g4dn.2xlarge by Amazon Web Services

g4dn.2xlarge is a Graphics intensive [Instance store volumes] [Network and EBS optimized] Gen4 2xlarge server offered by Amazon Web Services with 8 vCPUs, 32 GiB of memory and 225 GB of storage. The pricing starts at 0.0893 USD per hour.
8 vCPU
32 GiB Memory
225 GB Storage
1 GPU
Spare SCore
6,760
(All-cores)
1,285
(Single-core)

Specifications

Server Metadata

Vendor ID
aws
Name
g4dn.2xlarge
Description
Graphics intensive [Instance store volumes] [Network and EBS optimized] Gen4 2xlarge
Family
g4dn
Hw Virt
Status
active
Observed At
2026-11-01T20:44:46.789730

Availability

REGION / IDSPOTONDEMAND
Northern Virgina (US) / us-east-1
0.31646 USD/h 0.752 USD/h
Oregon (US) / us-west-2
0.3463 USD/h 0.752 USD/h
Ohio (US) / us-east-2
0.392967 USD/h 0.752 USD/h
Stockholm (SE) / eu-north-1
0.096833 USD/h 0.798 USD/h
Mumbai (IN) / ap-south-1
0.364833 USD/h 0.828 USD/h
Quebec (CA) / ca-central-1
0.3505 USD/h 0.835 USD/h
Dublin (IE) / eu-west-1
0.410267 USD/h 0.838 USD/h
Paris (FR) / eu-west-3
0.2682 USD/h 0.879 USD/h
Milan (IT) / eu-south-1
0.1988 USD/h 0.88 USD/h
London (GB) / eu-west-2
0.287567 USD/h 0.88 USD/h
California (US) / us-west-1
0.3841 USD/h 0.902 USD/h
Seoul (KR) / ap-northeast-2
0.447267 USD/h 0.925 USD/h
Frankfurt (DE) / eu-central-1
0.385933 USD/h 0.94 USD/h
Sydney (AU) / ap-southeast-2
0.429567 USD/h 0.978 USD/h
Cape Town (ZA) / af-south-1
0.4328 USD/h 0.998 USD/h
Osaka (JP) / ap-northeast-3
0.38425 USD/h 1.015 USD/h
Tokyo (JP) / ap-northeast-1
0.497133 USD/h 1.015 USD/h
Singapore (SG) / ap-southeast-1
0.530833 USD/h 1.052 USD/h
Hong Kong (HK) / ap-east-1
0.347833 USD/h 1.158 USD/h
Sao Paulo (BR) / sa-east-1
0.132167 USD/h 1.278 USD/h

Processor

vCPUs
8
Hypervisor
nitro
CPU Allocation
Dedicated
CPU Cores
4
CPU Speed
2.5 GHz
CPU Architecture
x86_64
CPU Manufacturer
Intel
CPU Family
Xeon
CPU Model
8259CL
CPU L1D Cache
32 KiB
CPU L1D Cache Total
128 KiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
128 KiB
CPU L2 Cache
1 MiB
CPU L2 Cache Total
4 MiB
CPU L3 Cache
36 MiB
CPU L3 Cache Total
36 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, rep_good, nopl, xtopology, nonstop_tsc, cpuid, aperfmperf, tsc_known_freq, pni, pclmulqdq, ssse3, fma, cx16, pcid, sse4_1, sse4_2, x2apic, movbe, popcnt, tsc_deadline_timer, aes, xsave, avx, f16c, rdrand, hypervisor, lahf_lm, abm, 3dnowprefetch, pti, fsgsbase, tsc_adjust, bmi1, avx2, smep, bmi2, erms, invpcid, mpx, avx512f, avx512dq, rdseed, adx, smap, clflushopt, clwb, avx512cd, avx512bw, avx512vl, xsaveopt, xsavec, xgetbv1, xsaves, ida, arat, pku, ospke, avx512_vnni
Ecpus
5.3
Scalability
66.25

System Resources and Accelerators

MEMORY
Memory Amount
32 GiB
Memory Amount Actual
32 GiB
Memory Generation
DDR4
Memory Speed
3200 Mhz
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
T4
Accelerators
  • NVIDIA Turing Tesla T4 (Memory amount: 15360, Firmware version: 590.48.01, BIOS version: 90.04.96.00.A0, Clock rate: 1590)
STORAGE
Storage Size
225 GB
Storage Type
nvme ssd
Storages
  • 225 GB nvme ssd
NETWORK
Network Speed Baseline
10 Gbps
Network Speed Max
25 Gbps
Network Storage Speed Baseline
1.15 Gbps
Network Storage Speed Max
3.5 Gbps
Inbound Traffic
0 GB/month
Outbound Traffic
0 GB/month
IPv4
0

CPU and System Topology

Server Description

An accelerator-optimized instance featuring an NVIDIA T4 GPU and local NVMe storage for efficient machine learning inference and graphics workloads.

GPU AcceleratedGeneral PurposeStorage & Database

Amazon Web Services g4dn.2xlarge is a graphics-optimized instance designed for GPU-accelerated workloads. Powered by an Intel Xeon 8259CL processor with 8 vCPUs and 32.0 GB of DDR4 RAM, it features a dedicated NVIDIA T4 GPU with 16 GB of VRAM. It also includes 225 GB of local NVMe SSD storage and a 10 Gbps baseline network bandwidth. While CPU-only benchmarks, such as Geekbench and stress-ng, place the instance in the lower performance tiers, its GPU capabilities deliver top-tier results in LLM inference prompt processing and strong performance in text generation for small and medium models. This hardware configuration makes the instance a cost-effective choice for machine learning inference, graphics rendering, and video transcoding, though it presents a tradeoff for CPU-heavy sequential tasks.

Economics

Average Price per Region

Prices per Zone

Lowest Prices

Performance

Workload Profiles

0.44Score
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) 676,521 ops/sec 1,918,343 ops/sec 0.36 50.00% 50.00% -40.00%
Redis RPS (pipeline=16, SET) 4,284,617 ops/sec 11,704,734 ops/sec 0.36 20.00% 20.00% -18.00%
PassMark Memory Mark (composite) 2,315 2,416 0.96 10.00% 10.00% -0.41%
Memory bandwidth (read, 16 MB ~ L3) 53,303 MB/sec 108,466 MB/sec 0.48 10.00% 10.00% -7.10%
PassMark single-thread CPU 1,834 Mops/s 2,365 Mops/s 0.75 10.00% 10.00% -2.80%

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
6.3 GB/s
Single GPU slot
Per-GPU GPU-to-host bandwidth
6.6 GB/s
Per-GPU host-to-GPU bandwidth
6.3 GB/s
Per-GPU host-device latency
657 ns
Benchmark Environment
Driver Version
 535.183.01
CUDA Version (driver)
 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: 1,125

Geekbench Multi-Core

Score: 4,267

Passmark CPU Scores

BENCHMARKSCORE
Mark
 7,707
Compression
 100,086
Encryption
 3,133
Extended Instructions
 6,149
Floating Point Maths
 12,135
Integer Maths
 22,575
Physics
 922
Prime Numbers
 45
Single Threaded
 1,834
String Sorting
 11,900

Passmark Memory Scores

BENCHMARKSCORE
Memory Mark
 2,315
Database Operations
 2,744
Memory Latency
 55
Memory Read Cached
 22,451
Memory Read Uncached
 10,797
Memory Write
 10,104

Stress-ng Stressors

CPU integerSingleMultiScaling
atomic
 217.6 219.68 1.01×
bitops
 100,301 460,098 4.59×
branch
 142,809 865,541 6.06×
funccall
 511.9 2,103 4.11×
funcret
 10,661 60,727 5.70×
hash
 118,678 684,164 5.76×
intmath
 21,024 125,266 5.96×
nop
 177,185 611,212 3.45×
prime
 293,468 1,310,312 4.46×
rotate
 89,548 488,295 5.45×
str
 680,190 2,973,273 4.37×
wcs
 435,100 1,884,634 4.33×
CryptographySingleMultiScaling
crypt
 2,300 8,947 3.89×
CompressionSingleMultiScaling
jpeg
 548.46 2,565 4.68×
zlib
 38.34 280.56 7.32×
HypervisorSingleMultiScaling
clock
 26,046 123,034 4.72×
clone
 314.82 1,129 3.59×
context
 1,817 9,717 5.35×
cpu-sched
 59.59 506.9 8.51×
eventfd
 47,178 432,099 9.16×
fault
 13,385 65,015 4.86×
fork
 4,185 21,581 5.16×
getrandom
 41,659 204,674 4.91×
hrtimers
 170,848 660,789 3.87×
mprotect
 830,533 1,158,342 1.39×
mtx
 13,905,978 207,773,412 14.94×
munmap
 2,909 14,004 4.81×
null
 192,357 950,641 4.94×
pipe
 907,525 2,572,461 2.83×
priv-instr
 405,727 2,109,498 5.20×
resched
 1,272,183 498,245 0.39×
switch
 84,478 1,325,705 15.69×
timermix
 161,107 1,015,033 6.30×
tlb-shootdown
 8,073 16,201 2.01×
tsc
 923,893 5,141,043 5.56×
usersyscall
 275,295 1,423,590 5.17×
vdso
 76,888,982 429,707,774 5.59×
vfork
 29,157 117,282 4.02×
x86syscall
 2,122,854 10,749,761 5.06×
yield
 4,013,383 4,016,228 1.00×
zero
 1,410,902 6,890,926 4.88×
CPU float/SIMDSingleMultiScaling
eigen
 14,964 74,117 4.95×
expmath
 2,524 15,377 6.09×
fma
 565,739 3,698,571 6.54×
fp
 115.2 580.48 5.04×
fractal
 4.6 27.2 5.91×
hyperbolic
 2,627 16,458 6.27×
logmath
 4,269 25,161 5.89×
matrix
 2,869 18,593 6.48×
matrix-3d
 629.07 1,875 2.98×
monte-carlo
 96.97 515.28 5.31×
mpfr
 1,141 4,884 4.28×
powmath
 1,468 8,955 6.10×
trig
 1,583 8,391 5.30×
veccmp
 14,457 59,553 4.12×
vecfp
 1,216 6,012 4.94×
vecmath
 2,024 12,770 6.31×
vecshuf
 522.64 2,242 4.29×
vecwide
 21,057 97,346 4.62×
AlgorithmsSingleMultiScaling
bitonicsort
 6.6 31.2 4.73×
bsearch
 220.75 1,030 4.66×
bubblesort
 0.6 1.8 3.00×
fibsearch
 230.25 1,053 4.57×
heapsort
 8.2 39 4.76×
hsearch
 4,723 22,988 4.87×
insertionsort
 7.6 30.4 4.00×
malloc
 2,182,300 16,015,148 7.34×
mcontend
 144.88 1,712 11.82×
mergesort
 22 118.2 5.37×
qsort
 13.8 63 4.57×
skiplist
 4,518 25,594 5.66×
tree
 1.4 6.4 4.57×
MemorySingleMultiScaling
cache
 348,418 618,087 1.77×
cachehammer
 377,731 2,685,197 7.11×
cacheline
 34.12 278.46 8.16×
l1cache
 207.26 924.02 4.46×
memcpy
 328.06 1,407 4.29×
memrate
 1.17 5.36 4.58×
memthrash
 183.81 305.99 1.66×
ptr-chase
 31,914,035 90,006,558 2.82×
spinmem
 268.99 422.37 1.57×
stream
 30.86 1,192 38.61×
vm
 54,963 291,145 5.30×

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
g4dn.xlarge 416 GiB1
g4dn.4xlarge 1664 GiB1
g4dn.8xlarge 32128 GiB1
g4dn.12xlarge 48192 GiB4
g4dn.16xlarge 64256 GiB1
g4dn.metal 96384 GiB8

Similar Servers

INSTANCEVENDORvCPUsMEMORYAccelerators

g4dn.2xlarge FAQs