SPEC® ACCEL_ACC Result

Copyright 2014-2015 Standard Performance Evaluation Corporation

ASUS (Test Sponsor: NVIDIA Corporation)

NVIDIA Tesla K40c

ASUS P9X79 Motherboard

SPECaccel_acc_base = 2.98

SPECaccel_acc_energy_base = 3.49

SPECaccel_acc_peak = 3.15

SPECaccel_acc_energy_peak = 3.63

ACCEL license: 019 Test date: Feb-2014
Test sponsor: NVIDIA Corporation Hardware Availability: Nov-2013
Tested by: NVIDIA Corporation Software Availability: Feb-2014
Benchmark results graph
Hardware
CPU Name: Intel Core i7-3930K
CPU Characteristics:
CPU MHz: 3200
CPU MHz Maximum: 3800
FPU: Integrated
CPU(s) enabled: 6 cores, 1 chip, 6 cores/chip, 2 threads/core
CPU(s) orderable: 1 chip
Primary Cache: 32 KB I + 32 KB D on chip per core
Secondary Cache: 256 KB I+D on chip per core
L3 Cache: 12 MB I+D on chip per chip
Other Cache: None
Memory: 8 GB (2 x 4 GB 2Rx4 PC3-14900R-9, running at 1600
MHz)
Disk Subsystem: 1000 GB Seagate ST1000DM003 7200 RPM SATA
Other Hardware: None
Accelerator
Accel Model Name: Tesla K40c
Accel Vendor: NVIDIA
Accel Name: NVIDIA Tesla K40c
Type of Accel: GPU
Accel Connection: PCIe 3.0 16x
Does Accel Use ECC: No
Accel Description: GPU Boost set to maximum graphic clock frequency
of 875 MHz. See notes below.
Accel Driver: NVIDIA UNIX x86_64 Kernel Module 319.60
Software
Operating System: Red Hat Enterprise Linux Server release 6.4
(Santiago)
2.6.32-358.el6.x86_64
Compiler: PGI Accelerator Server Complete, Release 14.2
File System: ext4
System State: Run level 3 (multi-user)
Other Software: FFTW 3.3.3
Power
Power Supply: 1200 W
Power Supply Details: Thermaltake SMART M1200W
Max. Power (W): 411.52
Idle Power (W): 94.85
Min. Temperature (C): 25.63
Power Analyzer
Power Analyzer: Power Analyzer
Hardware Vendor: Xitron Technologies, Inc.
Model: 2801
Serial Number: 28011109005
Input Connection: RS232 via USB-adapter
Metrology Institute: NIST
Calibration By: Micro Precision Calibration, Inc.
Calibration Label: 220081222038459
Calibration Date: 02.20.2014
PTDaemon Version: 1.6.2 (372e138a; 2013-12-04)
Setup Description: connected to the single power supply
that powers the system
Current Ranges Used: 2.0A
Voltage Range Used: 135V
Temperature Meter
Temperature Meter: Temperature Meter
Hardware Vendor: Digi
Model: DigiWATCHPORT_H
Serial Number: WS34682143
Input Connection: USB
PTDaemon Version: 1.6.2 (372e138a; 2013-12-04)
Setup Description: Position 5mm above intake fan

Base Results Table

Benchmark Seconds Ratio Energy
(kJ)
Maximum
Power
Average
Power
Energy
Ratio
Seconds Ratio Energy
(kJ)
Maximum
Power
Average
Power
Energy
Ratio
Seconds Ratio Energy
(kJ)
Maximum
Power
Average
Power
Energy
Ratio
Results appear in the order in which they were run. Bold underlined text indicates a median measurement.
303.ostencil 45.2  3.21  16.0  366   354   3.31  45.2  3.21  16.2  366   359   3.26  45.1  3.21  16.4  366   364   3.22 
304.olbm 93.2  4.88  27.1  301   291   5.75  93.1  4.89  27.3  308   293   5.72  93.2  4.88  27.1  301   291   5.76 
314.omriq 298    3.21  108    412   362   3.54  298    3.21  108    389   362   3.54  298    3.21  108    388   362   3.55 
350.md 90.0  2.80  31.2  352   347   3.01  90.1  2.80  31.1  352   346   3.02  90.0  2.80  31.2  353   347   3.01 
351.palm 183    2.03  41.0  235   224   2.61  180    2.06  40.4  242   225   2.64  180    2.05  40.3  235   224   2.65 
352.ep 328    1.62  83.4  274   254   1.97  328    1.62  83.1  256   253   1.98  328    1.62  82.9  255   253   1.98 
353.clvrleaf 129    3.45  41.5  326   321   3.80  129    3.45  41.3  326   320   3.82  129    3.45  41.3  326   320   3.82 
354.cg 140    2.92  39.4  320   283   3.55  140    2.92  39.4  328   282   3.55  140    2.92  39.3  319   282   3.56 
355.seismic 125    2.96  35.6  296   285   3.62  125    2.96  35.5  293   284   3.64  125    2.96  35.5  293   284   3.64 
356.sp 98.0  2.82  27.8  288   284   3.34  98.0  2.82  27.8  288   284   3.34  98.0  2.82  27.8  288   283   3.34 
357.csp 123    2.19  33.4  281   270   2.64  123    2.19  33.3  273   269   2.65  123    2.19  33.3  285   270   2.64 
359.miniGhost 107    3.45  30.2  307   282   3.94  107    3.46  30.1  307   282   3.95  107    3.46  30.0  307   281   3.95 
360.ilbdc 76.4  4.80  23.8  323   312   5.45  76.5  4.80  23.8  323   312   5.45  76.4  4.80  23.8  322   312   5.45 
363.swim 83.7  2.75  21.9  263   262   3.46  84.0  2.74  22.0  270   262   3.46  84.1  2.74  21.9  263   261   3.46 
370.bt 68.3  3.26  18.8  277   276   4.02  68.3  3.26  18.8  277   275   4.02  68.4  3.26  18.8  276   275   4.02 

Peak Results Table

Benchmark Seconds Ratio Energy
(kJ)
Maximum
Power
Average
Power
Energy
Ratio
Seconds Ratio Energy
(kJ)
Maximum
Power
Average
Power
Energy
Ratio
Seconds Ratio Energy
(kJ)
Maximum
Power
Average
Power
Energy
Ratio
Results appear in the order in which they were run. Bold underlined text indicates a median measurement.
303.ostencil 45.2  3.21  16.3  366   360   3.25  45.2  3.21  16.4  379   364   3.22  45.2  3.21  16.2  366   359   3.26 
304.olbm 93.2  4.88  27.1  301   291   5.76  93.1  4.89  27.2  300   292   5.75  93.2  4.88  27.2  300   292   5.74 
314.omriq 298    3.21  108    405   362   3.55  298    3.21  108    389   362   3.55  298    3.21  108    389   361   3.55 
350.md 86.1  2.93  30.6  361   355   3.07  86.1  2.93  30.7  361   356   3.06  86.1  2.93  30.8  361   357   3.06 
351.palm 140    2.64  33.1  246   236   3.23  140    2.64  32.6  251   233   3.28  140    2.65  32.5  245   233   3.29 
352.ep 328    1.62  83.1  272   253   1.98  328    1.62  82.9  255   253   1.98  328    1.62  83.1  255   253   1.98 
353.clvrleaf 129    3.45  41.2  326   319   3.82  129    3.44  41.3  326   319   3.82  129    3.44  41.3  326   320   3.81 
354.cg 140    2.92  39.3  319   281   3.56  140    2.92  39.6  323   283   3.54  140    2.92  39.4  319   283   3.55 
355.seismic 125    2.96  35.5  301   284   3.63  125    2.96  35.5  293   284   3.64  125    2.96  35.5  295   284   3.64 
356.sp 98.0  2.82  27.9  288   284   3.33  98.0  2.82  27.8  288   283   3.34  98.0  2.82  27.8  289   283   3.34 
357.csp 123    2.19  33.2  273   269   2.65  124    2.19  33.2  273   268   2.65  123    2.19  33.2  273   269   2.66 
359.miniGhost 99.4  3.71  28.9  343   291   4.11  99.4  3.71  28.9  341   291   4.11  99.4  3.71  29.0  339   292   4.10 
360.ilbdc 76.5  4.80  23.9  321   313   5.43  76.4  4.80  23.8  330   312   5.45  76.4  4.80  23.8  321   312   5.46 
363.swim 52.5  4.38  15.7  299   298   4.85  52.5  4.38  15.6  299   298   4.86  52.5  4.38  15.7  299   298   4.85 
370.bt 68.3  3.27  18.7  277   274   4.04  68.3  3.26  18.8  277   274   4.03  68.3  3.27  18.8  277   275   4.03 

Platform Notes

 Sysinfo program /local/home/SPECACCEL/Docs/sysinfo
 $Rev: 6874 $ $Date:: 2013-11-20 #$ 0953404ef7e75a5f9bbb534c6de3f831
 running on sbe02 Sun Feb 23 07:09:46 2014

 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 : Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
       1 "physical id"s (chips)
       12 "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 : 6
       siblings  : 12
       physical 0: cores 0 1 2 3 4 5
    cache size : 12288 KB

 From /proc/meminfo
    MemTotal:        8130700 kB
    HugePages_Total:       0
    Hugepagesize:       2048 kB

 /usr/bin/lsb_release -d
    Red Hat Enterprise Linux Server release 6.4 (Santiago)

 From /etc/*release* /etc/*version*
    redhat-release: Red Hat Enterprise Linux Server release 6.4 (Santiago)
    system-release: Red Hat Enterprise Linux Server release 6.4 (Santiago)
    system-release-cpe: cpe:/o:redhat:enterprise_linux:6server:ga:server

 uname -a:
    Linux sbe02 2.6.32-358.el6.x86_64 #1 SMP Tue Jan 29 11:47:41 EST 2013 x86_64
    x86_64 x86_64 GNU/Linux

 run-level 3 Feb 22 13:29

 SPEC is set to: /local/home/SPECACCEL
    Filesystem    Type    Size  Used Avail Use% Mounted on
    /dev/mapper/VolGroup-lv_home
                  ext4    860G   52G  764G   7% /local
 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.


 (End of data from sysinfo program)
 Information from pgaccelinfo
 CUDA Driver Version:           5050
 NVRM version:                  NVIDIA UNIX x86_64 Kernel Module  319.60
                                Wed Sep 25 14:28:26 PDT 2013
 Device Number:                 0
 Device Name:                   Tesla K40c
 Device Revision Number:        3.5
 Global Memory Size:            12079136768
 Number of Multiprocessors:     15
 Number of SP Cores:            2880
 Number of DP Cores:            960
 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:                    875 MHz
 Max. Clock Rate:               875 MHz
 Execution Timeout:             No
 Integrated Device:             No
 Can Map Host Memory:           Yes
 Compute Mode:                  default
 Concurrent Kernels:            Yes
 ECC Enabled:                   No
 Memory Clock Rate:             3004 MHz
 Memory Bus Width:              384 bits
 L2 Cache Size:                 1572864 bytes
 Max Threads Per SMP:           2048
 Async Engines:                 2
 Unified Addressing:            Yes

General Notes

 ECC disabled using the command "nvidia-smi -e 0"
 GPU Boost mode enabled by setting the device to persistant mode: "nvidia-smi -pm 1"
 and then setting the memory and graphic clock using: "nvidia-smi -ac <MEM>,<GRH>".
 For this run, the memory clock was not changed from the default 3004 MHz.
 The graphic clock was set to the maximum frequency of 875 MHz.
 nvidai-smi -ac 3004,875
 Kit built system using a CoolMaster HAF X case

Base Compiler Invocation

C benchmarks:

 pgcc 

Fortran benchmarks:

 pgfortran 

Benchmarks using both Fortran and C:

 pgcc   pgfortran 

Base Optimization Flags

C benchmarks:

 -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5 

Fortran benchmarks:

 -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5 

Benchmarks using both Fortran and C:

353.clvrleaf:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5 
359.miniGhost:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5   -Mnomain 

Peak Compiler Invocation

C benchmarks:

 pgcc 

Fortran benchmarks:

 pgfortran 

Benchmarks using both Fortran and C:

 pgcc   pgfortran 

Peak Portability Flags

351.palm:  -DSPEC_HOST_FFTW3 

Peak Optimization Flags

C benchmarks:

 -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5 

Fortran benchmarks:

350.md:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5   -ta=tesla:maxregcount:48 
351.palm:  -fast   -Mfprelaxed   -acc=noautopar   -ta=tesla:cc35   -ta=tesla:cuda5.5   -ta=tesla:fastmath   -lfftw3 
355.seismic:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5 
356.sp:  Same as 355.seismic 
360.ilbdc:  Same as 355.seismic 
363.swim:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5   -ta=tesla:pin 

Benchmarks using both Fortran and C:

353.clvrleaf:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5 
359.miniGhost:  -fast   -Mfprelaxed   -acc   -ta=tesla:cc35   -ta=tesla:cuda5.5   -ta=tesla:maxregcount:32   -Mnomain 

Peak Other Flags

Fortran benchmarks:

351.palm:  -I/local/home/SPECACCEL/fftw333/include   -L/local/home/SPECACCEL/fftw333/lib 

The flags file that was used to format this result can be browsed at
http://www.spec.org/accel/flags/pgi2014_flags.20150303.html.

You can also download the XML flags source by saving the following link:
http://www.spec.org/accel/flags/pgi2014_flags.20150303.xml.