1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Titan X pascal benchmark on SYS-4028GR-TR

Discussion in 'Linux-Bench Results and Discussion' started by dhenzjhen, Nov 2, 2016.

  1. dhenzjhen

    dhenzjhen Member

    Joined:
    Sep 14, 2016
    Messages:
    28
    Likes Received:
    35
    [​IMG]

    Motherboard: X10DRG-O / Product Name: SYS-4028GR-TR
    BIOS: 7/27/2016
    IPMI: 3.44
    CPU: E5 2689 3.1Ghz V4 x 2
    Memory: Samsung 16GB x 24
    GPU: Nvidia Titan X Pascal x 10
    OS: Redhat Linux 7.2 x64
    Driver: 367.57
    Software: HPL binary from Nvidia (NDA)


    #./run_me_10_gpu
    =======================================================================
    HPLinpack 2.1 -- High-Performance Linpack benchmark -- October 26, 2012
    Written by A. Petitet and R. Clint Whaley, Innovative Computing Laboratory, UTK
    Modified by Piotr Luszczek, Innovative Computing Laboratory, UTK
    Modified by Julien Langou, University of Colorado Denver
    =======================================================================

    An explanation of the input/output parameters follows:
    T/V : Wall time / encoded variant.
    N : The order of the coefficient matrix A.
    NB : The partitioning blocking factor.
    P : The number of process rows.
    Q : The number of process columns.
    Time : Time in seconds to solve the linear system.
    Gflops : Rate of execution for solving the linear system.

    The following parameter values will be used:

    N : 117760 0 0 0
    NB : 384
    PMAP : Row-major process mapping
    P : 5
    Q : 2
    PFACT : Left
    NBMIN : 8
    NDIV : 2
    RFACT : Left
    BCAST : 1ring
    DEPTH : 0
    SWAP : Spread-roll (long)
    L1 : no-transposed form
    U : transposed form
    EQUIL : no
    ALIGN : 8 double precision words

    --------------------------------------------------------------------------------

    - The matrix A is randomly generated for each test.
    - The following scaled residual check will be computed:
    ||Ax-b||_oo / ( eps * ( || x ||_oo * || A ||_oo + || b ||_oo ) * N )
    - The relative machine precision (eps) is taken to be 1.110223e-16
    - Computational tests pass if scaled residuals are less than 16.0

    gpu_dgemm_split from environment variable 1.000
    test_loops from environment variable 1

    ******** TESTING SYSTEM PARAMETERS ********
    PARAM [UNITS] MIN MAX AVG
    ----- ------- --- --- ---
    CPU :
    CPU_BW [GB/s ] 4.1 4.9 4.5
    CPU_FP [GFLPS] 60.8 68.6 64.7
    PCIE :
    H2D_BW [GB/s ] 2.3 2.4 2.3
    D2H_BW [GB/s ] 2.4 2.9 2.7
    BID_BW [GB/s ] 4.0 4.2 4.1
    GPU :
    GPU_BW [GB/s ] 382 385 384
    GPU_FP [GFLPS]
    NB = 128 382 393 388
    NB = 256 387 396 391
    NB = 384 388 397 393
    NB = 512 389 398 394
    NB = 640 390 399 394
    NB = 768 390 399 394
    NB = 896 390 399 394
    NB = 1024 389 399 394
    NET :
    NET_BW [MB/s ]
    8 B 3 5 4
    64 B 52 108 63
    512 B 156 188 171
    4 KB 696 743 719
    32 KB 1478 1706 1590
    256 KB 2579 2708 2631
    2048 KB 3099 3452 3199
    16384 KB 2990 3304 3074
    NET_LAT [ us ] 0.5 1.6 0.8

    displaying Prog:%complete, N:columns, Time:seconds
    iGF:instantaneous GF, GF:avg GF, GF_per: process GF


    Per-Process Host Memory Estimate: 11.50 GB (MAX) 11.26 GB (MIN)

    PCOL: 0 GPU_COLS: 59009 CPU_COLS: 0
    PCOL: 1 GPU_COLS: 58753 CPU_COLS: 0
    test_loop: 1 of 1
    2016-11-02 15:52:02.227

    Prog= 1.94% N_left= 116992 Time= 5.96 Time_left= 300.80 iGF= 3548.99 GF= 3548.99 iGF_per= 354.90 GF_per= 354.90
    Prog= 3.86% N_left= 116224 Time= 11.63 Time_left= 289.42 iGF= 3687.18 GF= 3616.31 iGF_per=
    I uploaded the rest of the time calculations here http://chmod755.sdf-us.org/titanxpascal

    ================================================================================
    T/V N NB P Q Time Gflops
    --------------------------------------------------------------------------------
    WR00L2L8 117760 384 5 2 316.71 3.438e+03
    --------------------------------------------------------------------------------
    ||Ax-b||_oo/(eps*(||A||_oo*||x||_oo+||b||_oo)*N)= 0.0028690 ...... PASSED

    Per-Process Host Memory Estimate: 0.00 GB (MAX) 0.00 GB (MIN)

    PCOL: 0 GPU_COLS: 1 CPU_COLS: 0
    PCOL: 1 GPU_COLS: 1 CPU_COLS: 0
    test_loop: 1 of 1
    2016-11-02 15:57:33.703
    2016-11-02 15:57:33.703
    ================================================================================
    T/V N NB P Q Time Gflops
    --------------------------------------------------------------------------------
    WR00L2L8 0 384 5 2 0.00 0.000e+00
    --------------------------------------------------------------------------------
    ||Ax-b||_oo/(eps*(||A||_oo*||x||_oo+||b||_oo)*N)= 0.0000000 ...... PASSED

    Per-Process Host Memory Estimate: 0.00 GB (MAX) 0.00 GB (MIN)

    PCOL: 1 GPU_COLS: 1 CPU_COLS: 0
    PCOL: 0 GPU_COLS: 1 CPU_COLS: 0
    test_loop: 1 of 1
    2016-11-02 15:57:33.802
    2016-11-02 15:57:33.802
    ================================================================================
    T/V N NB P Q Time Gflops
    --------------------------------------------------------------------------------
    WR00L2L8 0 384 5 2 0.00 0.000e+00
    --------------------------------------------------------------------------------
    ||Ax-b||_oo/(eps*(||A||_oo*||x||_oo+||b||_oo)*N)= 0.0000000 ...... PASSED

    Per-Process Host Memory Estimate: 0.00 GB (MAX) 0.00 GB (MIN)

    PCOL: 0 GPU_COLS: 1 CPU_COLS: 0
    PCOL: 1 GPU_COLS: 1 CPU_COLS: 0
    test_loop: 1 of 1
    2016-11-02 15:57:33.898
    2016-11-02 15:57:33.898
    ================================================================================
    T/V N NB P Q Time Gflops
    --------------------------------------------------------------------------------
    WR00L2L8 0 384 5 2 0.00 0.000e+00
    --------------------------------------------------------------------------------
    ||Ax-b||_oo/(eps*(||A||_oo*||x||_oo+||b||_oo)*N)= 0.0000000 ...... PASSED
    ================================================================================

    Finished 4 tests with the following results:
    4 tests completed and passed residual checks,
    0 tests completed and failed residual checks,
    0 tests skipped because of illegal input values.
    --------------------------------------------------------------------------------

    End of Tests.
    ================================================================================


    Note: this is just a quick benchmark
     
    #1
    jerrytsao and Patrick like this.
  2. dhenzjhen

    dhenzjhen Member

    Joined:
    Sep 14, 2016
    Messages:
    28
    Likes Received:
    35
    # nvidia-smi
    Wed Nov 2 15:59:56 2016
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 367.57 Driver Version: 367.57 |
    |-------------------------------+----------------------+----------------------+
    | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
    | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
    |===============================+======================+=============|
    | 0 TITAN X (Pascal) Off | 0000:04:00.0 Off | N/A |
    | 23% 42C P0 56W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 1 TITAN X (Pascal) Off | 0000:05:00.0 Off | N/A |
    | 24% 44C P0 59W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 2 TITAN X (Pascal) Off | 0000:06:00.0 Off | N/A |
    | 24% 44C P0 59W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 3 TITAN X (Pascal) Off | 0000:07:00.0 Off | N/A |
    | 25% 45C P0 57W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 4 TITAN X (Pascal) Off | 0000:08:00.0 Off | N/A |
    | 27% 49C P0 58W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 5 TITAN X (Pascal) Off | 0000:0B:00.0 Off | N/A |
    | 23% 42C P0 57W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 6 TITAN X (Pascal) Off | 0000:0C:00.0 Off | N/A |
    | 24% 45C P0 58W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 7 TITAN X (Pascal) Off | 0000:0D:00.0 Off | N/A |
    | 24% 44C P0 57W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 8 TITAN X (Pascal) Off | 0000:0E:00.0 Off | N/A |
    | 24% 44C P0 57W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+
    | 9 TITAN X (Pascal) Off | 0000:0F:00.0 Off | N/A |
    | 23% 43C P0 58W / 250W | 0MiB / 12189MiB | 0% Default |
    +-------------------------------+----------------------+----------------------+

    +-----------------------------------------------------------------------------+
    | Processes: GPU Memory |
    | GPU PID Type Process name Usage |
    |=====================================================================|
    | No running processes found |
    +-----------------------------------------------------------------------------+
     
    #2
    MiniKnight, William and Patrick like this.
  3. William

    William Active Member

    Joined:
    May 7, 2015
    Messages:
    516
    Likes Received:
    181
    WOW, nice job DJ !
     
    #3
    dhenzjhen likes this.
  4. MiniKnight

    MiniKnight Well-Known Member

    Joined:
    Mar 30, 2012
    Messages:
    2,185
    Likes Received:
    580
    Is that 3.438TFLOPS or am I reading it wrong?
     
    #4
  5. dhenzjhen

    dhenzjhen Member

    Joined:
    Sep 14, 2016
    Messages:
    28
    Likes Received:
    35
    yessir thats a 3400Gflops equals to 3.4Tflops.
     
    #5
    Last edited: Nov 2, 2016
Similar Threads: Titan pascal
Forum Title Date
Linux-Bench Results and Discussion Titan X pascal NCCL tesing on SYS-4028GR-TR Nov 7, 2016

Share This Page