Intel OpenCL NEO Ubuntu 19.04
Intel Core i7-6770HQ OpenCL Linux benchmarking by Michael Larabel for a future article.


Beignet OpenCL 2.0: 

	Processor: Intel Core i7-6770HQ @ 3.50GHz (4 Cores / 8 Threads), Motherboard: Intel NUC6i7KYB (KYSKLi70.86A.0037.2016.0603.1032 BIOS), Chipset: Intel Xeon E3-1200 v5/E3-1500, Memory: 32768MB, Disk: Samsung SSD 950 PRO 512GB, Graphics: Intel Iris Pro 580 3GB (950MHz), Audio: Realtek ALC233, Monitor: DELL P2415Q, Network: Intel I219-LM + Intel 8260

	OS: Ubuntu 19.04, Kernel: 5.0.0-13-generic (x86_64), Desktop: GNOME Shell 3.32.0, Display Server: X Server 1.20.4, Display Driver: modesetting 1.20.4, OpenGL: 4.5 Mesa 19.1.0-devel (git-bdd273d 2019-05-06 disco-oibaf-ppa), OpenCL: OpenCL 2.0 beignet 1.3, Compiler: GCC 8.3.0, File-System: ext4, Screen Resolution: 3840x2160

NEO OpenCL 2.1: 

	Processor: Intel Core i7-6770HQ @ 3.50GHz (4 Cores / 8 Threads), Motherboard: Intel NUC6i7KYB (KYSKLi70.86A.0037.2016.0603.1032 BIOS), Chipset: Intel Xeon E3-1200 v5/E3-1500, Memory: 32768MB, Disk: Samsung SSD 950 PRO 512GB, Graphics: Intel Iris Pro 580 3GB (950MHz), Audio: Realtek ALC233, Monitor: DELL P2415Q, Network: Intel I219-LM + Intel 8260

	OS: Ubuntu 19.04, Kernel: 5.0.0-13-generic (x86_64), Desktop: GNOME Shell 3.32.0, Display Server: X Server 1.20.4, Display Driver: modesetting 1.20.4, OpenGL: 4.5 Mesa 19.1.0-devel (git-bdd273d 2019-05-06 disco-oibaf-ppa), OpenCL: OpenCL 2.1, Compiler: GCC 8.3.0, File-System: ext4, Screen Resolution: 3840x2160


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: Triad
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 8.54  |=========================
NEO OpenCL 2.1 ..... 17.18 |===================================================


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: FFT SP
GFLOPS > Higher Is Better
Beignet OpenCL 2.0 . 28.63 |===================================
NEO OpenCL 2.1 ..... 41.24 |===================================================


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: MD5 Hash
GHash/s > Higher Is Better
Beignet OpenCL 2.0 . 0.89 |=================================================
NEO OpenCL 2.1 ..... 0.94 |====================================================


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: Max SP Flops
GFLOPS > Higher Is Better
Beignet OpenCL 2.0 . 845  |==============
NEO OpenCL 2.1 ..... 3176 |====================================================


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: Bus Speed Download
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 10.73 |====================
NEO OpenCL 2.1 ..... 27.63 |===================================================


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: Bus Speed Readback
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 18.62 |==================================
NEO OpenCL 2.1 ..... 27.93 |===================================================


SHOC Scalable HeterOgeneous Computing 2015-11-10
Target: OpenCL - Benchmark: Texture Read Bandwidth
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 141 |=====================================================
NEO OpenCL 2.1 ..... 128 |================================================


ViennaCL 1.4.2
OpenCL LU Factorization
GFLOPS > Higher Is Better
Beignet OpenCL 2.0 . 12.92 |=================================
NEO OpenCL 2.1 ..... 20.25 |===================================================


cl-mem 2017-01-13
Benchmark: Copy
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 46.65 |===================================================


cl-mem 2017-01-13
Benchmark: Read
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 43.09 |===================================================


cl-mem 2017-01-13
Benchmark: Write
GB/s > Higher Is Better
Beignet OpenCL 2.0 . 48.92 |===================================================


LeelaChessZero 0.20.1
Backend: OpenCL
Nodes Per Second > Higher Is Better
Beignet OpenCL 2.0 . 116 |========================================
NEO OpenCL 2.1 ..... 155 |=====================================================


Darktable 2.6.0
Test: Boat - Acceleration: OpenCL
Seconds < Lower Is Better
Beignet OpenCL 2.0 . 16.59 |===================================================
NEO OpenCL 2.1 ..... 15.64 |================================================


Darktable 2.6.0
Test: Masskrug - Acceleration: OpenCL
Seconds < Lower Is Better
Beignet OpenCL 2.0 . 12.02 |===================================================
NEO OpenCL 2.1 ..... 10.67 |=============================================


Darktable 2.6.0
Test: Server Rack - Acceleration: OpenCL
Seconds < Lower Is Better
Beignet OpenCL 2.0 . 0.28 |============================
NEO OpenCL 2.1 ..... 0.52 |====================================================


Darktable 2.6.0
Test: Server Room - Acceleration: OpenCL
Seconds < Lower Is Better
Beignet OpenCL 2.0 . 8.41 |====================================================
NEO OpenCL 2.1 ..... 5.30 |=================================


Blender 2.79a
Blend File: BMW27 - Compute: OpenCL
Seconds < Lower Is Better
Beignet OpenCL 2.0 . 695 |=====================================================
NEO OpenCL 2.1 ..... 679 |====================================================


Blender 2.79a
Blend File: Barbershop - Compute: OpenCL
Seconds < Lower Is Better
Beignet OpenCL 2.0 . 2779 |====================================================
NEO OpenCL 2.1 ..... 2745 |===================================================


Xsbench OpenCL 2017-07-06

Lookups/s > Higher Is Better
Beignet OpenCL 2.0 . 23624949 |================================================


JuliaGPU 1.2pts1
OpenCL Device: GPU
Samples/sec > Higher Is Better
Beignet OpenCL 2.0 . 49218726 |===================================
NEO OpenCL 2.1 ..... 66870065 |================================================


clpeak
OpenCL Test: Kernel Latency
us < Lower Is Better
Beignet OpenCL 2.0 . 31.25 |===============================================
NEO OpenCL 2.1 ..... 33.80 |===================================================


clpeak
OpenCL Test: Single-Precision Float
GFLOPS > Higher Is Better
Beignet OpenCL 2.0 . 1054 |===================================================
NEO OpenCL 2.1 ..... 1077 |====================================================


clpeak
OpenCL Test: Global Memory Bandwidth
GBPS > Higher Is Better
Beignet OpenCL 2.0 . 25.64 |===================================================
NEO OpenCL 2.1 ..... 25.78 |===================================================


clpeak
OpenCL Test: Transfer Bandwidth enqueueReadBuffer
GBPS > Higher Is Better
Beignet OpenCL 2.0 . 9.37 |===================================================
NEO OpenCL 2.1 ..... 9.59 |====================================================


clpeak
OpenCL Test: Transfer Bandwidth enqueueWriteBuffer
GBPS > Higher Is Better
Beignet OpenCL 2.0 . 28.94 |==================================================
NEO OpenCL 2.1 ..... 29.47 |===================================================


CoMD OpenCL 2017-07-06
Average Atom Update Rate
us/atom/task > Higher Is Better
Beignet OpenCL 2.0 . 3.73 |====================================================
NEO OpenCL 2.1 ..... 3.61 |==================================================