NC48ads_A100_v4 by Microsoft Azure

NC48ads_A100_v4 is a NCADSA100v4 family [AMD processor] [Local Disk] [Premium Storage capable] 48 vCPUs server offered by Microsoft Azure with 48 vCPUs, 440 GiB of memory and 2.05 TB of storage. The pricing starts at 1.3575 USD per hour.
48 vCPU
440 GiB Memory
2.05 TB Storage
2 GPU
Spare SCore
123,608
(All-cores)
2,913
(Single-core)

Specifications

Server Metadata

Vendor ID
azure
Server ID
Standard_NC48ads_A100_v4
Name
NC48ads_A100_v4
API Reference
Standard_NC48ads_A100_v4
Description
NCADSA100v4 family [AMD processor] [Local Disk] [Premium Storage capable] 48 vCPUs
Family
NCADSA100v4
Hw Virt
Average Time To Start
49
Status
active
Observed At
2026-10-31T08:29:24.782210

Availability

REGION / IDSPOTONDEMAND
East US (US) / eastus
1.3575 USD/h 7.346 USD/h
East US 2 (US) / eastus2
1.3575 USD/h 7.346 USD/h
West US 2 (US) / westus2
1.3575 USD/h 7.346 USD/h
West US 3 (US) / westus3
1.3575 USD/h 7.346 USD/h
Central US (US) / centralus
1.534 USD/h 8.301 USD/h
Canada Central (CA) / canadacentral
1.629 USD/h 8.815 USD/h
North Central US (US) / northcentralus
1.629 USD/h 8.815 USD/h
North Europe (IE) / northeurope
1.629 USD/h 8.815 USD/h
South Central US (US) / southcentralus
1.629 USD/h 8.815 USD/h
France Central (FR) / francecentral
1.697 USD/h 9.183 USD/h
UK South (GB) / uksouth
1.697 USD/h 9.183 USD/h
Germany West Central (DE) / germanywestcentral
1.7648 USD/h 9.55 USD/h
Italy North (IT) / italynorth
1.7648 USD/h 9.55 USD/h
Poland Central (PL) / polandcentral
1.7648 USD/h 9.55 USD/h
Southeast Asia (SG) / southeastasia
1.7648 USD/h 9.55 USD/h
Spain Central (ES) / spaincentral
1.7648 USD/h 9.55 USD/h
West Europe (NL) / westeurope
1.7648 USD/h 9.55 USD/h
West US (US) / westus
1.7648 USD/h 9.55 USD/h
Sweden Central (SE) / swedencentral
1.8747 USD/h 9.55 USD/h
Korea Central (KR) / koreacentral
1.8327 USD/h 9.917 USD/h
Central India (IN) / centralindia
1.9005 USD/h 10.284 USD/h
Jio India Central (IN) / jioindiacentral
- 10.284 USD/h
Jio India West (IN) / jioindiawest
- 10.284 USD/h
Switzerland North (CH) / switzerlandnorth
1.9413 USD/h 10.505 USD/h
Australia East (AU) / australiaeast
1.9685 USD/h 10.652 USD/h
Japan East (JP) / japaneast
1.9685 USD/h 10.652 USD/h
Brazil South (BR) / brazilsouth
2.7151 USD/h 14.692 USD/h

Processor

vCPUs
48
Hypervisor
Microsoft Hyper-V
CPU Allocation
Dedicated
CPU Cores
48
CPU Speed
3.7 GHz
CPU Architecture
x86_64
CPU Manufacturer
AMD
CPU Family
EPYC
CPU Model
7V13
CPU L1D Cache
32 KiB
CPU L1D Cache Total
2 MiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
2 MiB
CPU L2 Cache
512 KiB
CPU L2 Cache Total
24 MiB
CPU L3 Cache
32 MiB
CPU L3 Cache Total
192 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, rdseed, adx, smap, clflushopt, clwb, sha_ni, xsaveopt, xsavec, xgetbv1, xsaves, user_shstk, clzero, xsaveerptr, rdpru, arat, umip, vaes, vpclmulqdq, rdpid, fsrm
Ecpus
42.4
Scalability
88.33

System Resources and Accelerators

MEMORY
Memory Amount
440 GiB
Memory Amount Actual
440 GiB
ACCELERATORS
Accelerator Count
2
Accelerator Memory Min
80 GiB
Accelerator Memory Total
160 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Ampere
Accelerator Model
A100
Accelerators
  • NVIDIA Ampere NVIDIA A100 80GB PCIe (Memory amount: 81920, Firmware version: 595.71.05, BIOS version: 92.00.68.00.01, Clock rate: 1410)
  • NVIDIA Ampere NVIDIA A100 80GB PCIe (Memory amount: 81920, Firmware version: 595.71.05, BIOS version: 92.00.68.00.01, Clock rate: 1410)
STORAGE
Storage Size
2051 GB
Storage Type
nvme ssd
Storages
  • 131 GB ssd (temp disk)
  • 1920 GB nvme ssd
NETWORK
Inbound Traffic
0 GB/month
Outbound Traffic
0 GB/month
IPv4
0

CPU and System Topology

Server Description

A dual-GPU accelerated platform featuring AMD EPYC processors and local NVMe storage, optimized for large language model inference and parallel workloads.

GPU AcceleratedMemory OptimizedStorage & Database

Microsoft Azure NC48ads_A100_v4 is a GPU-accelerated server from the NCADSA100v4 family, featuring 48 dedicated AMD EPYC 7V13 physical cores running at 3.7 GHz. It is equipped with 440.0 GB of RAM, two NVIDIA A100 GPUs providing 160 GB of total VRAM, and a 2051 GB local NVMe SSD. Benchmarks indicate top-tier performance in LLM inference, with prompt processing reaching 501.25 tokens/sec for 70B models, alongside top-10% scores in image processing and memory performance. Multi-core CPU execution, ray tracing, and database operations also rank in the top 25%. However, large-block memory bandwidth and single-threaded text processing remain average. This server is optimized for deep learning, large-scale LLM inference, image processing, and database workloads.

Economics

Average Price per Region

Prices per Zone

Lowest Prices

Performance

Workload Profiles

2.3Score
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) 5,060,349 ops/sec 1,918,343 ops/sec 2.7 50.00% 50.00% +64.00%
Redis RPS (pipeline=16, SET) 41,365,760 ops/sec 11,704,734 ops/sec 3.4 20.00% 20.00% +28.00%
PassMark Memory Mark (composite) 3,324 2,416 1.4 10.00% 10.00% +3.40%
Memory bandwidth (read, 16 MB ~ L3) 192,880 MB/sec 108,466 MB/sec 1.7 10.00% 10.00% +5.40%
PassMark single-thread CPU 2,505 Mops/s 2,365 Mops/s 1 10.00% 10.00% 0.00%

GPU Bandwidth and Latency

Node (all GPUs)
Node baseline directional bandwidth
39.1 GB/s
Node GPU-to-host bandwidth
52.7 GB/s
Node host-to-GPU bandwidth
53.4 GB/s
Peer-to-peer
GPU-to-GPU P2P duplex bandwidth
260.7 GB/s
GPU-to-GPU P2P gather bandwidth
277.3 GB/s
GPU-to-GPU P2P latency
811 ns
GPU-to-GPU P2P bandwidth
277.3 GB/s
Single GPU slot
Per-GPU GPU-to-host bandwidth
26.3 GB/s
Per-GPU host-to-GPU bandwidth
26.7 GB/s
Per-GPU host-device latency
811 ns
Benchmark Environment
Driver Version
 535.183.01
CUDA Version (driver)
 12.6
nvbandwidth Version
 v0.5
GPU Count
 2
P2P Supported
Yes
Efficiency
SM/CE copy efficiency
1.00

Memory Bandwidth

Compression

OpenSSL

Geekbench Single-Core

Score: 1,726

Geekbench Multi-Core

Score: 17,886

Passmark CPU Scores

BENCHMARKSCORE
Mark
 64,635
Compression
 952,885
Encryption
 51,373
Extended Instructions
 66,673
Floating Point Maths
 193,430
Integer Maths
 238,843
Physics
 9,458
Prime Numbers
 728
Single Threaded
 2,505
String Sorting
 136,193

Passmark Memory Scores

BENCHMARKSCORE
Memory Mark
 3,324
Database Operations
 26,966
Memory Latency
 54
Memory Read Cached
 25,746
Memory Read Uncached
 22,985
Memory Write
 20,589

Stress-ng Stressors

CPU integerSingleMultiScaling
atomic
 531.03 178.24 0.34×
bitops
 151,337 6,017,494 39.76×
branch
 173,883 7,363,721 42.35×
funccall
 569.56 23,424 41.13×
funcret
 17,920 733,907 40.96×
hash
 147,376 6,012,574 40.80×
intmath
 62,109 2,615,065 42.10×
nop
 303,037 12,804,474 42.25×
prime
 418,994 16,392,571 39.12×
rotate
 81,941 3,497,360 42.68×
str
 839,032 34,100,806 40.64×
wcs
 619,180 21,870,974 35.32×
CryptographySingleMultiScaling
crypt
 3,510 130,998 37.33×
CompressionSingleMultiScaling
jpeg
 787.38 28,966 36.79×
zlib
 51 2,446 47.97×
HypervisorSingleMultiScaling
clock
 3,783 94,419 24.96×
clone
 943.45 22,426 23.77×
context
 3,718 153,371 41.25×
cpu-sched
 53.68 3,052 56.85×
eventfd
 25,848 2,230,285 86.29×
fault
 7,660 183,508 23.96×
fork
 3,573 56,037 15.68×
getrandom
 49,042 1,962,286 40.01×
hrtimers
 68,839 1,946,175 28.27×
mprotect
 615,205 10,068,790 16.37×
mtx
 50,310,169 4,469,958,600 88.85×
munmap
 2,628 54,217 20.63×
null
 461,670 19,193,209 41.57×
pipe
 1,705,071 20,909,669 12.26×
priv-instr
 427,073 13,682,376 32.04×
resched
 6,013,040 3,828,855 0.64×
switch
 64,136 8,382,926 130.71×
timermix
 42,396 1,988,153 46.90×
tlb-shootdown
 13,098 155,438 11.87×
tsc
 710,394 29,422,920 41.42×
usersyscall
 335,823 11,998,585 35.73×
vdso
 83,463,100 3,537,594,145 42.39×
vfork
 19,092 163,500 8.56×
x86syscall
 5,831,736 244,162,766 41.87×
yield
 38,391,810 31,713,658 0.83×
zero
 3,059,373 129,159,354 42.22×
CPU float/SIMDSingleMultiScaling
eigen
 14,617 569,645 38.97×
expmath
 2,515 105,076 41.79×
fma
 818,255 33,044,325 40.38×
fp
 241.32 10,078 41.76×
fractal
 7 283.25 40.46×
hyperbolic
 3,549 147,595 41.59×
logmath
 4,169 174,499 41.86×
matrix
 2,899 111,039 38.30×
matrix-3d
 96.45 3,265 33.85×
monte-carlo
 215.85 8,922 41.34×
mpfr
 1,502 58,127 38.69×
powmath
 1,247 52,522 42.12×
trig
 1,597 65,877 41.24×
veccmp
 17,343 662,976 38.23×
vecfp
 1,565 62,178 39.73×
vecmath
 3,662 153,086 41.80×
vecshuf
 567.33 22,102 38.96×
vecwide
 24,823 896,707 36.12×
AlgorithmsSingleMultiScaling
bitonicsort
 9.4 341.59 36.34×
bsearch
 279.98 11,224 40.09×
bubblesort
 0.8 38.4 48.00×
fibsearch
 343.59 14,261 41.51×
heapsort
 10 389.39 38.94×
hsearch
 8,319 309,749 37.24×
insertionsort
 12.4 401.19 32.35×
malloc
 2,817,167 74,737,343 26.53×
mcontend
 88.87 6,960 78.32×
mergesort
 29.8 1,178 39.54×
qsort
 19.8 758.78 38.32×
skiplist
 6,586 263,714 40.04×
tree
 2 57.59 28.80×
MemorySingleMultiScaling
cache
 507,153 1,042,421 2.06×
cachehammer
 301,994 9,914,937 32.83×
cacheline
 23.98 1,846 76.98×
l1cache
 155.94 6,942 44.52×
memcpy
 391.15 14,255 36.44×
memrate
 3.15 18.59 5.90×
memthrash
 106.2 917.81 8.64×
ptr-chase
 68,684,754 486,524,318 7.08×
spinmem
 260.79 1,735 6.65×
stream
 57.69 14,347 248.69×
vm
 178,309 298,913 1.68×

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
NC24ads_A100_v4 24220 GiB1
NC96ads_A100_v4 96880 GiB4

Similar Servers

INSTANCEVENDORvCPUsMEMORYAccelerators

NC48ads_A100_v4 FAQs