699 614,456 1.73×
    cachehammer
    		 93,520 224,160 2.40×
    cacheline
    		 59.35 976.48 16.45×
    l1cache
    		 151.57 2,420 15.97×
    memcpy
    		 368.99 5,897 15.98×
    memrate
    		 3.34 27.93 8.36×
    memthrash
    		 137.54 309.29 2.25×
    ptr-chase
    		 20,726,431 159,522,009 7.70×
    spinmem
    		 449.98 809.3 1.80×
    stream
    		 85.2 8,916 104.64×
    vm
    		 56,407 873,421 15.48×

    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
    g5g.xlarge 48 GiB1
    g5g.2xlarge 816 GiB1
    g5g.8xlarge 3264 GiB1
    g5g.16xlarge 64128 GiB2
    g5g.metal 64128 GiB2

    Similar Servers

    INSTANCEVENDORvCPUsMEMORYAccelerators

    g5g.4xlarge FAQs