GPU-12xCPU-128GB-2xL40S by UpCloud

GPU-12xCPU-128GB-2xL40S is a GPU (12 vCPUs, 128 GiB RAM, 2x L40S) server offered by UpCloud with 12 vCPUs, 128 GiB of memory and 0 GB of storage. The pricing starts at 1.7461 USD per hour.
12 vCPU
128 GiB Memory
2 GPU
Spare SCore
50,416
(All-cores)
4,216
(Single-core)

Specifications

Server Metadata

Vendor ID
upcloud
Name
GPU-12xCPU-128GB-2xL40S
Description
GPU (12 vCPUs, 128 GiB RAM, 2x L40S)
Family
GPU
Hw Virt
Average Time To Start
84.5
Status
active
Observed At
2026-10-24T09:12:07.389089

Availability

REGION / IDSPOTONDEMAND
Helsinki #2 (FI) / fi-hel2
1.7461 USD/h 2.3318 USD/h

Processor

vCPUs
12
Hypervisor
KVM
CPU Allocation
Shared
CPU Cores
12
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
768 KiB
CPU L1I Cache
64 KiB
CPU L1I Cache Total
768 KiB
CPU L2 Cache
512 KiB
CPU L2 Cache Total
6 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, 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, avx512_vp2intersect, arch_capabilities
Ecpus
12
Scalability
100

System Resources and Accelerators

MEMORY
Memory Amount
128 GiB
Memory Amount Actual
128 GiB
ACCELERATORS
Accelerator Count
2
Accelerator Memory Min
48 GiB
Accelerator Memory Total
96 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: 595.71.05, BIOS version: 95.02.66.00.15, Clock rate: 2520)
  • NVIDIA Ada Lovelace NVIDIA L40S (Memory amount: 46068, Firmware version: 595.71.05, BIOS version: 95.02.66.00.15, Clock rate: 2520)
STORAGE
Storage Size
0 GB
Storages
    NETWORK
    Inbound Traffic
    0 GB/month
    Outbound Traffic
    24576 GB/month
    IPv4
    1

    CPU and System Topology

    Server Description

    A GPU-accelerated virtual server featuring dual NVIDIA L40S GPUs and 96 GB VRAM, optimized for machine learning and LLM inference.

    GPU AcceleratedCompute Optimized

    UpCloud GPU-12xCPU-128GB-2xL40S is a GPU-accelerated server featuring 12 vCPUs, 128 GB of RAM, and two NVIDIA L40S GPUs with 96 GB of total VRAM. Powered by an AMD EPYC 9575F processor running at 2.0 GHz on the x86_64 architecture, this KVM-based instance offers top-tier single-core CPU performance and strong memory bandwidth. While the CPU allocation is shared and local storage is absent, the server delivers exceptional GPU capabilities, as evidenced by top-tier text generation speeds in small LLM inference benchmarks. This configuration is highly efficient for workloads that prioritize GPU acceleration over local disk storage, making it well-suited for machine learning, LLM inference, and data science applications.

    Economics

    Average Price per Region

    Prices per Zone

    Lowest Prices

    Performance

    Workload Profiles

    2.6Score
    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) 4,797,883 ops/sec 1,919,289 ops/sec 2.5 50.00% 50.00% +58.00%
    Redis RPS (pipeline=16, SET) 30,846,454 ops/sec 11,728,058 ops/sec 2.6 20.00% 20.00% +21.00%
    PassMark Memory Mark (composite) 2,888 2,416 1.2 10.00% 10.00% +1.80%
    Memory bandwidth (read, 16 MB ~ L3) 930,580 MB/sec 108,466 MB/sec 8.5 10.00% 10.00% +24.00%
    PassMark single-thread CPU 4,212 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
    		45.3 GB/s
    Node GPU-to-host bandwidth
    		52.7 GB/s
    Node host-to-GPU bandwidth
    		53.7 GB/s
    Single GPU slot
    Per-GPU GPU-to-host bandwidth
    		26.4 GB/s
    Per-GPU host-to-GPU bandwidth
    		26.8 GB/s
    Per-GPU host-device latency
    		588 ns
    Benchmark Environment
    Driver Version
    		 595.71.05
    CUDA Runtime Version
    		 13.2
    nvbandwidth Version
    		 v0.8
    GPU Count
    		 2
    P2P Supported
    No
    Efficiency
    SM/CE copy efficiency
    		1.00

    Memory Bandwidth

    Compression

    OpenSSL

    Geekbench Single-Core

    Score: 2,712

    Geekbench Multi-Core

    Score: 17,329

    Passmark CPU Scores

    BENCHMARKSCORE
    Mark
    		 41,243
    Compression
    		 444,625
    Encryption
    		 21,305
    Extended Instructions
    		 39,856
    Floating Point Maths
    		 102,410
    Integer Maths
    		 102,278
    Physics
    		 6,665
    Prime Numbers
    		 402
    Single Threaded
    		 4,212
    String Sorting
    		 60,216

    Passmark Memory Scores

    BENCHMARKSCORE
    Memory Mark
    		 2,888
    Database Operations
    		 17,639
    Memory Latency
    		 85
    Memory Read Cached
    		 38,192
    Memory Read Uncached
    		 35,383
    Memory Write
    		 35,374

    Stress-ng Stressors

    CPU integerSingleMultiScaling
    atomic
    		 401.29 448.76 1.12×
    bitops
    		 234,414 2,841,779 12.12×
    branch
    		 186,012 2,144,498 11.53×
    funccall
    		 922.66 10,860 11.77×
    funcret
    		 28,058 340,523 12.14×
    hash
    		 218,686 2,597,773 11.88×
    intmath
    		 81,535 978,168 12.00×
    nop
    		 408,245 5,084,861 12.46×
    prime
    		 623,371 7,088,633 11.37×
    rotate
    		 144,551 1,708,182 11.82×
    str
    		 1,686,680 18,116,309 10.74×
    wcs
    		 1,061,646 11,707,416 11.03×
    CryptographySingleMultiScaling
    crypt
    		 5,509 63,552 11.54×
    CompressionSingleMultiScaling
    jpeg
    		 1,302 14,801 11.37×
    zlib
    		 82.4 906.82 11.01×
    HypervisorSingleMultiScaling
    clock
    		 85,525 292,422 3.42×
    clone
    		 613.24 4,376 7.14×
    context
    		 8,982 100,468 11.19×
    cpu-sched
    		 58.24 754.5 12.96×
    eventfd
    		 86,254 2,308,315 26.76×
    fault
    		 39,091 164,371 4.20×
    fork
    		 6,764 59,152 8.74×
    getrandom
    		 105,600 1,200,719 11.37×
    hrtimers
    		 819,640 3,180,870 3.88×
    mprotect
    		 712,062 6,223,708 8.74×
    mtx
    		 47,624,283 1,064,413,516 22.35×
    munmap
    		 5,577 49,370 8.85×
    null
    		 1,485,190 17,204,651 11.58×
    pipe
    		 735,696 11,263,491 15.31×
    priv-instr
    		 1,434,836 10,156,912 7.08×
    resched
    		 10,681,189 6,179,733 0.58×
    switch
    		 104,526 5,235,182 50.09×
    timermix
    		 351,531 3,671,561 10.44×
    tlb-shootdown
    		 24,048 108,908 4.53×
    tsc
    		 917,222 11,014,235 12.01×
    usersyscall
    		 1,093,822 10,079,167 9.21×
    vdso
    		 101,051,404 1,206,400,739 11.94×
    vfork
    		 44,415 190,515 4.29×
    x86syscall
    		 19,528,324 229,565,789 11.76×
    yield
    		 18,857,041 21,493,292 1.14×
    zero
    		 8,188,828 93,099,090 11.37×
    CPU float/SIMDSingleMultiScaling
    eigen
    		 18,018 212,628 11.80×
    expmath
    		 4,062 48,925 12.05×
    fma
    		 1,546,599 18,711,591 12.10×
    fp
    		 231.45 2,580 11.15×
    fractal
    		 9.4 109.4 11.64×
    hyperbolic
    		 5,546 65,582 11.83×
    logmath
    		 5,895 68,905 11.69×
    matrix
    		 8,868 107,465 12.12×
    matrix-3d
    		 1,458 11,251 7.72×
    monte-carlo
    		 474.89 5,645 11.89×
    mpfr
    		 2,586 29,223 11.30×
    powmath
    		 1,918 22,409 11.68×
    trig
    		 2,503 29,300 11.71×
    veccmp
    		 30,487 353,631 11.60×
    vecfp
    		 2,500 29,810 11.93×
    vecmath
    		 4,964 58,106 11.70×
    vecshuf
    		 1,040 11,911 11.46×
    vecwide
    		 40,822 462,460 11.33×
    AlgorithmsSingleMultiScaling
    bitonicsort
    		 13.4 147.2 10.99×
    bsearch
    		 389.03 4,383 11.27×
    bubblesort
    		 1.4 16.6 11.86×
    fibsearch
    		 442.76 5,080 11.47×
    heapsort
    		 14 167.99 12.00×
    hsearch
    		 13,427 152,475 11.36×
    insertionsort
    		 16.2 188.2 11.62×
    malloc
    		 5,078,875 34,299,616 6.75×
    mcontend
    		 138.02 1,865 13.51×
    mergesort
    		 40.6 480.99 11.85×
    qsort
    		 28.2 316.59 11.23×
    skiplist
    		 8,633 99,113 11.48×
    tree
    		 2.4 25.39 10.58×
    MemorySingleMultiScaling
    cache
    		 1,240,643 1,317,830 1.06×
    cachehammer
    		 267,956 3,056,123 11.41×
    cacheline
    		 60.24 752.5 12.49×
    l1cache
    		 714.38 8,251 11.55×
    memcpy
    		 828.11 9,338 11.28×
    memrate
    		 1.14 10.43 9.15×
    memthrash
    		 168.88 778.48 4.61×
    ptr-chase
    		 52,094,468 149,596,546 2.87×
    spinmem
    		 430.98 788.53 1.83×
    stream
    		 442.4 21,492 48.58×
    vm
    		 245,381 293,280 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-16xCPU-80GB-1xRTXPRO6000 1680 GiB1
    GPU-16xCPU-192GB-1xL40S 16192 GiB1
    GPU-16xCPU-192GB-2xL40S 16192 GiB2
    GPU-16xCPU-192GB-3xL40S 16192 GiB3
    GPU-20xCPU-256GB-1xL40S 20256 GiB1

    Similar Servers

    INSTANCEVENDORvCPUsMEMORYAccelerators

    GPU-12xCPU-128GB-2xL40S FAQs