D48a_v4 by Microsoft Azure

D48a_v4 is a DAv4 family [AMD processor] 48 vCPUs server offered by Microsoft Azure with 48 vCPUs, 192 GiB of memory and 1.23 TB of storage. The pricing starts at 0.2737 USD per hour.
48 vCPU
192 GiB Memory
1.23 TB Storage
Spare SCore
64,906
(All-cores)
2,707
(Single-core)

Specifications

Server Metadata

Vendor ID
azure
Server ID
Standard_D48a_v4
Name
D48a_v4
API Reference
Standard_D48a_v4
Description
DAv4 family [AMD processor] 48 vCPUs
Family
DAv4
Hw Virt
Average Time To Start
19
Status
active
Observed At
2026-11-01T20:55:40.156940

Availability

REGION / IDSPOTONDEMAND
North Central US (US) / northcentralus
0.4258 USD/h -
West US 2 (US) / westus2
0.4258 USD/h -
Poland Central (PL) / polandcentral
0.4613 USD/h -
North Europe (IE) / northeurope
0.4746 USD/h -
Indonesia Central (ID) / indonesiacentral
0.479 USD/h -
Korea South (KR) / koreasouth
0.4912 USD/h -
Israel Central (IL) / israelcentral
0.4967 USD/h -
Italy North (IT) / italynorth
0.4967 USD/h -
Canada Central (CA) / canadacentral
0.4982 USD/h -
New Zealand North (NZ) / newzealandnorth
0.503 USD/h -
UAE Central (AE) / uaecentral
0.6804 USD/h -
Switzerland West (CH) / switzerlandwest
0.7915 USD/h -
Central India (IN) / centralindia
0.2737 USD/h 1.481 USD/h
Jio India Central (IN) / jioindiacentral
- 1.481 USD/h
India South Central (IN) / indiasouthcentral
0.4032 USD/h 2.182 USD/h
East US 2 (US) / eastus2
0.4258 USD/h 2.304 USD/h
East US (US) / eastus
0.4859 USD/h 2.304 USD/h
West US 3 (US) / westus3
- 2.304 USD/h
Spain Central (ES) / spaincentral
0.448 USD/h 2.424 USD/h
Sweden Central (SE) / swedencentral
0.4524 USD/h 2.448 USD/h
Austria East (AT) / austriaeast
0.4613 USD/h 2.496 USD/h
Belgium Central (BE) / belgiumcentral
0.4613 USD/h 2.496 USD/h
Denmark East (DK) / denmarkeast
0.4613 USD/h 2.496 USD/h
Germany West Central (DE) / germanywestcentral
0.4613 USD/h 2.496 USD/h
Mexico Central (ZA) / mexicocentral
- 2.534 USD/h
Canada East (CA) / canadaeast
0.4761 USD/h 2.568 USD/h
Malaysia West (MY) / malaysiawest
0.479 USD/h 2.592 USD/h
Central US (US) / centralus
0.481 USD/h 2.604 USD/h
UK South (GB) / uksouth
0.4923 USD/h 2.664 USD/h
France Central (FR) / francecentral
0.4967 USD/h 2.688 USD/h
West US (US) / westus
- 2.688 USD/h
West Europe (NL) / westeurope
0.5608 USD/h 2.76 USD/h
West Central US (US) / westcentralus
0.511 USD/h 2.765 USD/h
UK West (GB) / ukwest
0.5152 USD/h 2.788 USD/h
Qatar Central (QA) / qatarcentral
0.5221 USD/h 2.825 USD/h
Korea Central (KR) / koreacentral
0.5234 USD/h 2.832 USD/h
Australia Central 2 (AU) / australiacentral2
0.288 USD/h 2.88 USD/h
Australia East (AU) / australiaeast
0.5322 USD/h 2.88 USD/h
Southeast Asia (SG) / southeastasia
0.5322 USD/h 2.88 USD/h
Australia Central (AU) / australiacentral
- 2.88 USD/h
Chile Central (CL) / chilecentral
- 2.903 USD/h
West India (IN) / westindia
0.545 USD/h 2.949 USD/h
Japan East (JP) / japaneast
0.55 USD/h 2.976 USD/h
Japan West (JP) / japanwest
0.55 USD/h 2.976 USD/h
Australia Southeast (AU) / australiasoutheast
0.5535 USD/h 2.995 USD/h
South Africa North (ZA) / southafricanorth
0.5633 USD/h 3.048 USD/h
East Asia (HK) / eastasia
- 3.168 USD/h
South India (IN) / southindia
0.6004 USD/h 3.249 USD/h
Norway East (NO) / norwayeast
- 3.295 USD/h
Switzerland North (CH) / switzerlandnorth
- 3.295 USD/h
France South (FR) / francesouth
0.6457 USD/h 3.494 USD/h
Brazil South (BR) / brazilsouth
0.6808 USD/h 3.672 USD/h
South Africa West (ZA) / southafricawest
0.7322 USD/h 3.962 USD/h
Norway West (NO) / norwaywest
0.7915 USD/h 4.283 USD/h
Brazil Southeast (US) / brazilsoutheast
0.8822 USD/h 4.774 USD/h

Processor

vCPUs
48
Hypervisor
Microsoft Hyper-V
CPU Allocation
Dedicated
CPU Cores
24
CPU Speed
3.525 GHz
CPU Architecture
x86_64
CPU Manufacturer
AMD
CPU Family
EPYC
CPU Model
7763
CPU L1D Cache
32 KiB
CPU L1D Cache Total
768 KiB
CPU L1I Cache
32 KiB
CPU L1I Cache Total
768 KiB
CPU L2 Cache
512 KiB
CPU L2 Cache Total
12 MiB
CPU L3 Cache
32 MiB
CPU L3 Cache Total
96 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, rep_good, nopl, cpuid, extd_apicid, 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, osvw, topoext, vmmcall, fsgsbase, bmi1, avx2, smep, bmi2, erms, rdseed, adx, smap, clflushopt, clwb, sha_ni, xsaveopt, xsavec, xgetbv1, xsaves, xsaveerptr, arat, umip, vaes, vpclmulqdq, rdpid, fsrm
Ecpus
24
Scalability
50

System Resources and Accelerators

MEMORY
Memory Amount
192 GiB
Memory Amount Actual
192 GiB
ACCELERATORS
Accelerator Count
0
Accelerator Memory Min
0 MiB
Accelerator Memory Total
0 MiB
Accelerators
    STORAGE
    Storage Size
    1229 GB
    Storage Type
    ssd
    Storages
    • 1229 GB ssd (temp disk)
    NETWORK
    Inbound Traffic
    0 GB/month
    Outbound Traffic
    0 GB/month
    IPv4
    0

    CPU and System Topology

    Server Description

    A balanced x86_64 server featuring forty-eight dedicated AMD EPYC vCPUs, high-capacity local SSD storage, and strong multi-core processing capabilities.

    General PurposeStorage & Database

    Microsoft Azure D48a_v4 is a general-purpose server powered by an AMD EPYC 7763 processor with 48 dedicated vCPUs, 24 physical cores, and 192.0 GB of RAM. Running on Microsoft Hyper-V, it includes a bundled 1229 GB local SSD. Performance benchmarks show strong multi-core capabilities, with stress-ng and Geekbench multi-core scores in the top 25%, and a Passmark memory score in the top 10%. However, memory bandwidth and large-model LLM text generation are limited, with the 70B model text generation falling into the bottom 25%. This instance is qualitatively cost-efficient for multi-threaded processing tasks, such as image processing, ray tracing, and data decompression, but is not optimal for large-scale LLM inference.

    Economics

    Average Price per Region

    Prices per Zone

    Lowest Prices

    Performance

    Workload Profiles

    1.7Score
    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) 3,779,091 ops/sec 1,918,343 ops/sec 2 50.00% 50.00% +41.00%
    Redis RPS (pipeline=16, SET) 27,626,947 ops/sec 11,704,734 ops/sec 2.3 20.00% 20.00% +18.00%
    PassMark Memory Mark (composite) 3,005 2,416 1.2 10.00% 10.00% +1.80%
    Memory bandwidth (read, 16 MB ~ L3) 106,527 MB/sec 108,466 MB/sec 1 10.00% 10.00% 0.00%
    PassMark single-thread CPU 2,396 Mops/s 2,365 Mops/s 1 10.00% 10.00% 0.00%

    Memory Bandwidth

    Compression

    OpenSSL

    Geekbench Single-Core

    Score: 1,613

    Geekbench Multi-Core

    Score: 16,290

    Passmark CPU Scores

    BENCHMARKSCORE
    Mark
    		 52,695
    Compression
    		 760,175
    Encryption
    		 51,713
    Extended Instructions
    		 44,984
    Floating Point Maths
    		 110,080
    Integer Maths
    		 199,886
    Physics
    		 7,153
    Prime Numbers
    		 380
    Single Threaded
    		 2,396
    String Sorting
    		 106,101

    Passmark Memory Scores

    BENCHMARKSCORE
    Memory Mark
    		 3,005
    Database Operations
    		 20,103
    Memory Latency
    		 62
    Memory Read Cached
    		 24,472
    Memory Read Uncached
    		 20,738
    Memory Write
    		 20,346

    Stress-ng Stressors

    CPU integerSingleMultiScaling
    atomic
    		 506.64 170.59 0.34×
    bitops
    		 141,916 3,870,413 27.27×
    branch
    		 164,746 7,209,898 43.76×
    funccall
    		 540.62 16,718 30.92×
    funcret
    		 17,171 519,073 30.23×
    hash
    		 135,581 4,946,941 36.49×
    intmath
    		 56,997 1,872,050 32.84×
    nop
    		 283,158 6,788,399 23.97×
    prime
    		 393,955 11,528,447 29.26×
    rotate
    		 77,296 1,844,665 23.86×
    str
    		 826,649 25,097,480 30.36×
    wcs
    		 579,338 15,726,731 27.15×
    CryptographySingleMultiScaling
    crypt
    		 3,237 75,821 23.43×
    CompressionSingleMultiScaling
    jpeg
    		 743.45 23,970 32.24×
    zlib
    		 71.32 2,139 29.99×
    HypervisorSingleMultiScaling
    clock
    		 3,767 91,351 24.25×
    clone
    		 1,066 20,730 19.45×
    context
    		 3,423 137,124 40.06×
    cpu-sched
    		 69.69 3,109 44.61×
    eventfd
    		 25,423 1,462,915 57.54×
    fault
    		 7,345 171,141 23.30×
    fork
    		 4,382 50,864 11.61×
    getrandom
    		 45,311 1,379,415 30.44×
    hrtimers
    		 69,402 1,813,741 26.13×
    mprotect
    		 668,432 8,460,516 12.66×
    mtx
    		 35,356,532 2,627,274,629 74.31×
    munmap
    		 2,850 46,071 16.16×
    null
    		 434,141 16,822,900 38.75×
    pipe
    		 1,708,497 20,825,167 12.19×
    priv-instr
    		 394,954 14,046,435 35.56×
    resched
    		 6,306,002 2,249,150 0.36×
    switch
    		 59,812 7,174,031 119.94×
    timermix
    		 42,145 1,992,917 47.29×
    tlb-shootdown
    		 11,826 108,016 9.13×
    tsc
    		 668,517 30,524,301 45.66×
    usersyscall
    		 313,003 11,256,432 35.96×
    vdso
    		 73,732,160 2,886,391,287 39.15×
    vfork
    		 23,147 173,624 7.50×
    x86syscall
    		 5,423,772 209,170,273 38.57×
    yield
    		 27,344,352 26,612,130 0.97×
    zero
    		 2,902,772 103,530,011 35.67×
    CPU float/SIMDSingleMultiScaling
    eigen
    		 14,038 343,968 24.50×
    expmath
    		 2,503 91,598 36.60×
    fma
    		 779,041 29,274,859 37.58×
    fp
    		 227.84 7,866 34.52×
    fractal
    		 6.8 265.74 39.08×
    hyperbolic
    		 3,298 119,244 36.16×
    logmath
    		 3,873 132,631 34.25×
    matrix
    		 2,720 95,647 35.16×
    matrix-3d
    		 87.81 3,427 39.03×
    monte-carlo
    		 204.53 7,702 37.66×
    mpfr
    		 1,404 40,702 29.00×
    powmath
    		 1,194 46,773 39.16×
    trig
    		 1,497 55,005 36.75×
    veccmp
    		 16,063 502,508 31.28×
    vecfp
    		 1,480 49,482 33.43×
    vecmath
    		 3,438 131,594 38.28×
    vecshuf
    		 534.31 14,836 27.77×
    vecwide
    		 23,694 718,275 30.31×
    AlgorithmsSingleMultiScaling
    bitonicsort
    		 8.8 219.39 24.93×
    bsearch
    		 263.41 8,314 31.56×
    bubblesort
    		 0.8 28.4 35.50×
    fibsearch
    		 321.55 9,141 28.43×
    heapsort
    		 9.4 271.59 28.89×
    hsearch
    		 7,670 210,049 27.38×
    insertionsort
    		 11.6 260.2 22.43×
    malloc
    		 3,140,514 67,014,765 21.34×
    mcontend
    		 94.12 6,792 72.16×
    mergesort
    		 27.4 967.56 35.31×
    qsort
    		 18 595.18 33.07×
    skiplist
    		 6,230 225,150 36.14×
    tree
    		 1.8 39.8 22.11×
    MemorySingleMultiScaling
    cache
    		 479,263 1,253,223 2.61×
    cachehammer
    		 242,581 8,343,291 34.39×
    cacheline
    		 24.37 2,164 88.78×
    l1cache
    		 134.08 3,720 27.74×
    memcpy
    		 367.27 8,255 22.48×
    memrate
    		 2.75 13.95 5.07×
    memthrash
    		 130.21 699.24 5.37×
    ptr-chase
    		 65,489,108 387,300,604 5.91×
    spinmem
    		 272.79 1,557 5.71×
    stream
    		 29.16 6,765 232.01×
    vm
    		 148,064 298,035 2.01×

    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
    D2a_v4 28 GiB0
    D4a_v4 416 GiB0
    D8a_v4 832 GiB0
    D16a_v4 1664 GiB0
    D32a_v4 32128 GiB0
    D64a_v4 64256 GiB0
    D96a_v4 96384 GiB0

    Similar Servers

    INSTANCEVENDORvCPUsMEMORYAccelerators

    D48a_v4 FAQs