Intel Core i7-8700K testing of OpenCL Linux drivers by Michael Larabel for a future article.
Compare your own system(s) to this result file with the
Phoronix Test Suite by running the command:
phoronix-test-suite benchmark 1905107-HV-COFFEELAK00
Coffeelake Beignet vs. OpenCL NEO Intel,
"LeelaChessZero 0.20.1 - Backend: OpenCL",
Higher Results Are Better
"Beignet Git",47.278,42.0454,46.6298,68.2958,55.6809,53.3208
"Intel OpenCL NEO",106.477,105.682,105.029
"LuxMark 3.1 - OpenCL Device: GPU - Scene: Hotel",
Higher Results Are Better
"Beignet Git",2,3,134,2,4,5,3,3,4,2
"Intel OpenCL NEO",603,624,612
"Blender 2.79a - Blend File: Barbershop - Compute: OpenCL",
Lower Results Are Better
"Beignet Git",
"Intel OpenCL NEO",
"Blender 2.79a - Blend File: Fishy Cat - Compute: OpenCL",
Lower Results Are Better
"Beignet Git",
"Intel OpenCL NEO",
"Blender 2.79a - Blend File: Pabellon Barcelona - Compute: OpenCL",
Lower Results Are Better
"Beignet Git",
"Intel OpenCL NEO",
"Blender 2.79a - Blend File: Classroom - Compute: OpenCL",
Lower Results Are Better
"Beignet Git",
"Intel OpenCL NEO",
"Blender 2.79a - Blend File: BMW27 - Compute: OpenCL",
Lower Results Are Better
"Beignet Git",
"Intel OpenCL NEO",
"clpeak - OpenCL Test: Single-Precision Float",
Higher Results Are Better
"Beignet Git",56.91,56.91,39.62,56.66,100.38,56.66,56.66,39.5,56.66,39.67,56.67,30.23
"Intel OpenCL NEO",458.81,458.6,458.63
"clpeak - OpenCL Test: Global Memory Bandwidth",
Higher Results Are Better
"Beignet Git",35.41,7.05,33.44,33.41,33.41,33.42,33.48,33.45,33.43,33.47,33.45,33.45
"Intel OpenCL NEO",33.45,33.22,33.16
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: Max SP Flops",
Higher Results Are Better
"Beignet Git",467.229,467.206,467.221
"Intel OpenCL NEO",1452.4,1511.51,1555.53,1457.53,1445.58,1629.06,1523.43,1404.74,1473.09,1430.54,1470.74,1426.82,1522,1367.24,1464.28
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: Texture Read Bandwidth",
Higher Results Are Better
"Beignet Git",60.2543,59.2088,59.7621
"Intel OpenCL NEO",57.0993,57.0952,57.0891
"Parboil 2.5 - Test: OpenCL LBM",
Lower Results Are Better
"Beignet Git",78.597572,78.582695,78.501534
"Intel OpenCL NEO",69.558151,68.334702,69.546814
"Rodinia 2.4 - Test: OpenCL Heartwall",
Lower Results Are Better
"Beignet Git",12.52343416214,42.403953075409,35.008208036423,42.935297966003,35.135001897812,27.999330997467,34.09322810173,35.999274969101,35.790456056595,42.10465502739,42.986444950104,35.081465005875
"Intel OpenCL NEO",9.2038531303406,9.0929470062256,9.1912491321564
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: MD5 Hash",
Higher Results Are Better
"Beignet Git",0.373,0.373,0.373
"Intel OpenCL NEO",0.3853,0.3853,0.3852
"Xsbench OpenCL 2017-07-06 - ",
Higher Results Are Better
"Beignet Git",11454155,11459518,11452836
"Darktable 2.4.2 - Test: Boat - Acceleration: OpenCL",
Lower Results Are Better
"Beignet Git",13.804,13.89,13.903
"Intel OpenCL NEO",13.774,13.907,13.934
"ViennaCL 1.4.2 - OpenCL LU Factorization",
Higher Results Are Better
"Beignet Git",7.45085,7.46709,8.24211,9.04299,8.26838,7.81095,7.92373,12.564,7.60631,8.69682,7.42672,9.74913
"Intel OpenCL NEO",21.1944,21.2256,21.2841
"CoMD OpenCL 2017-07-06 - Average Atom Update Rate",
Higher Results Are Better
"Beignet Git",2.58,2.57,2.59
"Intel OpenCL NEO",2.57,2.58,2.58
"clpeak - OpenCL Test: Transfer Bandwidth enqueueReadBuffer",
Higher Results Are Better
"Beignet Git",15.46,15.4,15.39
"Intel OpenCL NEO",15.64,15.27,14.62,15.61,15.57
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: Triad",
Higher Results Are Better
"Beignet Git",10.4807,6.998,10.7522,10.1357,10.0389,10.7158,9.6515,10.5579,10.35,10.4438,6.9536,7.1363,9.8079,7.022,6.9507
"Intel OpenCL NEO",13.7875,13.9617,13.9745
"Darktable 2.4.2 - Test: Masskrug - Acceleration: OpenCL",
Lower Results Are Better
"Beignet Git",6.462,6.477,6.496
"Intel OpenCL NEO",6.454,6.498,6.476
"clpeak - OpenCL Test: Transfer Bandwidth enqueueWriteBuffer",
Higher Results Are Better
"Beignet Git",40.96,40.51,40.98
"Intel OpenCL NEO",41.47,41.17,41.29
"Darktable 2.4.2 - Test: Server Room - Acceleration: OpenCL",
Lower Results Are Better
"Beignet Git",4.745,4.749,4.769
"Intel OpenCL NEO",4.788,4.755,4.789
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: FFT SP",
Higher Results Are Better
"Beignet Git",15.9807,15.9801,15.9802
"Intel OpenCL NEO",17.6678,17.6844,17.5922
"clpeak - OpenCL Test: Kernel Latency",
Lower Results Are Better
"Beignet Git",21.98,21.93,21.19
"Intel OpenCL NEO",23.66,24.02,23.82
"Parboil 2.5 - Test: OpenCL BFS",
Lower Results Are Better
"Beignet Git",0.959125,0.965082,0.962884
"Intel OpenCL NEO",1.022457,1.032213,1.059206
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: Bus Speed Readback",
Higher Results Are Better
"Beignet Git",37.7729,40.3298,40.392,37.7729,39.0677,39.1845
"Intel OpenCL NEO",30.1062,28.0724,29.4562,29.4271
"SHOC Scalable HeterOgeneous Computing 2015-11-10 - Target: OpenCL - Benchmark: Bus Speed Download",
Higher Results Are Better
"Beignet Git",20.6575,20.448,20.337
"Intel OpenCL NEO",29.8081,29.3072,29.4257
"Darktable 2.4.2 - Test: Server Rack - Acceleration: OpenCL",
Lower Results Are Better
"Beignet Git",0.183,0.183,0.186
"Intel OpenCL NEO",0.186,0.187,0.19