c9g.24xlarge by Amazon Web Services

c9g.24xlarge is a Compute optimized [AWS Graviton processors] Gen9 24xlarge server offered by Amazon Web Services with 96 vCPUs, 192 GiB of memory and 0 GB of storage. The pricing starts at 0.6311 USD per hour.
96 vCPU
192 GiB Memory
Spare SCore
192,014
(All-cores)
2,007
(Single-core)

Specifications

Server Metadata

Vendor ID
aws
Name
c9g.24xlarge
Description
Compute optimized [AWS Graviton processors] Gen9 24xlarge
Family
Compute Optimized
Series
c9g
Hw Virt
Average Time To Start
8
Status
active
Observed At
2026-11-14T19:09:28.102021

Availability

Region / IDSpotOndemand
(MY) / ap-southeast-5
1.5529 USD/h
4.0742 USD/h
Ohio (US) / us-east-2
0.834533 USD/h
4.1587 USD/h
Northern Virgina (US) / us-east-1
0.963625 USD/h
4.1731 USD/h
Oregon (US) / us-west-2
1.406325 USD/h
4.1731 USD/h
Seoul (KR) / ap-northeast-2
1.020233 USD/h
4.6963 USD/h
Frankfurt (DE) / eu-central-1
1.3275 USD/h
4.7453 USD/h
Singapore (SG) / ap-southeast-1
1.2155 USD/h
4.7942 USD/h
Tokyo (JP) / ap-northeast-1
1.236733 USD/h
5.2349 USD/h

Processor

vCPUs
96
Hypervisor
nitro
CPU Allocation
Dedicated
CPU Cores
96
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
6 MiB
CPU L1I Cache
64 KiB
CPU L1I Cache Total
6 MiB
CPU L2 Cache
2 MiB
CPU L2 Cache Total
192 MiB
CPU L3 Cache
96 MiB
CPU L3 Cache Total
96 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
95.7
Scalability
99.69

System Resources and Accelerators

Memory
Memory Amount
192 GiB
Memory Amount Actual
189 GiB
Accelerators
Accelerator Count
0
Accelerator Memory Min
0 MiB
Accelerator Memory Total
0 MiB
Accelerators
    Storage
    Storage Size
    0 GB
    Storages
      Compatible Storage IDs
      standard, gp2, gp3, st1, sc1, io1, io2
      Network
      Network Speed Baseline
      50 Gbps
      Network Speed Max
      50 Gbps
      Network Storage Speed Baseline
      36 Gbps
      Network Storage Speed Max
      36 Gbps
      Inbound Traffic
      0 GB/month
      Outbound Traffic
      0 GB/month
      IPv4
      0

      CPU and System Topology

      Server Description

      An arm64 compute-optimized instance delivering 96 dedicated physical cores and 50 Gbps networking for high-concurrency workloads and multi-threaded data processing.

      Amazon Web Services c9g.24xlarge is a compute-optimized server equipped with 96 dedicated ARMv9 Neoverse-V3 cores running at 3.3 GHz, 192.0 GB of RAM, and 50 Gbps baseline network bandwidth. Built on the AWS Nitro hypervisor, it provides 2.0 GB of memory per core along with 96 MB of shared L3 cache and hardware AES acceleration, featuring no local storage or GPUs. Benchmark evaluations show strong multi-threaded scaling, including a stress-ng multi-core score of 192014.4, a PassMark database score of 57730.61, and Redis throughput of 269.45 million SET operations per second. The server provides high physical compute density and network throughput, optimizing resource utilization for parallel execution. The instance is suited for static web serving, key-value caching, multi-threaded compression, and CPU-driven LLM prompt processing.

      Economics

      Average Price per Region

      Prices per Zone

      Lowest Prices

      Performance

      Workload Profiles

      8.8Score
      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) 32,844,236 ops/sec 1,933,252 ops/sec 17 50.00% 50.00% +310.00%
      Redis RPS (pipeline=16, SET) 192,446,762 ops/sec 11,814,424 ops/sec 16 20.00% 20.00% +74.00%
      PassMark Memory Mark (composite) 4,604 2,408 1.9 10.00% 10.00% +6.60%
      Memory bandwidth (read, 16 MB ~ L3) 425,661 MB/sec 108,674 MB/sec 3.9 10.00% 10.00% +15.00%
      PassMark single-thread CPU 2,464 Mops/s 2,365 Mops/s 1 10.00% 10.00% 0.00%

      Memory Bandwidth

      Compression

      OpenSSL

      Passmark CPU Scores

      BENCHMARKSCORE
      Mark
      		 77,220
      Compression
      		 890,417
      Encryption
      		 28,519
      Extended Instructions
      		 129,079
      Floating Point Maths
      		 732,775
      Integer Maths
      		 586,420
      Physics
      		 21,539
      Prime Numbers
      		 1,127
      Single Threaded
      		 2,464
      String Sorting
      		 489,031

      Passmark Memory Scores

      BENCHMARKSCORE
      Memory Mark
      		 4,604
      Database Operations
      		 57,731
      Memory Latency
      		 28
      Memory Read Cached
      		 24,850
      Memory Read Uncached
      		 24,526
      Memory Write
      		 24,951

      Stress-ng Stressors

      CPU integerSingleMultiScaling
      atomic
      		 181.13 9.15 0.05×
      bitops
      		 165,033 15,815,464 95.83×
      branch
      		 319,905 30,629,666 95.75×
      funccall
      		 2,281 218,196 95.65×
      funcret
      		 41,061 3,753,236 91.41×
      hash
      		 181,349 17,339,631 95.61×
      intmath
      		 39,471 3,785,088 95.89×
      nop
      		 429,700 41,117,317 95.69×
      prime
      		 600,481 57,460,086 95.69×
      rotate
      		 149,067 14,245,766 95.57×
      str
      		 936,637 89,344,342 95.39×
      wcs
      		 280,876 26,712,024 95.10×
      CryptographySingleMultiScaling
      crypt
      		 4,280 417,155 97.47×
      CompressionSingleMultiScaling
      jpeg
      		 914.78 55,927 61.14×
      zlib
      		 45.02 4,445 98.73×
      HypervisorSingleMultiScaling
      clock
      		 46,386 262,672 5.66×
      clone
      		 311.24 17,442 56.04×
      context
      		 4,258 380,880 89.46×
      cpu-sched
      		 69 6,185 89.63×
      eventfd
      		 143,016 4,645,635 32.48×
      fault
      		 28,164 374,428 13.29×
      fork
      		 6,025 70,372 11.68×
      getrandom
      		 73,976 7,081,317 95.72×
      hrtimers
      		 485,121 7,054,899 14.54×
      mprotect
      		 1,223,223 19,717,978 16.12×
      mtx
      		 12,335,551 5,753,043,364 466.38×
      munmap
      		 4,400 112,354 25.54×
      null
      		 683,421 65,700,335 96.13×
      pipe
      		 942,954 50,417,163 53.47×
      priv-instr
      		 1,120,822 10,857,657 9.69×
      resched
      		 1,703,633 16,738,324 9.83×
      switch
      		 251,831 64,640,362 256.68×
      timermix
      		 588,527 37,454,504 63.64×
      tlb-shootdown
      		 15,452 271,599 17.58×
      vdso
      		 59,588,050 6,101,437,556 102.39×
      vfork
      		 48,799 366,789 7.52×
      yield
      		 158,483,547 170,005,801 1.07×
      zero
      		 3,281,772 314,433,322 95.81×
      CPU float/SIMDSingleMultiScaling
      eigen
      		 2,245 214,647 95.63×
      expmath
      		 424.96 40,675 95.72×
      fma
      		 851,809 81,549,477 95.74×
      fp
      		 175.2 16,852 96.19×
      fractal
      		 9.2 867.17 94.26×
      hyperbolic
      		 408.13 39,058 95.70×
      logmath
      		 491.16 46,994 95.68×
      matrix
      		 7,978 762,382 95.56×
      matrix-3d
      		 1,477 17,708 11.99×
      monte-carlo
      		 269.59 25,988 96.40×
      mpfr
      		 1,887 180,638 95.74×
      powmath
      		 281.91 26,968 95.66×
      trig
      		 553.35 52,952 95.69×
      veccmp
      		 26,340 2,519,522 95.65×
      vecfp
      		 1,860 178,038 95.70×
      vecmath
      		 4,813 460,285 95.63×
      vecshuf
      		 1,029 98,717 95.90×
      vecwide
      		 30,105 3,097,843 102.90×
      AlgorithmsSingleMultiScaling
      bitonicsort
      		 11.2 1,075 95.98×
      bsearch
      		 422.43 40,436 95.72×
      bubblesort
      		 1.4 134.4 96.00×
      fibsearch
      		 414.3 39,608 95.60×
      heapsort
      		 15.4 1,459 94.76×
      hsearch
      		 10,849 1,028,867 94.84×
      insertionsort
      		 7.4 753.8 101.86×
      malloc
      		 4,204,865 194,984,206 46.37×
      mcontend
      		 215.89 241,812 1120.07×
      mergesort
      		 38.6 3,657 94.74×
      qsort
      		 26.8 2,547 95.03×
      skiplist
      		 7,658 732,722 95.68×
      tree
      		 1.8 95.64 53.13×
      MemorySingleMultiScaling
      cache
      		 702,599 2,569,248 3.66×
      cachehammer
      		 2,292,340 34,609,144 15.10×
      cacheline
      		 55.97 5,442 97.23×
      l1cache
      		 389.92 35,897 92.06×
      memcpy
      		 1,016 91,686 90.20×
      memrate
      		 6.72 53.45 7.95×
      memthrash
      		 289.93 537.1 1.85×
      ptr-chase
      		 28,743,070 877,849,481 30.54×
      spinmem
      		 6,542 95,539 14.60×
      stream
      		 54.37 129,194 2376.20×
      vm
      		 198,039 3,126,178 15.79×

      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
      c6g.medium 12 GiB0
      c6gd.medium 12 GiB0
      c6gn.medium 12 GiB0
      c7a.medium 12 GiB0
      c7g.medium 12 GiB0
      c7gd.medium 12 GiB0
      c7gn.medium 12 GiB0

      Similar Servers

      INSTANCEVENDORvCPUsMEMORYAccelerators

      c9g.24xlarge FAQs