Copyright 2015-2019 Standard Performance Evaluation Corporation

Dell (Test Sponsor: Indiana University)

Tesla V100X-4Q

PowerEdge C4140 Server (KVM virtual machine)

SPECaccel_acc_base = 4.02

SPECaccel_acc_peak = Not Run

ACCEL license: 3440A Test date: Jul-2019
Test sponsor: Indiana University Hardware Availability: May-2019
Tested by: Indiana University Software Availability: Apr-2019
Benchmark results graph
CPU Name: Intel Xeon Gold 6130

CPU Characteristics: Intel Turbo Boost on, SMT off. 2 of 32 cores
allocated to KVM virtual machine.
CPU MHz: 2100
CPU MHz Maximum: 3700
FPU: Integrated
CPU(s) enabled: 32 cores, 2 chips, 16 cores/chip
CPU(s) orderable: 1,2 chips
Primary Cache: 32 KB I + 32 KB D on chip per core
Secondary Cache: 1 MB I+D on chip per core
L3 Cache: 22 MB I+D on chip per chip
Other Cache: None
Memory: 256 GB (16 x 16 GB 2Rx8 PC4-2666V-R)
10 GB (10/256 GB Allocated for KVM)
Disk Subsystem: None
Other Hardware: None
Accel Model Name: Tesla V100
Accel Vendor: NVIDIA Corporation
Accel Name: Tesla V100X-4Q
Type of Accel: GPU
Accel Connection: PCIe
Does Accel Use ECC: Yes
Accel Description: a quarter of virtualized V100-SMX2-16GB with
NVLink (Persistence Mode enabled) allocated to
KVM virtual machine.
Accel Driver: NVIDIA UNIX x86_64 Kernel Module 418.70
Operating System: VM:
CentOS Linux release 7.6.1810 (Core)
Red Hat Enterprise Linux Server release 7.6
(Maipo) 3.10.0-957.21.3.el7.x86_64
Compiler: PGI Community Edition, Release 19.4
File System: cephfs nfsv4 (ganesha) over 100Gbits/s Ethernet
System State: Run level 3 (multi-user)
Other Software: KVM Version 2.12, CUDA 10.1

Results Table

Benchmark Base Peak
Seconds Ratio Seconds Ratio Seconds Ratio Seconds Ratio Seconds Ratio Seconds Ratio
Results appear in the order in which they were run. Bold underlined text indicates a median measurement.
303.ostencil 32.5  4.47  32.5  4.46  32.5  4.47 
304.olbm 125    3.64  125    3.65  125    3.63 
314.omriq 155    6.16  155    6.16  155    6.18 
350.md 41.6  6.05  41.3  6.10  41.6  6.06 
351.palm 143    2.59  140    2.65  141    2.62 
352.ep 193    2.74  193    2.75  193    2.75 
353.clvrleaf 154    2.89  155    2.87  156    2.86 
354.cg 50.0  8.15  48.4  8.43  51.2  7.98 
355.seismic 98.8  3.75  99.1  3.73  98.5  3.76 
356.sp 81.8  3.37  81.9  3.37  83.3  3.31 
357.csp 66.3  4.07  66.0  4.09  66.2  4.08 
359.miniGhost 117    3.15  117    3.14  119    3.10 
360.ilbdc 102    3.61  101    3.62  102    3.59 
363.swim 64.5  3.57  85.3  2.69  87.7  2.62 
370.bt 30.5  7.32  30.5  7.31  30.5  7.31 

Platform Notes

 Sysinfo program /home/lijunj/junjie_benchmarks/spec/accel-1.2-4q/Docs/sysinfo
 $Rev: 6965 $ $Date:: 2015-04-21 #$ c05a7f14b1b1765e3fe1df68447e8a35
 running on v100x-4q.novalocal Thu Jul 11 03:04:21 2019

 This section contains SUT (System Under Test) info as seen by
 some common utilities.  To remove or add to this section, see:

 From /proc/cpuinfo
    model name : Intel(R) Xeon(R) Gold 6130 CPU @ 2.10GHz
       2 "physical id"s (chips)
       2 "processors"
    cores, siblings (Caution: counting these is hw and system dependent.  The
    following excerpts from /proc/cpuinfo might not be reliable.  Use with
       cpu cores : 1
       siblings  : 1
       physical 0: cores 0
       physical 1: cores 0
    cache size : 16384 KB

 From /proc/meminfo
    MemTotal:       11105732 kB
    HugePages_Total:       0
    Hugepagesize:       2048 kB

 From /etc/*release* /etc/*version*
    centos-release: CentOS Linux release 7.6.1810 (Core)
    centos-release-upstream: Derived from Red Hat Enterprise Linux 7.6 (Source)
       NAME="CentOS Linux"
       VERSION="7 (Core)"
       ID_LIKE="rhel fedora"
       PRETTY_NAME="CentOS Linux 7 (Core)"
    redhat-release: CentOS Linux release 7.6.1810 (Core)
    system-release: CentOS Linux release 7.6.1810 (Core)
    system-release-cpe: cpe:/o:centos:centos:7

 uname -a:
    Linux v100x-4q.novalocal 3.10.0-957.12.1.el7.x86_64 #1 SMP Mon Apr 29
    14:59:59 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

 run-level 3 Jul 9 12:18

 SPEC is set to: /home/lijunj/junjie_benchmarks/spec/accel-1.2-4q
    Filesystem                                                        Type  Size
    Used Avail Use% Mounted on nfs4  128G
    54G   75G  42% /home/lijunj
 Additional information from dmidecode:

    Warning: Use caution when you interpret this section. The 'dmidecode' program
    reads system data which is "intended to allow hardware to be accurately
    determined", but the intent may not be met, as there are frequent changes to

Information from pgaccelinfo:
CUDA Driver Version:           10010
NVRM version:                  NVIDIA UNIX x86_64 Kernel Module  418.70
Device Number:                 0
Device Name:                   GRID V100X-4Q
Device Revision Number:        7.0
Global Memory Size:            4294967296
Number of Multiprocessors:     80
Concurrent Copy and Execution: Yes
Total Constant Memory:         65536
Total Shared Memory per Block: 49152
Registers per Block:           65536
Warp Size:                     32
Maximum Threads per Block:     1024
Maximum Block Dimensions:      1024, 1024, 64
Maximum Grid Dimensions:       2147483647 x 65535 x 65535
Maximum Memory Pitch:          2147483647B
Texture Alignment:             512B
Clock Rate:                    1530 MHz
Execution Timeout:             No
Integrated Device:             No
Can Map Host Memory:           Yes
Compute Mode:                  default
Concurrent Kernels:            Yes
ECC Enabled:                   No
Memory Clock Rate:             877 MHz
Memory Bus Width:              4096 bits
L2 Cache Size:                 6291456 bytes
Max Threads Per SMP:           2048
Async Engines:                 2
Unified Addressing:            Yes
Managed Memory:                No
Preemption Supported:          Yes
Cooperative Launch:            Yes
  Multi-Device:                Yes
PGI Default Target:            -ta=tesla:cc70

General Notes

 Four V100-SMX2-16GB GPUs were installed on the host system,
 three of them were idle while only one is active for running SPEC Accel.

 CPUs and GPUs are connected via PCIe, while the four GPUs are connected
 through NVLink.  The NVlink connection is not used in this test.

 Stacksize set to 'unlimited':
      ulimit -s unlimited

 Spectre & Meltdown:
   Yes: The test sponsor attests, as of date of publication, that CVE-2017-5754 (Meltdown)
   is mitigated in the system as tested and documented.
   Yes: The test sponsor attests, as of date of publication, that CVE-2017-5753 (Spectre variant 1)
   is mitigated in the system as tested and documented.
   Yes: The test sponsor attests, as of date of publication, that CVE-2017-5715 (Spectre variant 2)
   is mitigated in the system as tested and documented.

Base Compiler Invocation

C benchmarks:


Fortran benchmarks:


Benchmarks using both Fortran and C:

 pgcc   pgfortran 

Base Optimization Flags

C benchmarks:

 -fast   -Mfprelaxed   -acc   -ta=tesla:cc70   -ta=tesla:cuda10.1 

Fortran benchmarks:

 -fast   -Mfprelaxed   -acc   -ta=tesla:cc70   -ta=tesla:cuda10.1 

Benchmarks using both Fortran and C:

353.clvrleaf:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc70   -ta=tesla:cuda10.1 
359.miniGhost:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc70   -ta=tesla:cuda10.1   -Mnomain 

The flags file that was used to format this result can be browsed at

You can also download the XML flags source by saving the following link: