a3-megagpu-8g by Google Cloud Platform

a3-megagpu-8g is a Accelerator Optimized: 8 NVIDIA H100 GPU, 208 vCPUs, 1872GB RAM, 16 local SSD server offered by Google Cloud Platform with 208 vCPUs, 1872 GiB of memory and 6.43 TB of storage. The pricing starts at 27.8636 USD per hour.
208 vCPU
1872 GiB Memory
6.43 TB Storage
8 GPU
Spare SCore
167,246
(All-cores)
2,051
(Single-core)

Specifications

Server Metadata

Vendor ID
gcp
Server ID
1720366
Name
a3-megagpu-8g
Description
Accelerator Optimized: 8 NVIDIA H100 GPU, 208 vCPUs, 1872GB RAM, 16 local SSD
Family
a3
Hw Virt
Average Time To Start
27
Status
active
Observed At
2026-11-01T09:19:14.320239

Availability

REGION / IDSPOTONDEMAND
Columbus (US) / us-east5
41.4869 USD/h 93.4007 USD/h
Council Bluffs (US) / us-central1
45.5986 USD/h 93.4007 USD/h
The Dalles (US) / us-west1
45.5986 USD/h 93.4007 USD/h
Ashburn (US) / us-east4
34.5831 USD/h 93.4665 USD/h
St. Ghislain (BE) / europe-west1
47.9638 USD/h 102.777 USD/h
Las Vegas (US) / us-west4
27.8636 USD/h 105.1799 USD/h
Frankfurt (DE) / europe-west3
65.4982 USD/h 110.226 USD/h
Sydney (AU) / australia-southeast1
51.7962 USD/h 116.8166 USD/h
Tokyo (JP) / asia-northeast1
34.3045 USD/h 119.9556 USD/h
Jurong West (SG) / asia-southeast1
65.5043 USD/h 121.2894 USD/h
Eemshaven (NL) / europe-west4
72.6988 USD/h 121.2894 USD/h

Processor

vCPUs
208
CPU Allocation
Dedicated
CPU Cores
104
CPU Speed
2.7 GHz
CPU Architecture
x86_64
CPU Manufacturer
Intel
CPU Family
Xeon
CPU Model
8481C
CPU L1D Cache
48 KiB
CPU L1D Cache Total
5 MiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
3 MiB
CPU L2 Cache
2 MiB
CPU L2 Cache Total
208 MiB
CPU L3 Cache
105 MiB
CPU L3 Cache Total
210 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, tsc_known_freq, pni, pclmulqdq, ssse3, fma, cx16, pcid, sse4_1, sse4_2, x2apic, movbe, popcnt, aes, xsave, avx, f16c, rdrand, hypervisor, lahf_lm, abm, 3dnowprefetch, ssbd, ibrs, ibpb, stibp, ibrs_enhanced, fsgsbase, tsc_adjust, bmi1, avx2, smep, bmi2, erms, invpcid, rtm, avx512f, avx512dq, rdseed, adx, smap, avx512ifma, clflushopt, clwb, avx512cd, sha_ni, avx512bw, avx512vl, xsaveopt, xsavec, xgetbv1, xsaves, avx_vnni, avx512_bf16, arat, avx512vbmi, umip, avx512_vbmi2, gfni, vaes, vpclmulqdq, avx512_vnni, avx512_bitalg, avx512_vpopcntdq, la57, rdpid, cldemote, movdiri, movdir64b, fsrm, md_clear, serialize, amx_bf16, avx512_fp16, amx_tile, amx_int8, arch_capabilities
Ecpus
81.5
Scalability
39.18

System Resources and Accelerators

MEMORY
Memory Amount
2 TB
Memory Amount Actual
2 TB
ACCELERATORS
Accelerator Count
8
Accelerator Memory Min
80 GiB
Accelerator Memory Total
640 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Hopper
Accelerator Model
H100
Accelerators
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
  • NVIDIA Hopper NVIDIA H100 80GB HBM3 (Memory amount: 81559, Firmware version: 590.48.01, BIOS version: 96.00.CF.00.01, Clock rate: 1980)
STORAGE
Storage Size
6432 GB
Storage Type
nvme ssd
Storages
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
  • 402 GB nvme ssd
NETWORK
Inbound Traffic
0 GB/month
Outbound Traffic
0 GB/month
IPv4
0

CPU and System Topology

Server Description

An accelerator-optimized platform featuring eight NVIDIA H100 GPUs, massive system memory, and high-speed local NVMe storage for intensive parallel computing.

GPU AcceleratedMemory OptimizedStorage & Database

Google Cloud Platform a3-megagpu-8g is an accelerator-optimized server featuring 208 dedicated vCPUs on Intel Xeon 8481C processors, 1872.0 GB of RAM, and eight NVIDIA H100 GPUs with 640 GB of total VRAM. It includes 6432 GB of local NVMe SSD storage. Benchmarks show top-tier multi-core CPU performance, exceptional memory bandwidth, and high database throughput, though single-core performance remains average. The server supports advanced instruction sets like Intel AMX and AVX-512. This hardware configuration is highly suited for demanding workloads such as large language model inference, ray tracing, high-performance computing, and large-scale database operations.

Economics

Average Price per Region

Prices per Zone

Lowest Prices

Performance

Workload Profiles

7.1Score
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) 26,650,892 ops/sec 1,918,343 ops/sec 14 50.00% 50.00% +270.00%
Redis RPS (pipeline=16, SET) 125,372,253 ops/sec 11,704,734 ops/sec 11 20.00% 20.00% +62.00%
PassMark Memory Mark (composite) 2,331 2,416 0.96 10.00% 10.00% -0.41%
Memory bandwidth (read, 16 MB ~ L3) 459,338 MB/sec 108,466 MB/sec 4.2 10.00% 10.00% +15.00%
PassMark single-thread CPU 3,141 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
97.0 GB/s
Node GPU-to-host bandwidth
114.2 GB/s
Node host-to-GPU bandwidth
110.8 GB/s
Peer-to-peer
GPU-to-GPU P2P duplex bandwidth
383.4 GB/s
GPU-to-GPU P2P gather bandwidth
398.9 GB/s
GPU-to-GPU P2P latency
1,258 ns
GPU-to-GPU P2P bandwidth
388.8 GB/s
Single GPU slot
Per-GPU GPU-to-host bandwidth
28.6 GB/s
Per-GPU host-to-GPU bandwidth
27.7 GB/s
Per-GPU host-device latency
1,220 ns
Benchmark Environment
Driver Version
 535.183.01
CUDA Version (driver)
 12.6
nvbandwidth Version
 v0.5
GPU Count
 8
P2P Supported
Yes
Efficiency
SM/CE copy efficiency
0.91

Memory Bandwidth

Compression

OpenSSL

Geekbench Single-Core

Score: 2,120

Geekbench Multi-Core

Score: 19,116

Passmark CPU Scores

BENCHMARKSCORE
Mark
 125,327
Compression
 2,707,581
Encryption
 169,091
Extended Instructions
 187,524
Floating Point Maths
 569,260
Integer Maths
 798,874
Physics
 15,053
Prime Numbers
 1,030
Single Threaded
 3,141
String Sorting
 344,257

Passmark Memory Scores

BENCHMARKSCORE
Memory Mark
 2,331
Database Operations
 33,320
Memory Latency
 72
Memory Read Cached
 27,494
Memory Read Uncached
 9,677
Memory Write
 8,535

Stress-ng Stressors

CPU integerSingleMultiScaling
atomic
 189.12 33.56 0.18×
bitops
 157,353 14,996,425 95.30×
branch
 146,455 18,061,250 123.32×
funccall
 790.28 77,430 97.98×
funcret
 13,113 1,670,390 127.38×
hash
 188,386 21,694,778 115.16×
intmath
 68,715 6,409,792 93.28×
nop
 281,190 18,259,384 64.94×
prime
 497,937 43,473,281 87.31×
rotate
 105,793 8,928,087 84.39×
str
 978,770 99,016,242 101.16×
wcs
 740,058 61,050,984 82.49×
CryptographySingleMultiScaling
crypt
 3,590 267,109 74.41×
CompressionSingleMultiScaling
jpeg
 882.91 80,066 90.68×
zlib
 53.16 8,126 152.87×
HypervisorSingleMultiScaling
clock
 22,882 205,871 9.00×
clone
 410.4 13,615 33.17×
context
 5,505 723,514 131.42×
cpu-sched
 63.9 12,823 200.67×
eventfd
 65,847 2,426,418 36.85×
fault
 32,241 161,182 5.00×
fork
 4,145 52,418 12.65×
getrandom
 62,861 6,905,944 109.86×
hrtimers
 332,446 21,905,820 65.89×
mprotect
 646,521 14,321,506 22.15×
mtx
 20,713,107 4,927,405,336 237.89×
munmap
 3,171 63,679 20.08×
null
 945,991 117,147,328 123.84×
pipe
 621,062 92,852,057 149.51×
priv-instr
 890,210 7,264,930 8.16×
resched
 10,413,086 25,235,870 2.42×
switch
 87,003 7,047,796 81.01×
timermix
 238,715 29,489,230 123.53×
tlb-shootdown
 7,076 178,959 25.29×
tsc
 615,566 81,178,770 131.88×
usersyscall
 750,694 7,337,221 9.77×
vdso
 95,629,221 11,725,472,878 122.61×
vfork
 23,228 114,446 4.93×
x86syscall
 10,544,893 1,406,327,672 133.37×
yield
 163,375,833 194,465,920 1.19×
zero
 6,832,495 859,273,041 125.76×
CPU float/SIMDSingleMultiScaling
eigen
 17,169 1,721,273 100.25×
expmath
 3,736 423,702 113.42×
fma
 1,465,891 148,717,999 101.45×
fp
 326.39 38,726 118.65×
fractal
 7.4 836.15 112.99×
hyperbolic
 4,308 483,001 112.11×
logmath
 5,521 689,897 124.95×
matrix
 6,727 512,633 76.21×
matrix-3d
 677.25 15,831 23.38×
monte-carlo
 311.6 37,950 121.79×
mpfr
 1,834 148,284 80.84×
powmath
 2,059 245,496 119.25×
trig
 2,790 268,816 96.35×
veccmp
 27,306 2,178,313 79.77×
vecfp
 2,413 209,134 86.67×
vecmath
 4,159 451,814 108.63×
vecshuf
 4,080 604,237 148.08×
vecwide
 86,214 7,672,787 89.00×
AlgorithmsSingleMultiScaling
bitonicsort
 9.4 831.98 88.51×
bsearch
 293.97 28,706 97.65×
bubblesort
 1 83.2 83.20×
fibsearch
 327.2 32,046 97.94×
heapsort
 10.6 1,037 97.83×
hsearch
 9,057 773,043 85.36×
insertionsort
 13 908.19 69.86×
malloc
 2,774,397 122,688,758 44.22×
mcontend
 53.7 28,905 538.26×
mergesort
 32.4 3,259 100.59×
qsort
 23.2 2,290 98.72×
skiplist
 5,759 716,823 124.47×
tree
 1.6 124.53 77.83×
MemorySingleMultiScaling
cache
 288,733 8,617,625 29.85×
cachehammer
 203,827 8,517,613 41.79×
cacheline
 44.37 2,211 49.84×
l1cache
 1,179 23,520 19.95×
memcpy
 469.56 33,659 71.68×
memrate
 0.17 24.86 146.24×
memthrash
 469.01 1,017 2.17×
ptr-chase
 26,780,886 1,545,046,556 57.69×
spinmem
 253.19 8,379 33.09×
stream
 7.08 40,531 5724.78×
vm
 131,204 363,871 2.77×

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
a3-highgpu-1g 26234 GiB1
a3-highgpu-2g 52468 GiB2
a3-highgpu-4g 104936 GiB4
a3-highgpu-8g 2081872 GiB8
a3-edgegpu-8g 2081872 GiB8
a3-edgegpu-8g-nolssd 2081872 GiB8
a3-ultragpu-8g 2242952 GiB8

Similar Servers

INSTANCEVENDORvCPUsMEMORYAccelerators

a3-megagpu-8g FAQs