NC40ads_H100_v5 by Microsoft Azure

NC40ads_H100_v5 is a NCadsH100v5 family [AMD processor] [Local Disk] [Premium Storage capable] 40 vCPUs server offered by Microsoft Azure with 40 vCPUs, 320 GiB of memory and 3.97 TB of storage. The pricing starts at 1.2899 USD per hour.
40 vCPU
320 GiB Memory
3.97 TB Storage
1 GPU
Spare SCore
123,325
(All-cores)
3,086
(Single-core)

Specifications

Server Metadata

Vendor ID
azure
Server ID
Standard_NC40ads_H100_v5
Name
NC40ads_H100_v5
API Reference
Standard_NC40ads_H100_v5
Description
NCadsH100v5 family [AMD processor] [Local Disk] [Premium Storage capable] 40 vCPUs
Family
NCadsH100v5
Hw Virt
Average Time To Start
28
Status
active
Observed At
2026-10-25T04:45:15.499270

Availability

REGION / IDSPOTONDEMAND
East US (US) / eastus
1.2899 USD/h 6.98 USD/h
East US 2 (US) / eastus2
3.2164 USD/h 6.98 USD/h
West US 2 (US) / westus2
3.2164 USD/h 6.98 USD/h
West US 3 (US) / westus3
3.9549 USD/h 6.98 USD/h
Indonesia Central (ID) / indonesiacentral
1.6604 USD/h 8.167 USD/h
Malaysia West (MY) / malaysiawest
2.0074 USD/h 8.167 USD/h
North Central US (US) / northcentralus
1.5479 USD/h 8.376 USD/h
North Europe (IE) / northeurope
1.5479 USD/h 8.376 USD/h
South Central US (US) / southcentralus
2.1394 USD/h 8.38 USD/h
Central US (US) / centralus
1.5865 USD/h 8.585 USD/h
Canada Central (CA) / canadacentral
1.5874 USD/h 8.59 USD/h
Korea South (KR) / koreasouth
1.6124 USD/h 8.725 USD/h
UK South (GB) / uksouth
1.6124 USD/h 8.725 USD/h
Germany West Central (DE) / germanywestcentral
1.6769 USD/h 9.074 USD/h
Southeast Asia (SG) / southeastasia
1.6769 USD/h 9.074 USD/h
Sweden Central (SE) / swedencentral
1.7123 USD/h 9.074 USD/h
Spain Central (ES) / spaincentral
1.7903 USD/h 9.074 USD/h
West US (US) / westus
4.7711 USD/h 9.074 USD/h
West Europe (NL) / westeurope
3.3469 USD/h 9.08 USD/h
UK West (GB) / ukwest
1.6891 USD/h 9.14 USD/h
Korea Central (KR) / koreacentral
2.1428 USD/h 9.423 USD/h
Central India (IN) / centralindia
4.9642 USD/h 9.772 USD/h
Jio India Central (IN) / jioindiacentral
- 9.772 USD/h
Jio India West (IN) / jioindiawest
- 9.772 USD/h
UAE North (AE) / uaenorth
1.8443 USD/h 9.98 USD/h
Switzerland North (CH) / switzerlandnorth
2.115 USD/h 9.981 USD/h
Australia East (AU) / australiaeast
1.8704 USD/h 10.121 USD/h
Japan East (JP) / japaneast
3.2833 USD/h 10.121 USD/h

Processor

vCPUs
40
Hypervisor
Microsoft Hyper-V
CPU Allocation
Dedicated
CPU Cores
40
CPU Speed
3.7 GHz
CPU Architecture
x86_64
CPU Manufacturer
AMD
CPU Family
EPYC
CPU Model
9V84
CPU L1D Cache
32 KiB
CPU L1D Cache Total
1 MiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
1 MiB
CPU L2 Cache
1 MiB
CPU L2 Cache Total
40 MiB
CPU L3 Cache
32 MiB
CPU L3 Cache Total
160 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, tsc_reliable, nonstop_tsc, cpuid, extd_apicid, aperfmperf, pni, pclmulqdq, ssse3, fma, cx16, pcid, sse4_1, sse4_2, movbe, popcnt, aes, xsave, avx, f16c, rdrand, hypervisor, lahf_lm, cmp_legacy, cr8_legacy, abm, sse4a, misalignsse, 3dnowprefetch, osvw, topoext, perfctr_core, vmmcall, fsgsbase, bmi1, avx2, smep, bmi2, erms, invpcid, avx512f, avx512dq, rdseed, adx, smap, avx512ifma, clflushopt, clwb, avx512cd, sha_ni, avx512bw, avx512vl, xsaveopt, xsavec, xgetbv1, xsaves, user_shstk, avx512_bf16, clzero, xsaveerptr, rdpru, arat, avx512vbmi, umip, avx512_vbmi2, gfni, vaes, vpclmulqdq, avx512_vnni, avx512_bitalg, avx512_vpopcntdq, rdpid, fsrm
Ecpus
40
Scalability
100

System Resources and Accelerators

MEMORY
Memory Amount
320 GiB
Memory Amount Actual
320 GiB
ACCELERATORS
Accelerator Count
1
Accelerator Memory Min
94 GiB
Accelerator Memory Total
94 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Hopper
Accelerator Model
H100
Accelerators
  • NVIDIA Hopper NVIDIA H100 NVL (Memory amount: 95830, Firmware version: 595.71.05, BIOS version: 96.00.9F.00.04, Clock rate: 1785)
STORAGE
Storage Size
3971 GB
Storage Type
nvme ssd
Storages
  • 131 GB ssd (temp disk)
  • 3840 GB nvme ssd
NETWORK
Inbound Traffic
0 GB/month
Outbound Traffic
0 GB/month
IPv4
0

CPU and System Topology

Server Description

A high-density GPU server combining dedicated AMD EPYC processors, local NVMe storage, and Hopper acceleration for demanding machine learning workloads.

GPU AcceleratedCompute OptimizedStorage & Database

Microsoft Azure NC40ads_H100_v5 is a high-performance, GPU-accelerated server featuring 40 dedicated AMD EPYC 9V84 vCPUs operating at 3.7 GHz, 320.0 GB of system memory, and a bundled 3971 GB NVMe SSD. It integrates a single NVIDIA Hopper H100 GPU with 94 GB of VRAM. Benchmarks indicate top-tier performance in Geekbench multi-core and image processing, alongside strong CPU and memory scores. The server achieves 100.89 million operations per second in Redis SET tests and demonstrates strong text generation speeds for small and medium LLMs. While large-block read memory bandwidth and static web serving throughput are average, the overall resource density makes this server highly efficient for machine learning inference, data processing, and database workloads.

Economics

Average Price per Region

Prices per Zone

Lowest Prices

Performance

Workload Profiles

2.6Score
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) 5,762,338 ops/sec 1,919,289 ops/sec 3.1 50.00% 50.00% +76.00%
Redis RPS (pipeline=16, SET) 50,913,350 ops/sec 11,728,058 ops/sec 4.2 20.00% 20.00% +33.00%
PassMark Memory Mark (composite) 3,177 2,416 1.3 10.00% 10.00% +2.70%
Memory bandwidth (read, 16 MB ~ L3) 234,506 MB/sec 108,466 MB/sec 2.1 10.00% 10.00% +7.70%
PassMark single-thread CPU 2,907 Mops/s 2,365 Mops/s 1.2 10.00% 10.00% +1.80%

GPU Bandwidth and Latency

Node (all GPUs)
Node baseline directional bandwidth
49.4 GB/s
Node GPU-to-host bandwidth
57.3 GB/s
Node host-to-GPU bandwidth
57.6 GB/s
Single GPU slot
Per-GPU GPU-to-host bandwidth
57.3 GB/s
Per-GPU host-to-GPU bandwidth
57.6 GB/s
Per-GPU host-device latency
834 ns
Benchmark Environment
Driver Version
 535.230.02
CUDA Runtime Version
 12.2
nvbandwidth Version
 v0.5
GPU Count
 1
P2P Supported
No
Efficiency
SM/CE copy efficiency
0.93

Memory Bandwidth

Compression

OpenSSL

Geekbench Single-Core

Score: 2,140

Geekbench Multi-Core

Score: 23,376

Passmark CPU Scores

BENCHMARKSCORE
Mark
 73,108
Compression
 1,058,740
Encryption
 52,570
Extended Instructions
 100,393
Floating Point Maths
 212,985
Integer Maths
 242,522
Physics
 10,739
Prime Numbers
 794
Single Threaded
 2,907
String Sorting
 152,284

Passmark Memory Scores

BENCHMARKSCORE
Memory Mark
 3,177
Database Operations
 30,607
Memory Latency
 65
Memory Read Cached
 27,953
Memory Read Uncached
 27,776
Memory Write
 25,954

Stress-ng Stressors

CPU integerSingleMultiScaling
atomic
 918.44 335.49 0.37×
bitops
 160,688 6,369,441 39.64×
branch
 213,081 8,441,804 39.62×
funccall
 618.37 24,680 39.91×
funcret
 20,866 829,930 39.77×
hash
 163,594 6,486,017 39.65×
intmath
 67,781 2,651,675 39.12×
nop
 629,294 25,070,386 39.84×
prime
 490,775 19,379,552 39.49×
rotate
 88,449 3,519,346 39.79×
str
 1,071,435 43,368,488 40.48×
wcs
 697,949 27,975,857 40.08×
CryptographySingleMultiScaling
crypt
 3,719 145,901 39.23×
CompressionSingleMultiScaling
jpeg
 912.64 34,408 37.70×
zlib
 49.88 2,423 48.57×
HypervisorSingleMultiScaling
clock
 22,073 211,909 9.60×
clone
 1,088 20,210 18.58×
context
 3,063 121,277 39.59×
cpu-sched
 72.15 2,534 35.12×
eventfd
 34,547 3,130,787 90.63×
fault
 11,104 203,718 18.35×
fork
 4,069 60,058 14.76×
getrandom
 52,307 2,082,657 39.82×
hrtimers
 226,826 4,946,060 21.81×
mprotect
 695,939 9,833,015 14.13×
mtx
 35,679,950 5,295,185,204 148.41×
munmap
 3,062 52,206 17.05×
null
 352,879 13,847,470 39.24×
pipe
 1,254,596 19,823,368 15.80×
priv-instr
 450,824 14,454,577 32.06×
resched
 6,765,877 4,010,771 0.59×
switch
 104,779 8,545,757 81.56×
timermix
 183,313 7,214,061 39.35×
tlb-shootdown
 12,948 140,901 10.88×
tsc
 786,532 31,013,547 39.43×
usersyscall
 334,499 11,684,390 34.93×
vdso
 94,589,090 3,735,531,702 39.49×
vfork
 23,322 223,562 9.59×
x86syscall
 4,175,298 165,498,428 39.64×
yield
 28,825,762 33,609,972 1.17×
zero
 2,789,515 109,637,312 39.30×
CPU float/SIMDSingleMultiScaling
eigen
 16,384 643,365 39.27×
expmath
 3,072 123,610 40.23×
fma
 919,722 36,147,107 39.30×
fp
 252 10,011 39.73×
fractal
 7.4 291.49 39.39×
hyperbolic
 3,825 152,095 39.76×
logmath
 4,568 180,641 39.55×
matrix
 4,968 195,536 39.36×
matrix-3d
 149.26 5,095 34.14×
monte-carlo
 182.16 7,269 39.90×
mpfr
 1,701 67,630 39.76×
powmath
 1,237 48,394 39.12×
trig
 1,846 74,596 40.41×
veccmp
 19,051 758,043 39.79×
vecfp
 1,815 71,616 39.45×
vecmath
 4,041 159,630 39.50×
vecshuf
 639.88 25,484 39.83×
vecwide
 28,642 1,137,793 39.73×
AlgorithmsSingleMultiScaling
bitonicsort
 10.4 412.79 39.69×
bsearch
 306.21 12,129 39.61×
bubblesort
 0.8 32.2 40.25×
fibsearch
 381.04 15,155 39.77×
heapsort
 10.6 424.59 40.06×
hsearch
 11,437 448,121 39.18×
insertionsort
 13.2 521.4 39.50×
malloc
 3,517,399 80,757,440 22.96×
mcontend
 93.89 6,217 66.22×
mergesort
 34.2 1,353 39.56×
qsort
 22 872.59 39.66×
skiplist
 7,118 282,095 39.63×
tree
 2.2 62.59 28.45×
MemorySingleMultiScaling
cache
 890,980 1,123,434 1.26×
cachehammer
 293,033 9,789,344 33.41×
cacheline
 50.98 3,911 76.71×
l1cache
 283.4 10,379 36.62×
memcpy
 448.98 17,872 39.81×
memrate
 3.74 21.67 5.79×
memthrash
 202.93 1,572 7.75×
ptr-chase
 61,627,308 432,299,366 7.01×
spinmem
 297.38 1,959 6.59×
stream
 136.57 29,476 215.83×
vm
 201,783 297,829 1.48×

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
NC80adis_H100_v5 80640 GiB2

Similar Servers

INSTANCEVENDORvCPUsMEMORYAccelerators

NC40ads_H100_v5 FAQs