g5.8xlarge by Amazon Web Services

g5.8xlarge is a Graphics intensive Gen5 8xlarge server offered by Amazon Web Services with 32 vCPUs, 128 GiB of memory and 900 GB of storage. The pricing starts at 0.3426 USD per hour.
32 vCPU
128 GiB Memory
900 GB Storage
1 GPU
Spare SCore
12,274
(All-cores)
766
(Single-core)

Specifications

Server Metadata

Vendor ID
aws
Name
g5.8xlarge
Description
Graphics intensive Gen5 8xlarge
Family
g5
Hw Virt
Average Time To Start
9
Status
active
Observed At
2026-11-02T08:25:35.977337

Availability

REGION / IDSPOTONDEMAND
Northern Virgina (US) / us-east-1
1.04032 USD/h 2.448 USD/h
Oregon (US) / us-west-2
0.930033 USD/h 2.448 USD/h
Ohio (US) / us-east-2
1.2195 USD/h 2.448 USD/h
Aragón (ES) / eu-south-2
2.5048 USD/h 2.5794 USD/h
Stockholm (SE) / eu-north-1
0.61505 USD/h 2.5964 USD/h
Quebec (CA) / ca-central-1
2.7181 USD/h 2.7181 USD/h
Dublin (IE) / eu-west-1
1.5176 USD/h 2.7327 USD/h
Tel Aviv (IL) / il-central-1
1.768267 USD/h 2.8693 USD/h
Mumbai (IN) / ap-south-1
2.03865 USD/h 2.9396 USD/h
United Arab Emirates / me-central-1
1.2293 USD/h 3.0028 USD/h
Seoul (KR) / ap-northeast-2
1.3834 USD/h 3.0101 USD/h
Frankfurt (DE) / eu-central-1
1.3554 USD/h 3.0612 USD/h
Paris (FR) / eu-west-3
1.8632 USD/h 3.1075 USD/h
London (GB) / eu-west-2
2.07015 USD/h 3.1075 USD/h
Sydney (AU) / ap-southeast-2
0.9467 USD/h 3.1829 USD/h
Jakarta (ID) / ap-southeast-3
0.7947 USD/h 3.4262 USD/h
Tokyo (JP) / ap-northeast-1
1.63015 USD/h 3.5503 USD/h
Hong Kong (HK) / ap-east-1
0.4482 USD/h 3.7689 USD/h
Sao Paulo (BR) / sa-east-1
1.36065 USD/h 4.1611 USD/h

Processor

vCPUs
32
Hypervisor
nitro
CPU Allocation
Dedicated
CPU Cores
16
CPU Speed
3.3 GHz
CPU Architecture
x86_64
CPU Manufacturer
AMD
CPU Family
EPYC
CPU Model
7R32
CPU L1D Cache
32 KiB
CPU L1D Cache Total
512 KiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
512 KiB
CPU L2 Cache
512 KiB
CPU L2 Cache Total
8 MiB
CPU L3 Cache
16 MiB
CPU L3 Cache Total
64 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, ht, syscall, nx, mmxext, fxsr_opt, pdpe1gb, rdtscp, lm, constant_tsc, rep_good, nopl, nonstop_tsc, cpuid, extd_apicid, aperfmperf, tsc_known_freq, pni, pclmulqdq, ssse3, fma, cx16, sse4_1, sse4_2, movbe, popcnt, aes, xsave, avx, f16c, rdrand, hypervisor, lahf_lm, cmp_legacy, cr8_legacy, abm, sse4a, misalignsse, 3dnowprefetch, topoext, ssbd, ibrs, ibpb, stibp, vmmcall, fsgsbase, bmi1, avx2, smep, bmi2, rdseed, adx, smap, clflushopt, clwb, sha_ni, xsaveopt, xsavec, xgetbv1, clzero, xsaveerptr, rdpru, wbnoinvd, arat, npt, nrip_save, rdpid
Ecpus
16
Scalability
50

System Resources and Accelerators

MEMORY
Memory Amount
128 GiB
Memory Amount Actual
128 GiB
Memory Generation
DDR4
Memory Speed
3200 Mhz
ACCELERATORS
Accelerator Count
1
Accelerator Memory Min
22 GiB
Accelerator Memory Total
22 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Ampere
Accelerator Model
A10G
Accelerators
  • NVIDIA Ampere NVIDIA A10G (Memory amount: 23028, Firmware version: 590.48.01, BIOS version: 94.02.91.00.07, Clock rate: 1710)
STORAGE
Storage Size
900 GB
Storage Type
nvme ssd
Storages
  • 900 GB nvme ssd
NETWORK
Network Speed Baseline
25 Gbps
Network Speed Max
25 Gbps
Network Storage Speed Baseline
16 Gbps
Network Storage Speed Max
16 Gbps
Inbound Traffic
0 GB/month
Outbound Traffic
0 GB/month
IPv4
0

CPU and System Topology

Server Description

A GPU-accelerated instance combining dedicated AMD EPYC processors, an NVIDIA A10G GPU, and local NVMe storage for parallel computing workloads.

GPU AcceleratedStorage & DatabaseGeneral Purpose

Amazon Web Services g5.8xlarge is a GPU-accelerated instance featuring 32 dedicated vCPUs from an AMD EPYC 7R32 processor running at 3.3 GHz, 128.0 GB of DDR4 memory, and a single NVIDIA Ampere A10G GPU with 22 GB of VRAM. It includes 900 GB of local NVMe SSD storage and a 25 Gbps baseline network bandwidth. While single-core CPU benchmarks place in the bottom 10% and multi-core CPU performance sits in the middle 50% average, the instance excels in memory performance and GPU-driven tasks. Specifically, LLM inference speeds for small and medium models rank in the top 10% tier. This hardware profile is suited for workloads requiring GPU acceleration, such as machine learning inference, ray tracing, and database operations utilizing local NVMe storage.

Economics

Average Price per Region

Prices per Zone

Lowest Prices

Performance

Workload Profiles

1.2Score
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) 2,617,513 ops/sec 1,918,343 ops/sec 1.4 50.00% 50.00% +18.00%
Redis RPS (pipeline=16, SET) 17,320,424 ops/sec 11,704,734 ops/sec 1.4 20.00% 20.00% +7.00%
PassMark Memory Mark (composite) 2,868 2,416 1.2 10.00% 10.00% +1.80%
Memory bandwidth (read, 16 MB ~ L3) 74,300 MB/sec 108,466 MB/sec 0.67 10.00% 10.00% -3.90%
PassMark single-thread CPU 2,136 Mops/s 2,365 Mops/s 0.88 10.00% 10.00% -1.30%

GPU Bandwidth and Latency

Node (all GPUs)
Node baseline directional bandwidth
11.0 GB/s
Node GPU-to-host bandwidth
13.2 GB/s
Node host-to-GPU bandwidth
13.5 GB/s
Single GPU slot
Per-GPU GPU-to-host bandwidth
13.2 GB/s
Per-GPU host-to-GPU bandwidth
13.5 GB/s
Per-GPU host-device latency
921 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: 1,335

Geekbench Multi-Core

Score: 11,223

Passmark CPU Scores

BENCHMARKSCORE
Mark
 34,661
Compression
 479,822
Encryption
 34,138
Extended Instructions
 29,725
Floating Point Maths
 64,394
Integer Maths
 107,967
Physics
 3,311
Prime Numbers
 162
Single Threaded
 2,136
String Sorting
 66,285

Passmark Memory Scores

BENCHMARKSCORE
Memory Mark
 2,868
Database Operations
 11,052
Memory Latency
 50
Memory Read Cached
 24,687
Memory Read Uncached
 12,944
Memory Write
 13,346

Stress-ng Stressors

CPU integerSingleMultiScaling
atomic
 402.49 344.56 0.86×
bitops
 120,283 2,436,822 20.26×
branch
 153,005 4,319,349 28.23×
funccall
 502.5 10,794 21.48×
funcret
 14,789 331,173 22.39×
hash
 131,987 3,288,539 24.92×
intmath
 19,592 337,427 17.22×
nop
 236,316 3,826,674 16.19×
prime
 345,677 7,334,100 21.22×
rotate
 59,935 996,088 16.62×
str
 625,557 12,995,371 20.77×
wcs
 468,350 8,572,629 18.30×
CryptographySingleMultiScaling
crypt
 2,809 48,218 17.16×
CompressionSingleMultiScaling
jpeg
 664.92 14,278 21.47×
zlib
 67.09 1,531 22.83×
HypervisorSingleMultiScaling
clock
 19,295 254,250 13.18×
clone
 129.55 2,105 16.25×
context
 3,365 91,052 27.06×
cpu-sched
 70.37 2,032 28.88×
eventfd
 28,310 1,600,201 56.52×
fault
 8,430 158,617 18.82×
fork
 2,887 43,700 15.13×
getrandom
 33,561 743,428 22.15×
hrtimers
 136,611 2,177,913 15.94×
mprotect
 578,684 5,557,710 9.60×
mtx
 21,827,118 1,286,103,175 58.92×
munmap
 2,009 35,674 17.76×
null
 419,587 10,779,646 25.69×
pipe
 1,762,207 9,436,929 5.36×
priv-instr
 435,114 10,880,376 25.01×
resched
 4,759,527 1,308,809 0.27×
switch
 129,256 3,775,508 29.21×
timermix
 143,823 3,786,117 26.32×
tlb-shootdown
 8,346 66,321 7.95×
tsc
 652,663 20,297,011 31.10×
usersyscall
 279,315 7,178,254 25.70×
vdso
 78,387,856 1,662,990,197 21.21×
vfork
 16,032 123,396 7.70×
x86syscall
 5,300,723 136,772,602 25.80×
yield
 14,320,987 14,989,095 1.05×
zero
 2,273,735 54,369,898 23.91×
CPU float/SIMDSingleMultiScaling
eigen
 12,390 232,942 18.80×
expmath
 2,782 69,967 25.15×
fma
 740,657 18,615,921 25.13×
fp
 208.21 4,803 23.07×
fractal
 6.4 160.47 25.07×
hyperbolic
 3,641 87,554 24.04×
logmath
 4,109 95,782 23.31×
matrix
 2,752 78,889 28.67×
matrix-3d
 112.94 2,511 22.23×
monte-carlo
 182.94 4,797 26.22×
mpfr
 1,276 26,431 20.72×
powmath
 1,341 34,448 25.68×
trig
 1,562 38,136 24.41×
veccmp
 17,136 332,382 19.40×
vecfp
 1,386 32,985 23.80×
vecmath
 3,367 86,571 25.71×
vecshuf
 484.46 9,587 19.79×
vecwide
 26,545 506,304 19.07×
AlgorithmsSingleMultiScaling
bitonicsort
 6.8 135.19 19.88×
bsearch
 228.52 4,640 20.30×
bubblesort
 0.8 12.8 16.00×
fibsearch
 259.63 5,403 20.81×
heapsort
 8.6 192.79 22.42×
hsearch
 5,340 104,091 19.49×
insertionsort
 6.6 126.6 19.18×
malloc
 2,582,417 42,212,341 16.35×
mcontend
 107.28 4,719 43.99×
mergesort
 26 628.77 24.18×
qsort
 16.6 406.78 24.50×
skiplist
 5,508 140,517 25.51×
tree
 1.8 19.18 10.66×
MemorySingleMultiScaling
cache
 472,872 942,808 1.99×
cachehammer
 291,993 7,332,395 25.11×
cacheline
 46.37 2,410 51.98×
l1cache
 197.77 3,079 15.57×
memcpy
 351.87 5,929 16.85×
memrate
 1.95 6.17 3.16×
memthrash
 133.85 676.94 5.06×
ptr-chase
 30,896,412 258,174,304 8.36×
spinmem
 251.98 1,649 6.54×
stream
 177.18 21,507 121.38×
vm
 55,178 295,497 5.36×

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

PostgreSQL heavy read-only throughput

Alternatives

Servers of the Same Family

INSTANCEvCPUsMEMORYAccelerators
g5.xlarge 416 GiB1
g5.2xlarge 832 GiB1
g5.4xlarge 1664 GiB1
g5.12xlarge 48192 GiB4
g5.16xlarge 64256 GiB1
g5.24xlarge 96384 GiB4
g5.48xlarge 192768 GiB8

Similar Servers

INSTANCEVENDORvCPUsMEMORYAccelerators

g5.8xlarge FAQs