Intel Arc Graphics vs. NVIDIA Linux Compute - October 2023

Benchmarks 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 2310268-NE-COMPUTEAR41
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts

Limit displaying results to tests within:

NVIDIA GPU Compute 4 Tests
OpenCL 3 Tests

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Geometric Means Per-Suite/Category
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Disable Color Branding
Prefer Vertical Bar Graphs
No Box Plots
On Line Graphs With Missing Data, Connect The Line Gaps

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs
Condense Test Profiles With Multiple Version Results Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Hide
Result
Result
Identifier
Performance Per
Dollar
Date
Run
  Test
  Duration
Arc A380
October 16 2023
  44 Minutes
Arc A580
October 16 2023
  21 Minutes
Arc A750
October 15 2023
  33 Minutes
Arc A770
October 14 2023
  20 Minutes
RTX 3060
October 25 2023
  16 Minutes
RTX 3060 Ti
October 25 2023
  13 Minutes
RTX 3070
October 24 2023
  13 Minutes
RTX 3070 Ti
October 25 2023
  11 Minutes
RTX 3080
October 26 2023
  10 Minutes
RTX 4060
October 24 2023
  14 Minutes
Invert Hiding All Results Option
  20 Minutes

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):


Intel Arc Graphics vs. NVIDIA Linux Compute - October 2023ProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorOSKernelDesktopDisplay ServerOpenGLOpenCLCompilerFile-SystemScreen ResolutionDisplay DriverArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060Intel Core i9-13900K @ 5.50GHz (24 Cores / 32 Threads)ASUS PRIME Z790-P WIFI (0812 BIOS)Intel Device 7a2732GBWestern Digital WD_BLACK SN850X 4000GB + Western Digital WD_BLACK SN850X 1000GBIntel Arc A380 DG2 6GB (2450MHz)Realtek ALC897ASUS VP28UUbuntu 23.106.6.0-060600rc5-generic (x86_64)GNOME Shell 45.0X Server + Wayland4.6 Mesa 23.3~git2310140600.29e2e9~oibaf~m (git-29e2e92 2023-10-14 mantic-oibaf-ppa)OpenCL 3.0GCC 13.2.0ext41920x1080Intel Arc A580 DG2 8GB (2400MHz)Intel Arc A750 DG2 8GB (2400MHz)Intel Arc A770 DG2 16GB (2400MHz)eVGA NVIDIA GeForce RTX 3060 12GBX Server 1.21.1.7NVIDIA 545.23.064.6.0OpenCL 3.0 CUDA 12.3.68NVIDIA GeForce RTX 3060 Ti 8GBNVIDIA GeForce RTX 3070 8GBNVIDIA GeForce RTX 3070 Ti 8GBNVIDIA GeForce RTX 3080 10GBMSI NVIDIA GeForce RTX 4060 8GBOpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseCompiler Details- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v Processor Details- Scaling Governor: intel_pstate performance (EPP: performance) - CPU Microcode: 0x119 - Thermald 2.5.4Python Details- Python 3.11.6Security Details- gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected Graphics Details- RTX 3060: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 94.06.14.40.46- RTX 3060 Ti: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 94.04.25.00.2c- RTX 3070: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 94.04.25.00.2b- RTX 3070 Ti: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 94.04.5b.00.02- RTX 3080: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 94.02.20.00.07- RTX 4060: BAR1 / Visible vRAM Size: 256 MiB - vBIOS Version: 95.07.31.00.e3OpenCL Details- RTX 3060: GPU Compute Cores: 3584- RTX 3060 Ti: GPU Compute Cores: 4864- RTX 3070: GPU Compute Cores: 5888- RTX 3070 Ti: GPU Compute Cores: 6144- RTX 3080: GPU Compute Cores: 8704- RTX 4060: GPU Compute Cores: 3072

Arc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060Result OverviewPhoronix Test Suite100%274%448%621%795%clpeakFluidX3DHashcatVkResampleSHOC Scalable HeterOgeneous Computing

Intel Arc Graphics vs. NVIDIA Linux Compute - October 2023fluidx3d: FP32-FP32shoc: OpenCL - Texture Read Bandwidthfluidx3d: FP32-FP16Cfluidx3d: FP32-FP16Sshoc: OpenCL - S3Dvkresample: 2x - Singlehashcat: MD5hashcat: SHA-512hashcat: TrueCrypt RIPEMD160 + XTSclpeak: Integer Computeclpeak: Integer 24-bit Computeclpeak: Global Memory Bandwidthclpeak: Single-Precision Computeshoc: OpenCL - FFT SPshoc: OpenCL - Triadshoc: OpenCL - MD5 Hashshoc: OpenCL - Bus Speed Readbackshoc: OpenCL - Bus Speed DownloadArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060622454.7451115109778.370162.113100888833332845000001010251484.661484.04120.433315.82422.11210.47206.437811.234211.77172532731.21734833882214.64720.611261869333337978166672815834065.044068.45387.879748.261170.3719.991018.542422.424123.45292470766.31439524212225.25419.173300868333339003833333127674833.884834.19396.8011371.111192.0420.505022.336022.392223.24522616983.09342714422226.19017.3183515513333310784000003837575530.715519.11397.6213008.681219.9120.021926.088222.437423.470021411270.5735434143172.71326.320247977000009776166672906206311.446336.87315.0512309.21886.87223.115615.105326.344724.026726162107.3746665116212.44221.0913465095714313553833334021387798.688000.27379.9915298.761123.7523.401821.196426.337323.998726282124.5350275191219.88020.22941589700000164761666749496310067.0510095.67390.4419674.251135.7323.380725.331126.343224.000735152076.6159216837286.62715.81242332871429165923333349878810643.4410700.73529.7920722.731509.2423.748926.226326.342724.016741822205.2877588226339.34712.76158402900000230095000067628814227.5514210.91638.2227261.611919.3323.919636.515726.341824.022216311790.4031583085156.10532.5372997046666712179333333657677559.917587.93238.3914783.09722.94212.343418.924913.192812.8371OpenBenchmarking.org

FluidX3D

FluidX3D is a speedy and memory efficient Boltzmann CFD (Computational Fluid Dynamics) software package implemented using OpenCL and intended for GPU acceleration. FluidX3D is developed by Moritz Lehmann and written free for non-commercial use. This is a test profile measuring the system OpenCL performance using the FluidX3D benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMLUPs/s, More Is BetterFluidX3D 2.9Test: FP32-FP32Arc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40609001800270036004500SE +/- 0.00, N = 3SE +/- 2.73, N = 3SE +/- 8.02, N = 3SE +/- 5.57, N = 3SE +/- 0.00, N = 3SE +/- 1.45, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 2.52, N = 3SE +/- 0.00, N = 3622253224702616214126162628351541821631

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Texture Read BandwidthArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40605001000150020002500SE +/- 2.94, N = 3SE +/- 0.06, N = 3SE +/- 21.20, N = 15SE +/- 1.66, N = 3SE +/- 2.22, N = 4SE +/- 3.45, N = 4SE +/- 4.90, N = 4SE +/- 5.97, N = 4SE +/- 4.76, N = 3SE +/- 2.12, N = 3454.75731.22766.31983.091270.572107.372124.532076.612205.281790.401. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt

FluidX3D

FluidX3D is a speedy and memory efficient Boltzmann CFD (Computational Fluid Dynamics) software package implemented using OpenCL and intended for GPU acceleration. FluidX3D is developed by Moritz Lehmann and written free for non-commercial use. This is a test profile measuring the system OpenCL performance using the FluidX3D benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMLUPs/s, More Is BetterFluidX3D 2.9Test: FP32-FP16CArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 406017003400510068008500SE +/- 0.33, N = 3SE +/- 20.04, N = 3SE +/- 14.29, N = 3SE +/- 45.67, N = 3SE +/- 2.73, N = 3SE +/- 7.51, N = 3SE +/- 1.00, N = 3SE +/- 3.53, N = 3SE +/- 7.80, N = 3SE +/- 0.67, N = 31115348339524271354346665027592177583158

OpenBenchmarking.orgMLUPs/s, More Is BetterFluidX3D 2.9Test: FP32-FP16SArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40602K4K6K8K10KSE +/- 10.37, N = 3SE +/- 8.41, N = 3SE +/- 0.88, N = 3SE +/- 30.66, N = 3SE +/- 0.33, N = 3SE +/- 2.33, N = 3SE +/- 0.33, N = 3SE +/- 0.88, N = 3SE +/- 3.06, N = 3SE +/- 0.00, N = 31097388242124422414351165191683782263085

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: S3DArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 406070140210280350SE +/- 0.55, N = 3SE +/- 0.72, N = 3SE +/- 1.04, N = 3SE +/- 0.84, N = 3SE +/- 0.01, N = 13SE +/- 0.03, N = 13SE +/- 0.05, N = 13SE +/- 0.07, N = 13SE +/- 0.05, N = 13SE +/- 0.02, N = 378.37214.65225.25226.19172.71212.44219.88286.63339.35156.111. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt

VkResample

VkResample is a Vulkan-based image upscaling library based on VkFFT. The sample input file is upscaling a 4K image to 8K using Vulkan-based GPU acceleration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: SingleArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40601428425670SE +/- 0.04, N = 3SE +/- 0.00, N = 6SE +/- 0.01, N = 6SE +/- 0.00, N = 6SE +/- 0.00, N = 5SE +/- 0.01, N = 5SE +/- 0.00, N = 5SE +/- 0.00, N = 5SE +/- 0.02, N = 5SE +/- 0.00, N = 362.1120.6119.1717.3226.3221.0920.2315.8112.7632.541. (CXX) g++ options: -O3

Hashcat

Hashcat is an open-source, advanced password recovery tool supporting GPU acceleration with OpenCL, NVIDIA CUDA, and Radeon ROCm. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: MD5Arc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 406013000M26000M39000M52000M65000MSE +/- 10038904.88, N = 6SE +/- 52502537.50, N = 6SE +/- 32959759.98, N = 6SE +/- 14282919.09, N = 6SE +/- 40085550.18, N = 6SE +/- 39649619.33, N = 7SE +/- 25896432.92, N = 7SE +/- 42136771.15, N = 7SE +/- 62354336.92, N = 6SE +/- 48062540.28, N = 310088883333261869333333008683333335155133333247977000003465095714341589700000423328714295840290000029970466667

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: SHA-512Arc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060500M1000M1500M2000M2500MSE +/- 862167.81, N = 6SE +/- 871174.18, N = 6SE +/- 6115031.57, N = 6SE +/- 594138.03, N = 5SE +/- 852610.37, N = 6SE +/- 3243497.77, N = 6SE +/- 2876620.32, N = 6SE +/- 2377767.39, N = 6SE +/- 6840650.55, N = 6SE +/- 88191.71, N = 3284500000797816667900383333107840000097761666713553833331647616667165923333323009500001217933333

OpenBenchmarking.orgH/s, More Is BetterHashcat 6.2.4Benchmark: TrueCrypt RIPEMD160 + XTSArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060140K280K420K560K700KSE +/- 149.30, N = 4SE +/- 319.81, N = 6SE +/- 1857.90, N = 6SE +/- 178.43, N = 7SE +/- 2070.36, N = 15SE +/- 2383.72, N = 8SE +/- 403.09, N = 8SE +/- 1315.08, N = 8SE +/- 2376.78, N = 8SE +/- 2383.51, N = 3101025281583312767383757290620402138494963498788676288365767

clpeak

Clpeak is designed to test the peak capabilities of OpenCL devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGIOPS, More Is Betterclpeak 1.1.2OpenCL Test: Integer ComputeArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40603K6K9K12K15KSE +/- 0.22, N = 4SE +/- 3.59, N = 3SE +/- 3.74, N = 3SE +/- 7.88, N = 3SE +/- 3.38, N = 13SE +/- 13.43, N = 13SE +/- 5.08, N = 13SE +/- 47.35, N = 13SE +/- 73.29, N = 13SE +/- 13.40, N = 31484.664065.044833.885530.716311.447798.6810067.0510643.4414227.557559.911. (CXX) g++ options: -O3

OpenBenchmarking.orgGIOPS, More Is Betterclpeak 1.1.2OpenCL Test: Integer 24-bit ComputeArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40603K6K9K12K15KSE +/- 0.13, N = 4SE +/- 0.41, N = 3SE +/- 2.06, N = 3SE +/- 3.66, N = 3SE +/- 8.59, N = 13SE +/- 42.07, N = 13SE +/- 17.10, N = 13SE +/- 41.98, N = 13SE +/- 48.38, N = 13SE +/- 0.58, N = 31484.044068.454834.195519.116336.878000.2710095.6710700.7314210.917587.931. (CXX) g++ options: -O3

OpenBenchmarking.orgGBPS, More Is Betterclpeak 1.1.2OpenCL Test: Global Memory BandwidthArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060140280420560700SE +/- 0.08, N = 5SE +/- 0.07, N = 8SE +/- 0.15, N = 8SE +/- 0.13, N = 8SE +/- 0.84, N = 9SE +/- 0.53, N = 11SE +/- 0.02, N = 11SE +/- 0.68, N = 15SE +/- 0.92, N = 11SE +/- 0.06, N = 3120.43387.87396.80397.62315.05379.99390.44529.79638.22238.391. (CXX) g++ options: -O3

OpenBenchmarking.orgGFLOPS, More Is Betterclpeak 1.1.2OpenCL Test: Single-Precision ComputeArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 40606K12K18K24K30KSE +/- 0.38, N = 3SE +/- 1.76, N = 3SE +/- 2.25, N = 3SE +/- 0.78, N = 3SE +/- 13.72, N = 13SE +/- 67.44, N = 13SE +/- 38.39, N = 13SE +/- 137.23, N = 13SE +/- 181.88, N = 13SE +/- 0.10, N = 33315.829748.2611371.1113008.6812309.2115298.7619674.2520722.7327261.6114783.091. (CXX) g++ options: -O3

SHOC Scalable HeterOgeneous Computing

The CUDA and OpenCL version of Vetter's Scalable HeterOgeneous Computing benchmark suite. SHOC provides a number of different benchmark programs for evaluating the performance and stability of compute devices. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: FFT SPArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060400800120016002000SE +/- 0.22, N = 12SE +/- 9.38, N = 15SE +/- 7.59, N = 13SE +/- 9.02, N = 15SE +/- 0.26, N = 12SE +/- 0.05, N = 12SE +/- 0.28, N = 12SE +/- 0.06, N = 12SE +/- 0.31, N = 12SE +/- 0.20, N = 3422.111170.371192.041219.91886.871123.751135.731509.241919.33722.941. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: TriadArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060612182430SE +/- 0.01, N = 12SE +/- 0.07, N = 12SE +/- 0.07, N = 12SE +/- 0.06, N = 15SE +/- 0.01, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.01, N = 13SE +/- 0.01, N = 310.4719.9920.5120.0223.1223.4023.3823.7523.9212.341. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt

OpenBenchmarking.orgGHash/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: MD5 HashArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060816243240SE +/- 0.0096, N = 10SE +/- 0.0057, N = 13SE +/- 0.0328, N = 13SE +/- 0.0148, N = 15SE +/- 0.0089, N = 12SE +/- 0.0208, N = 13SE +/- 0.0181, N = 13SE +/- 0.0040, N = 13SE +/- 0.1231, N = 14SE +/- 0.0125, N = 36.437818.542422.336026.088215.105321.196425.331126.226336.515718.92491. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed ReadbackArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060612182430SE +/- 0.00, N = 12SE +/- 0.00, N = 13SE +/- 0.01, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 311.2322.4222.3922.4426.3426.3426.3426.3426.3413.191. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt

OpenBenchmarking.orgGB/s, More Is BetterSHOC Scalable HeterOgeneous Computing 2020-04-17Target: OpenCL - Benchmark: Bus Speed DownloadArc A380Arc A580Arc A750Arc A770RTX 3060RTX 3060 TiRTX 3070RTX 3070 TiRTX 3080RTX 4060612182430SE +/- 0.00, N = 12SE +/- 0.00, N = 14SE +/- 0.02, N = 14SE +/- 0.00, N = 14SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 13SE +/- 0.00, N = 311.7723.4523.2523.4724.0324.0024.0024.0224.0212.841. (CXX) g++ options: -O2 -lSHOCCommonOpenCL -lSHOCCommon -lOpenCL -lrt