GPU-16xCPU-192GB-3xL40S by UpCloud

GPU-16xCPU-192GB-3xL40S is a GPU (16 vCPUs, 192 GiB RAM, 3x L40S) server offered by UpCloud with 16 vCPUs, 192 GiB of memory and 0 GB of storage. The pricing starts at 3.1452 USD per hour.
16 vCPU
192 GiB Memory
3 GPU
Spare SCore
66,960
(All-cores)
4,185
(Single-core)

Specifications

Server Metadata

Vendor ID
upcloud
Name
GPU-16xCPU-192GB-3xL40S
Description
GPU (16 vCPUs, 192 GiB RAM, 3x L40S)
Family
GPU
Hw Virt
Average Time To Start
58
Status
active
Observed At
2026-10-23T22:21:39.012016

Availability

REGION / IDSPOTONDEMAND
Helsinki #2 (FI) / fi-hel2
3.1452 USD/h 4.1974 USD/h

Processor

vCPUs
16
Hypervisor
KVM
CPU Allocation
Shared
CPU Cores
16
CPU Speed
2 GHz
CPU Architecture
x86_64
CPU Manufacturer
AMD
CPU Family
EPYC
CPU Model
9575F
CPU L1D Cache
64 KiB
CPU L1D Cache Total
1 MiB
CPU L1I Cache
64 KiB
CPU L1I Cache Total
1 MiB
CPU L2 Cache
512 KiB
CPU L2 Cache Total
8 MiB
CPU L3 Cache
16 MiB
CPU L3 Cache Total
16 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, tsc_known_freq, pni, pclmulqdq, ssse3, fma, cx16, pcid, sse4_1, sse4_2, x2apic, movbe, popcnt, tsc_deadline_timer, aes, xsave, avx, f16c, rdrand, hypervisor, lahf_lm, cmp_legacy, cr8_legacy, abm, sse4a, misalignsse, 3dnowprefetch, osvw, perfctr_core, ssbd, ibrs, ibpb, stibp, vmmcall, fsgsbase, tsc_adjust, bmi1, avx2, smep, bmi2, erms, invpcid, avx512f, avx512dq, rdseed, adx, smap, avx512ifma, clflushopt, clwb, avx512cd, sha_ni, avx512bw, avx512vl, xsaveopt, xsavec, xgetbv1, xsaves, avx_vnni, avx512_bf16, clzero, xsaveerptr, wbnoinvd, arat, avx512vbmi, umip, pku, ospke, avx512_vbmi2, gfni, vaes, vpclmulqdq, avx512_vnni, avx512_bitalg, avx512_vpopcntdq, la57, rdpid, movdiri, movdir64b, fsrm, avx512_vp2intersect, arch_capabilities
Ecpus
16
Scalability
100

System Resources and Accelerators

MEMORY
Memory Amount
192 GiB
Memory Amount Actual
192 GiB
ACCELERATORS
Accelerator Count
3
Accelerator Memory Min
48 GiB
Accelerator Memory Total
144 GiB
Accelerator Type
GPU
Accelerator Manufacturer
NVIDIA
Accelerator Family
Ada Lovelace
Accelerator Model
L40S
Accelerators
  • NVIDIA Ada Lovelace NVIDIA L40S (Memory amount: 46068, Firmware version: 595.84, BIOS version: 95.02.66.00.15, Clock rate: 2520)
  • NVIDIA Ada Lovelace NVIDIA L40S (Memory amount: 46068, Firmware version: 595.84, BIOS version: 95.02.66.00.15, Clock rate: 2520)
  • NVIDIA Ada Lovelace NVIDIA L40S (Memory amount: 46068, Firmware version: 595.84, BIOS version: 95.02.66.00.15, Clock rate: 2520)
STORAGE
Storage Size
0 GB
Storages
    NETWORK
    Inbound Traffic
    0 GB/month
    Outbound Traffic
    32768 GB/month
    IPv4
    1

    CPU and System Topology

    Server Description

    A high-density GPU acceleration platform featuring three Ada Lovelace GPUs and substantial system memory for parallel computing and model inference.

    GPU AcceleratedMemory Optimized

    UpCloud GPU-16xCPU-192GB-3xL40S is a GPU-accelerated virtual server featuring 16 vCPUs, 192 GB of system memory, and three NVIDIA L40S GPUs with 144 GB of total VRAM. Powered by an AMD EPYC 9575F processor on the x86_64 architecture, the instance utilizes a shared CPU allocation model under the KVM hypervisor. It does not include local storage. Performance benchmarks show top-tier single-core CPU execution and large-block memory write bandwidth, alongside strong multi-core capabilities. The server excels in LLM inference, demonstrating top-tier prompt processing and text generation speeds for small and medium models. This hardware profile is qualitatively cost-effective for GPU-heavy workloads, making it highly suitable for machine learning inference, image processing, and data science tasks that require substantial VRAM without requiring dedicated local storage.

    Economics

    Average Price per Region

    Prices per Zone

    Lowest Prices

    Performance

    Workload Profiles

    3.2Score
    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) 6,280,244 ops/sec 1,919,289 ops/sec 3.3 50.00% 50.00% +82.00%
    Redis RPS (pipeline=16, SET) 40,908,720 ops/sec 11,728,058 ops/sec 3.4 20.00% 20.00% +28.00%
    PassMark Memory Mark (composite) 2,573 2,416 1.1 10.00% 10.00% +0.96%
    Memory bandwidth (read, 16 MB ~ L3) 1,246,627 MB/sec 108,466 MB/sec 11 10.00% 10.00% +27.00%
    PassMark single-thread CPU 4,156 Mops/s 2,365 Mops/s 1.8 10.00% 10.00% +6.10%

    GPU Bandwidth and Latency

    Node (all GPUs)
    Node baseline directional bandwidth
    		67.9 GB/s
    Node GPU-to-host bandwidth
    		79.1 GB/s
    Node host-to-GPU bandwidth
    		80.6 GB/s
    Single GPU slot
    Per-GPU GPU-to-host bandwidth
    		26.4 GB/s
    Per-GPU host-to-GPU bandwidth
    		26.9 GB/s
    Per-GPU host-device latency
    		640 ns
    Benchmark Environment
    Driver Version
    		 595.84
    CUDA Runtime Version
    		 13.2
    nvbandwidth Version
    		 v0.8
    GPU Count
    		 3
    P2P Supported
    No
    Efficiency
    SM/CE copy efficiency
    		1.00

    Memory Bandwidth

    Compression

    OpenSSL

    Geekbench Single-Core

    Score: 2,922

    Geekbench Multi-Core

    Score: 20,598

    Passmark CPU Scores

    BENCHMARKSCORE
    Mark
    		 52,195
    Compression
    		 586,406
    Encryption
    		 28,264
    Extended Instructions
    		 53,392
    Floating Point Maths
    		 135,244
    Integer Maths
    		 135,068
    Physics
    		 8,691
    Prime Numbers
    		 506
    Single Threaded
    		 4,156
    String Sorting
    		 80,633

    Passmark Memory Scores

    BENCHMARKSCORE
    Memory Mark
    		 2,573
    Database Operations
    		 20,078
    Memory Latency
    		 97
    Memory Read Cached
    		 37,665
    Memory Read Uncached
    		 28,576
    Memory Write
    		 30,421

    Stress-ng Stressors

    CPU integerSingleMultiScaling
    atomic
    		 588.51 446.53 0.76×
    bitops
    		 254,505 4,077,461 16.02×
    branch
    		 205,733 3,178,055 15.45×
    funccall
    		 1,003 15,974 15.92×
    funcret
    		 29,799 490,228 16.45×
    hash
    		 241,651 3,836,580 15.88×
    intmath
    		 88,380 1,407,133 15.92×
    nop
    		 479,710 7,616,553 15.88×
    prime
    		 685,674 10,859,299 15.84×
    rotate
    		 158,635 2,516,485 15.86×
    str
    		 1,687,804 27,351,612 16.21×
    wcs
    		 1,162,906 18,204,516 15.65×
    CryptographySingleMultiScaling
    crypt
    		 6,087 93,805 15.41×
    CompressionSingleMultiScaling
    jpeg
    		 1,348 21,017 15.59×
    zlib
    		 76.66 1,278 16.68×
    HypervisorSingleMultiScaling
    clock
    		 85,559 180,032 2.10×
    clone
    		 267.62 3,542 13.23×
    context
    		 9,566 151,812 15.87×
    cpu-sched
    		 83.55 1,002 11.99×
    eventfd
    		 79,089 625,112 7.90×
    fault
    		 44,640 120,291 2.69×
    fork
    		 4,937 44,569 9.03×
    getrandom
    		 108,338 1,744,793 16.11×
    hrtimers
    		 228,702 1,436,096 6.28×
    mprotect
    		 477,890 7,056,067 14.77×
    mtx
    		 42,418,630 1,538,770,168 36.28×
    munmap
    		 3,830 36,477 9.52×
    null
    		 1,586,341 25,217,634 15.90×
    pipe
    		 638,210 9,459,745 14.82×
    priv-instr
    		 1,528,316 5,961,435 3.90×
    resched
    		 7,796,620 8,299,465 1.06×
    switch
    		 102,511 8,126,953 79.28×
    timermix
    		 368,483 5,741,838 15.58×
    tlb-shootdown
    		 16,578 134,360 8.10×
    tsc
    		 987,075 15,748,734 15.95×
    usersyscall
    		 1,117,661 6,055,649 5.42×
    vdso
    		 109,008,273 1,735,601,265 15.92×
    vfork
    		 31,826 132,331 4.16×
    x86syscall
    		 21,121,290 330,687,114 15.66×
    yield
    		 25,665,160 24,917,132 0.97×
    zero
    		 12,323,608 193,144,195 15.67×
    CPU float/SIMDSingleMultiScaling
    eigen
    		 19,290 310,770 16.11×
    expmath
    		 4,445 71,030 15.98×
    fma
    		 1,748,230 27,758,386 15.88×
    fp
    		 220.67 3,580 16.22×
    fractal
    		 9.8 155.6 15.88×
    hyperbolic
    		 6,080 96,620 15.89×
    logmath
    		 6,368 100,906 15.84×
    matrix
    		 10,375 163,829 15.79×
    matrix-3d
    		 1,625 23,176 14.27×
    monte-carlo
    		 507.01 8,064 15.91×
    mpfr
    		 2,694 42,671 15.84×
    powmath
    		 2,035 32,691 16.07×
    trig
    		 2,704 43,179 15.97×
    veccmp
    		 33,743 532,823 15.79×
    vecfp
    		 2,643 42,966 16.26×
    vecmath
    		 5,285 84,138 15.92×
    vecshuf
    		 1,115 17,765 15.94×
    vecwide
    		 44,965 712,494 15.85×
    AlgorithmsSingleMultiScaling
    bitonicsort
    		 14.4 223.59 15.53×
    bsearch
    		 400.39 6,470 16.16×
    bubblesort
    		 1.4 23.2 16.57×
    fibsearch
    		 473.55 7,566 15.98×
    heapsort
    		 15.8 253.59 16.05×
    hsearch
    		 14,525 235,748 16.23×
    insertionsort
    		 18 279.8 15.54×
    malloc
    		 5,497,666 26,531,688 4.83×
    mcontend
    		 83.82 2,361 28.17×
    mergesort
    		 44.2 694.18 15.71×
    qsort
    		 29.2 469.58 16.08×
    skiplist
    		 9,274 149,150 16.08×
    tree
    		 2.8 44.76 15.99×
    MemorySingleMultiScaling
    cache
    		 1,320,003 1,095,399 0.83×
    cachehammer
    		 153,256 2,982,457 19.46×
    cacheline
    		 45.36 730.83 16.11×
    l1cache
    		 750.41 12,097 16.12×
    memcpy
    		 906.23 14,322 15.80×
    memrate
    		 0.96 11.69 12.18×
    memthrash
    		 155.17 935.01 6.03×
    ptr-chase
    		 63,509,774 854,095,634 13.45×
    spinmem
    		 451.37 1,447 3.21×
    stream
    		 405.1 50,459 124.56×
    vm
    		 245,091 293,248 1.20×

    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
    GPU-8xCPU-64GB-1xL40S 864 GiB1
    GPU-12xCPU-128GB-1xL40S 12128 GiB1
    GPU-12xCPU-128GB-2xL40S 12128 GiB2
    GPU-16xCPU-80GB-1xRTXPRO6000 1680 GiB1
    GPU-16xCPU-192GB-1xL40S 16192 GiB1
    GPU-16xCPU-192GB-2xL40S 16192 GiB2
    GPU-20xCPU-256GB-1xL40S 20256 GiB1

    Similar Servers

    INSTANCEVENDORvCPUsMEMORYAccelerators

    GPU-16xCPU-192GB-3xL40S FAQs