/
SPEC2017

SPEC2017

Repository

https://dev-private-git.linaro.org/restricted-benchmarks/CPU2017.git/

 

Status

  1. IntRate/FpRate with no_fortran can run on WOA with clang build.

  2. Fortran test need to be support after flang is OK.

  3. FpSpeed tests run failed on clang

  4. Test case 526 clang result wrong need to be fixed, it looks x86 clang also failed on Windows.

 

Install

  1. Clone the code from the repo.

  2. Install.bat

Run

  1. shrc.bat

  2. download this 2 files and put them into config folder

 

3. ‘runcpu no_fortran -c clang-woa’ or ‘runcpu no_fortran -c msvc’

when we use clang to build, please make sure the clang version >= 15
clang 14 build failed on 523/623.

Compile Options

All compile options can be adjust in config file. It is in the config folder.

We can find the variable CXXOPTIMIZE/OPTIMIZE to modify the options.

 

Result

O0 Result:

 

MSVC 19.33.31629

O0

Clang15

O0

 

500.perlbench_r

1158.876899

1.373744

1175.391681

1.354442

-1.41%

502.gcc_r

876.47172

1.615568

874.259943

1.619656

0.25%

505.mcf_r

840.492653

1.922682

830.327836

1.946219

1.22%

520.omnetpp_r

1968.219366

0.666592

1941.018189

0.675934

1.40%

523.xalancbmk_r

1602.713984

0.658882

1670.802401

0.632032

-4.08%

525.x264_r

2240.751749

0.781434

1599.872048

1.094463

40.06%

531.deepsjeng_r

945.577264

1.211958

858.81092

1.334403

10.10%

541.leela_r

2542.71477

0.651272

2407.003167

0.687992

5.64%

548.exchange2_r

 

 

 

 

 

557.xz_r

1181.348912

0.914209

1099.669843

0.982113

7.43%

 

 

 

 

 

 

SPECrate2017_int_base

 

1.006158

 

1.067484

6.10%

 

 

 

 

 

 

503.bwaves_r

 

 

 

 

 

507.cactuBSSN_r

 

 

 

 

 

508.namd_r

1156.462228

0.821471

976.556179

0.972806

18.42%

510.parest_r

 

 

2897.094787

0.902974

 

511.povray_r

1648.060358

1.416817

1689.090041

1.382401

-2.43%

519.lbm_r

730.963926

1.441932

474.349233

2.221992

54.10%

521.wrf_r

 

 

 

 

 

526.blender_r

1131.855406

1.345578

978.278377

1.556817

 

527.cam4_r

 

 

 

 

 

538.imagick_r

1447.23497

1.718449

1248.919507

1.991321

15.88%

544.nab_r

939.029472

1.792276

846.085775

1.98916

10.99%

549.fotonik3d_r

 

 

 

 

 

554.roms_r

 

 

 

 

 

 

 

 

 

 

 

SPECrate2017_fp_base

 

1.381604

 

1.49433

8.16%

O1 result:

 

MSVC 19.33.31629

O1

Clang15

O1

 

500.perlbench_r

577.058571

2.758819

572.917326

2.77876

0.72%

502.gcc_r

401.088755

3.530391

394.771672

3.586884

1.60%

505.mcf_r

436.545381

3.701792

471.227337

3.429343

-7.36%

520.omnetpp_r

878.880531

1.492808

878.687742

1.493136

0.02%

523.xalancbmk_r

394.254187

2.678475

397.621992

2.655789

-0.85%

525.x264_r

665.708082

2.630282

378.619068

4.624701

75.83%

531.deepsjeng_r

389.288872

2.943829

358.815106

3.193845

8.49%

541.leela_r

624.580502

2.65138

572.785861

2.891133

9.04%

548.exchange2_r

 

 

 

 

 

557.xz_r

516.374698

2.091504

537.039062

2.011027

-3.85%

 

 

 

 

 

 

SPECrate2017_int_base

 

2.63815

 

2.829912

7.27%

 

 

 

 

 

 

503.bwaves_r

 

 

 

 

 

507.cactuBSSN_r

 

 

 

 

 

508.namd_r

318.319754

2.984421

253.464223

3.748063

25.59%

510.parest_r

 

 

492.154462

5.315404

 

511.povray_r

651.234819

3.585496

619.564597

3.768776

5.11%

519.lbm_r

293.301958

3.593566

236.65693

4.453704

23.94%

521.wrf_r

 

 

 

 

 

526.blender_r

352.828239

4.316548

 

 

 

527.cam4_r

 

 

 

 

 

538.imagick_r

684.180401

3.635006

435.202322

5.714583

57.21%

544.nab_r

439.243907

3.831584

353.448188

4.76166

24.27%

549.fotonik3d_r

 

 

 

 

 

554.roms_r

Comments