m9g.metal-48xl by Amazon Web Services

m9g.metal-48xl is a General purpose [AWS Graviton processors] Gen9 metal-48xl server offered by Amazon Web Services with 192 vCPUs, 768 GiB of memory and 0 GB of storage. The pricing starts at 0.9393 USD per hour.
192 vCPU
768 GiB Memory
Spare SCore
382,009
(All-cores)
2,010
(Single-core)

Specifications

Server Metadata

Vendor ID
aws
Name
m9g.metal-48xl
Description
General purpose [AWS Graviton processors] Gen9 metal-48xl
Family
m9g
Hw Virt
Average Time To Start
19.67
Status
active
Observed At
2026-10-28T16:57:27.612764

Availability

REGION / IDSPOTONDEMAND
Ohio (US) / us-east-2
1.187667 USD/h 9.3926 USD/h
Northern Virgina (US) / us-east-1
2.39055 USD/h 9.3926 USD/h
Oregon (US) / us-west-2
3.096575 USD/h 9.3926 USD/h
(MY) / ap-southeast-5
3.0347 USD/h 9.9802 USD/h
Aragón (ES) / eu-south-2
1.8705 USD/h 10.4621 USD/h
Dublin (IE) / eu-west-1
2.697033 USD/h 10.4678 USD/h
Frankfurt (DE) / eu-central-1
1.977433 USD/h 11.2531 USD/h
Sydney (AU) / ap-southeast-2
1.3393 USD/h 11.7408 USD/h
Singapore (SG) / ap-southeast-1
4.275233 USD/h 11.7408 USD/h
Tokyo (JP) / ap-northeast-1
2.5837 USD/h 12.1325 USD/h

Processor

vCPUs
192
CPU Allocation
Dedicated
CPU Cores
192
CPU Speed
3.3 GHz
CPU Architecture
arm64
CPU Manufacturer
AWS
CPU Family
ARMv9
CPU Model
Neoverse-V3
CPU L1D Cache
64 KiB
CPU L1D Cache Total
12 MiB
CPU L1I Cache
64 KiB
CPU L1I Cache Total
12 MiB
CPU L2 Cache
2 MiB
CPU L2 Cache Total
384 MiB
CPU L3 Cache
96 MiB
CPU L3 Cache Total
192 MiB
CPU Flags
fp, asimd, evtstrm, aes, pmull, sha1, sha2, crc32, atomics, fphp, asimdhp, cpuid, asimdrdm, jscvt, fcma, lrcpc, dcpop, sha3, asimddp, sha512, sve, asimdfhm, dit, uscat, ilrcpc, flagm, sb, paca, pacg, dcpodp, sve2, sveaes, svepmull, svebitperm, svesha3, flagm2, frint, svei8mm, svebf16, i8mm, bf16, dgh, rng, bti, ecv, afp, wfxt
Ecpus
190
Scalability
98.96

System Resources and Accelerators

MEMORY
Memory Amount
768 GiB
Memory Amount Actual
768 GiB
Memory Generation
DDR5
ACCELERATORS
Accelerator Count
0
Accelerator Memory Min
0 MiB
Accelerator Memory Total
0 MiB
Accelerators
    STORAGE
    Storage Size
    0 GB
    Storages
      NETWORK
      Network Speed Baseline
      100 Gbps
      Network Speed Max
      100 Gbps
      Network Storage Speed Baseline
      72 Gbps
      Network Storage Speed Max
      72 Gbps
      Inbound Traffic
      0 GB/month
      Outbound Traffic
      0 GB/month
      IPv4
      0

      CPU and System Topology

      Server Description

      A bare-metal ARM64 platform featuring 192 dedicated physical cores and DDR5 memory for highly parallelized, high-throughput enterprise workloads.

      General PurposeCompute Optimized

      Amazon Web Services m9g.metal-48xl is a bare-metal general-purpose server featuring 192 dedicated ARM64 physical cores powered by AWS Graviton Neoverse-V3 processors at 3.3 GHz. It is equipped with 768.0 GB of DDR5 memory, providing a balanced 4.0 GB per core, and offers a baseline network bandwidth of 100 Gbps. The server lacks local storage and GPU accelerators. In benchmarks, it delivers top-tier multi-core CPU and memory bandwidth performance, alongside high-throughput database and static web serving capabilities. However, single-core CPU performance is average, and LLM text generation speeds are relatively low. This hardware profile is configured for highly parallelized workloads, multi-threaded applications, large-scale database operations, and high-volume web serving.

      Economics

      Average Price per Region

      Prices per Zone

      Lowest Prices

      Performance

      Workload Profiles

      15Score
      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) 64,977,916 ops/sec 1,919,289 ops/sec 34 50.00% 50.00% +480.00%
      Redis RPS (pipeline=16, SET) 386,443,977 ops/sec 11,728,058 ops/sec 32 20.00% 20.00% +100.00%
      PassMark Memory Mark (composite) 4,354 2,416 1.8 10.00% 10.00% +6.10%
      Memory bandwidth (read, 16 MB ~ L3) 846,468 MB/sec 108,466 MB/sec 7.7 10.00% 10.00% +23.00%
      PassMark single-thread CPU 2,474 Mops/s 2,365 Mops/s 1 10.00% 10.00% 0.00%

      Memory Bandwidth

      Compression

      OpenSSL

      Passmark CPU Scores

      BENCHMARKSCORE
      Mark
      		 107,872
      Compression
      		 1,788,856
      Encryption
      		 56,703
      Extended Instructions
      		 244,429
      Floating Point Maths
      		 1,463,908
      Integer Maths
      		 1,172,073
      Physics
      		 42,561
      Prime Numbers
      		 2,291
      Single Threaded
      		 2,474
      String Sorting
      		 909,594

      Passmark Memory Scores

      BENCHMARKSCORE
      Memory Mark
      		 4,354
      Database Operations
      		 62,632
      Memory Latency
      		 31
      Memory Read Cached
      		 22,335
      Memory Read Uncached
      		 24,683
      Memory Write
      		 25,137

      Stress-ng Stressors

      CPU integerSingleMultiScaling
      atomic
      		 138.07 - -
      bitops
      		 165,698 11,042,821 66.64×
      branch
      		 318,086 22,900,810 72.00×
      funccall
      		 2,289 434,570 189.84×
      funcret
      		 40,009 7,054,070 176.31×
      hash
      		 181,976 34,587,342 190.07×
      intmath
      		 39,813 7,537,444 189.32×
      nop
      		 427,616 82,032,216 191.84×
      prime
      		 601,858 114,787,969 190.72×
      rotate
      		 149,638 28,360,778 189.53×
      str
      		 938,699 178,510,267 190.17×
      wcs
      		 285,141 52,735,048 184.94×
      CryptographySingleMultiScaling
      crypt
      		 4,377 802,373 183.31×
      CompressionSingleMultiScaling
      jpeg
      		 922.52 113,394 122.92×
      zlib
      		 37.14 9,066 244.11×
      HypervisorSingleMultiScaling
      clock
      		 45,449 149,237 3.28×
      clone
      		 498.19 23,021 46.21×
      context
      		 3,968 759,247 191.36×
      cpu-sched
      		 62.56 10,957 175.14×
      eventfd
      		 150,766 4,324,921 28.69×
      fault
      		 28,275 303,775 10.74×
      fork
      		 5,659 75,673 13.37×
      getrandom
      		 74,302 14,126,811 190.13×
      hrtimers
      		 207,512 5,916,315 28.51×
      mprotect
      		 995,162 14,696,269 14.77×
      mtx
      		 13,207,082 10,193,733,740 771.84×
      munmap
      		 4,260 104,729 24.58×
      null
      		 724,010 130,861,894 180.75×
      pipe
      		 973,655 85,206,550 87.51×
      priv-instr
      		 1,101,340 7,114,512 6.46×
      resched
      		 1,152,037 22,908,694 19.89×
      switch
      		 289,586 120,560,339 416.32×
      timermix
      		 976,927 180,282,876 184.54×
      tlb-shootdown
      		 15,640 213,703 13.66×
      vdso
      		 64,371,177 11,934,971,020 185.41×
      vfork
      		 41,901 279,197 6.66×
      yield
      		 358,369,861 365,572,244 1.02×
      zero
      		 3,273,777 618,915,212 189.05×
      CPU float/SIMDSingleMultiScaling
      eigen
      		 2,252 426,593 189.41×
      expmath
      		 426 80,839 189.76×
      fma
      		 856,054 163,142,108 190.57×
      fp
      		 175.2 33,569 191.60×
      fractal
      		 9.2 1,707 185.52×
      hyperbolic
      		 409.52 77,334 188.84×
      logmath
      		 492.75 93,791 190.34×
      matrix
      		 8,021 1,521,001 189.63×
      matrix-3d
      		 1,477 34,574 23.41×
      monte-carlo
      		 280.07 51,310 183.20×
      mpfr
      		 1,893 354,945 187.49×
      powmath
      		 282.93 53,590 189.41×
      trig
      		 555.54 105,109 189.20×
      veccmp
      		 26,427 4,918,052 186.10×
      vecfp
      		 1,867 350,291 187.66×
      vecmath
      		 4,830 916,512 189.75×
      vecshuf
      		 1,031 193,691 187.78×
      vecwide
      		 30,282 6,154,362 203.23×
      AlgorithmsSingleMultiScaling
      bitonicsort
      		 3.2 758.97 237.18×
      bsearch
      		 423.92 37,509 88.48×
      bubblesort
      		 1.4 268.8 192.00×
      fibsearch
      		 415.72 78,805 189.56×
      heapsort
      		 15.4 2,910 188.97×
      hsearch
      		 10,892 2,052,885 188.47×
      insertionsort
      		 7.6 1,502 197.60×
      malloc
      		 4,279,002 310,377,957 72.54×
      mcontend
      		 166.12 497,937 2997.45×
      mergesort
      		 38.6 7,362 190.72×
      qsort
      		 26.8 5,121 191.08×
      skiplist
      		 7,686 1,424,444 185.33×
      tree
      		 1.8 191.57 106.43×
      MemorySingleMultiScaling
      cache
      		 567,286 7,596,963 13.39×
      cachehammer
      		 2,186,351 19,674,568 9.00×
      cacheline
      		 51.37 4,176 81.30×
      l1cache
      		 390.17 60,536 155.15×
      memcpy
      		 994.73 174,396 175.32×
      memrate
      		 6.92 97.06 14.03×
      memthrash
      		 498.6 365.19 0.73×
      ptr-chase
      		 28,440,007 1,743,786,538 61.31×
      spinmem
      		 3,308 177,135 53.55×
      stream
      		 26.69 236,036 8843.61×
      vm
      		 197,036 5,420,865 27.51×

      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
      m9g.medium 14 GiB0
      m9g.large 28 GiB0
      m9g.xlarge 416 GiB0
      m9g.2xlarge 832 GiB0
      m9g.4xlarge 1664 GiB0
      m9g.8xlarge 32128 GiB0
      m9g.12xlarge 48192 GiB0

      Similar Servers

      INSTANCEVENDORvCPUsMEMORYAccelerators

      m9g.metal-48xl FAQs