SPEC® ACCEL™ OCL Result

Copyright 2015-2021 Standard Performance Evaluation Corporation

Lenovo Global Technology

NVIDIA Tesla A100-PCIE-40GB

ThinkSystem SR655

SPECaccel_ocl_base = 15.8 

SPECaccel_ocl_peak = 18.2 

ACCEL license: 28 Test date: May-2021
Test sponsor: Lenovo Global Technology Hardware Availability: Jun-2021
Tested by: Lenovo Global Technology Software Availability: Jun-2021
Benchmark results graph
Hardware
CPU Name: AMD EPYC 7763
CPU Characteristics: Turbo up to 3.5 GHz
CPU MHz: 2450
CPU MHz Maximum: 3500
FPU: Integrated
CPU(s) enabled: 64 cores, 1 chip, 64 cores/chip
CPU(s) orderable: 1 chip
Primary Cache: 32 KB I + 32 KB D on chip per core
Secondary Cache: 512 KB I+D on chip per core
L3 Cache: 256 MB I+D on chip per chip
32 MB shared / 8 cores
Other Cache: None
Memory: 256 GB (8 x 32 GB 2Rx8 PC4-3200AA-R)
Disk Subsystem: 1 x 480 GB 2.5" SSD
Other Hardware: None
Accelerator
Accel Model Name: NVIDIA Tesla A100-PCIE-40GB
Accel Vendor: NVIDIA Corporation
Accel Name: NVIDIA Tesla A100-PCIE-40GB
Type of Accel: GPU
Accel Connection: PCIe 4.0 16x
Does Accel Use ECC: Yes
Accel Description: NVIDIA Tesla A100-PCIE-40GB
Accel Driver: NVIDIA UNIX x86_64 Kernel Module 450.51.05
Software
Operating System: Red Hat Enterprise Linux release 8.3 (Ootpa)
4.18.0-240.el8.x86_64
Compiler: Nvidia HPC SDK Release 21.3
File System: xfs
System State: Run level 3
Other Software: CUDA 11.0 SDK

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.
101.tpacf 6.02  17.8   6.00  17.8   5.99  17.9   3.70   28.9   3.68   29.1   3.70   28.9  
103.stencil 4.84  25.8   4.82  25.9   4.82  25.9   4.84   25.8   4.82   25.9   4.82   25.9  
104.lbm 4.25  26.3   4.25  26.3   4.29  26.1   4.24   26.4   4.27   26.2   4.27   26.2  
110.fft 4.46  24.9   4.47  24.8   4.48  24.8   4.46   24.9   4.47   24.8   4.48   24.8  
112.spmv 12.5   11.8   12.5   11.7   12.5   11.8   12.5    11.8   12.5    11.8   12.5    11.8  
114.mriq 2.73  39.9   2.68  40.6   2.74  39.7   2.73   39.9   2.68   40.6   2.74   39.7  
116.histo 32.5   3.50  33.9   3.36  33.2   3.43  32.5    3.50  33.9    3.36  33.2    3.43 
117.bfs 5.36  21.8   5.33  22.0   5.35  21.9   5.85   20.0   5.87   19.9   5.83   20.1  
118.cutcp 3.36  29.5   3.39  29.2   3.34  29.6   3.36   29.5   3.39   29.2   3.34   29.6  
120.kmeans 31.5   3.17  31.9   3.14  31.6   3.17  31.6    3.16  31.7    3.15  31.6    3.17 
121.lavamd 4.44  24.5   4.53  24.0   4.58  23.8   4.44   24.5   4.53   24.0   4.58   23.8  
122.cfd 8.17  15.4   8.21  15.4   8.19  15.4   8.10   15.6   8.06   15.6   8.10   15.6  
123.nw 13.8   8.31  14.0   8.19  13.9   8.27  13.8    8.31  14.0    8.19  13.9    8.27 
124.hotspot 5.75  19.8   5.71  20.0   5.80  19.6   5.75   19.8   5.71   20.0   5.80   19.6  
125.lud 9.02  13.2   9.01  13.2   9.02  13.2   5.96   20.0   5.92   20.1   5.98   19.9  
126.ge 6.27  24.7   6.27  24.7   6.27  24.7   0.945  164     0.948  163     0.934  166    
127.srad 8.01  14.2   8.01  14.2   8.01  14.2   8.01   14.2   8.01   14.2   8.01   14.2  
128.heartwall 8.75  12.1   8.68  12.2   8.67  12.2   8.75   12.1   8.68   12.2   8.67   12.2  
140.bplustree 6.09  17.7   6.07  17.8   6.09  17.7   6.09   17.7   6.07   17.8   6.09   17.7  

Submit Notes

The config file option 'submit' was used.

Platform Notes

 Sysinfo program /home/ACCEL1.3/Docs/sysinfo
 $Rev: 6965 $ $Date:: 2015-04-21 #$ c05a7f14b1b1765e3fe1df68447e8a35
 running on amd2srh833 Tue May 11 20:26:42 2021

 This section contains SUT (System Under Test) info as seen by
 some common utilities.  To remove or add to this section, see:
   http://www.spec.org/accel/Docs/config.html#sysinfo

 From /proc/cpuinfo
    model name : AMD EPYC 7763 64-Core Processor
       1 "physical id"s (chips)
       64 "processors"
    cores, siblings (Caution: counting these is hw and system dependent.  The
    following excerpts from /proc/cpuinfo might not be reliable.  Use with
    caution.)
       cpu cores : 64
       siblings  : 64
       physical 0: cores 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
       22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
       47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63
    cache size : 512 KB

 From /proc/meminfo
    MemTotal:       263708564 kB
    HugePages_Total:       0
    Hugepagesize:       2048 kB

 From /etc/*release* /etc/*version*
    os-release:
       NAME="Red Hat Enterprise Linux"
       VERSION="8.3 (Ootpa)"
       ID="rhel"
       ID_LIKE="fedora"
       VERSION_ID="8.3"
       PLATFORM_ID="platform:el8"
       PRETTY_NAME="Red Hat Enterprise Linux 8.3 (Ootpa)"
       ANSI_COLOR="0;31"
    redhat-release: Red Hat Enterprise Linux release 8.3 (Ootpa)
    system-release: Red Hat Enterprise Linux release 8.3 (Ootpa)
    system-release-cpe: cpe:/o:redhat:enterprise_linux:8.3:ga

 uname -a:
    Linux amd2srh833 4.18.0-240.el8.x86_64 #1 SMP Wed Sep 23 05:13:10 EDT 2020
    x86_64 x86_64 x86_64 GNU/Linux

 run-level 3 Jan 13 11:28

 SPEC is set to: /home/ACCEL1.3
    Filesystem     Type  Size  Used Avail Use% Mounted on
    /dev/sda3      xfs   419G   76G  343G  19% /home
 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
    hardware, firmware, and the "DMTF SMBIOS" standard.

   BIOS Lenovo             CFE125L 03/26/2021
   Memory:
    8x Samsung M393A4G43AB3-CWE 32 GB 2 rank 3200 MT/s
    8x Unknown Unknown

 (End of data from sysinfo program)

General Notes

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 Runtime Environment

C benchmarks:

 OpenCL Platform: NVIDIA CUDA, OpenCL 1.2 CUDA 11.0.197   OpenCL Device #0: A100-PCIE-40GB, v 450.51.05 

C++ benchmarks:

 OpenCL Platform: NVIDIA CUDA, OpenCL 1.2 CUDA 11.0.197   OpenCL Device #0: A100-PCIE-40GB, v 450.51.05 

Base Compiler Invocation

C benchmarks:

 nvc 

C++ benchmarks:

 nvc++ 

Base Portability Flags

116.histo:  -DSPEC_LOCAL_MEMORY_HEADROOM=1 

Base Optimization Flags

C benchmarks:

 -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed 

C++ benchmarks:

 -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed 

Base Other Flags

C benchmarks:

 -I/usr/local/cuda-11.0/include   -L/usr/local/cuda-11.0/lib64   -lOpenCL 

C++ benchmarks:

 -I/usr/local/cuda-11.0/include   -L/usr/local/cuda-11.0/lib64   -lOpenCL 

Peak Runtime Environment

C benchmarks:

 OpenCL Platform: NVIDIA CUDA, OpenCL 1.2 CUDA 11.0.197   OpenCL Device #0: A100-PCIE-40GB, v 450.51.05 

C++ benchmarks:

 OpenCL Platform: NVIDIA CUDA, OpenCL 1.2 CUDA 11.0.197   OpenCL Device #0: A100-PCIE-40GB, v 450.51.05 

Peak Compiler Invocation

C benchmarks:

 nvc 

C++ benchmarks:

 nvc++ 

Peak Portability Flags

116.histo:  -DSPEC_LOCAL_MEMORY_HEADROOM=1 

Peak Optimization Flags

C benchmarks:

110.fft:  basepeak = yes 
114.mriq:  basepeak = yes 
116.histo:  basepeak = yes 
117.bfs:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=64   -DSPEC_ACCEL_WG_SIZE_1_0=64 
118.cutcp:  basepeak = yes 
121.lavamd:  basepeak = yes 
124.hotspot:  basepeak = yes 
127.srad:  basepeak = yes 
128.heartwall:  basepeak = yes 
140.bplustree:  basepeak = yes 

C++ benchmarks:

101.tpacf:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=1024 
103.stencil:  basepeak = yes 
104.lbm:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=32   -DSPEC_ACCEL_WG_SIZE_0_1=1   -DSPEC_ACCEL_WG_SIZE_0_2=1 
112.spmv:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=96 
120.kmeans:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=288 
122.cfd:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_3_0=288 
123.nw:  basepeak = yes 
125.lud:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=32 
126.ge:  -fast   -Mstack_arrays   -Mnouniform   -Mfprelaxed   -DSPEC_ACCEL_WG_SIZE_0_0=512   -DSPEC_ACCEL_WG_SIZE_1_0=1   -DSPEC_ACCEL_WG_SIZE_1_1=512 

Peak Other Flags

C benchmarks:

 -I/usr/local/cuda-11.0/include   -L/usr/local/cuda-11.0/lib64   -lOpenCL 

C++ benchmarks:

 -I/usr/local/cuda-11.0/include   -L/usr/local/cuda-11.0/lib64   -lOpenCL 

The flags file that was used to format this result can be browsed at
https://www.spec.org/accel/flags/nvidia_flags.20210608.html.

You can also download the XML flags source by saving the following link:
https://www.spec.org/accel/flags/nvidia_flags.20210608.xml.