Contents
Introduction
OpenCL is a way to use the GPU in some graphics cards for additional general-purpose processing. Support was committed for OpenCL in FreeBSD Ports in revision r397198
Architecture
OpenCL providers on FreeBSD are installed as "ocl-icd" modules. ocl-icd stands for "OpenCL - Installable Client Driver". This provides a flexible setup, not locking consuming software into using a specific OpenCL implementation, and theoretically allowing more than one provider to be used at the same time. On the other hand, users must choose and install the OpenCL providers they wish to use.
OpenCL Providers
The following OpenCL Providers are available:
Clover: Mesa's OpenCL implementation, targeting all Gallium drivers, only Radeon GPUs in our case.
IntelĀ® Graphics Compute Runtime: For Intel GPU's, Ivy Bridge and later (intel-compute-runtime GitHub Repository)
PortableCL (pocl): A software (CPU) OpenCL implementation.
There is also the information utility, clinfo, which can be compared with glxinfo but for OpenCL. You can benchmark your OpenCL device computational efficiency using clpeak utility.
Notes:
- For either Beignet or Clover to work, a graphics card supported by the kernel KMS drivers is needed.
- Beignet does not support Sandy Bridge class GPUs, only Ivy Bridge and higher.
Related OpenCL Ports
Port |
Notes |
devel/clinfo |
glxinfo-like program for OpenCL |
devel/ocl-icd |
OpenCL - Installable Client Driver library |
benchmarks/clpeak |
Benchmarking tool to measure peak capabilities of opencl devices |
lang/clover |
OpenCL driver for Radeon GPUs |
lang/intel-compute-runtime |
OpenCL driver for Intel GPUs |
lang/pocl |
Another OpenCL driver that uses the CPU for computations |
Examples
Below are Clover and Intel-Compute-Runtime utilities example outputs.
Clover
> uname -a
FreeBSD magellan.dumbbell.fr 11.0-CURRENT FreeBSD 11.0-CURRENT #0 67cdc8d(kms-drm-update-38): Thu Feb 5 09:39:28 CET 2015 root@magellan.dumbbell.fr:/usr/obj/home/dumbbell/Projects/freebsd/src/GIT/sys/GENERIC amd64
drmn0: <ATI Mobility Radeon HD 5870> on vgapci0
info: [drm] initializing kernel modesetting (JUNIPER 0x1002:0x68A0 0x1558:0x8687).
drmn0: info: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
drmn0: info: GTT: 512M 0x0000000040000000 - 0x000000005FFFFFFF
> clinfo
Number of platforms 1
Platform Name Clover
Platform Vendor Mesa
Platform Version OpenCL 1.1 MESA 10.4.3
Platform Profile FULL_PROFILE
Platform Extensions cl_khr_icd
Platform Extensions function suffix MESA
Platform Name Clover
Number of devices 1
Device Name AMD JUNIPER
Device Vendor X.Org
Device Vendor ID 0x1002
Device Version OpenCL 1.1 MESA 10.4.3
Driver Version 10.4.3
Device OpenCL C Version OpenCL C 1.1
Device Type GPU
Device Profile FULL_PROFILE
Max compute units 1
Max clock frequency 0MHz
Max work item dimensions 3
Max work item sizes 256x256x256
Max work group size 256
Preferred work group size multiple 1
Preferred / native vector sizes
char 16 / 16
short 8 / 8
int 4 / 4
long 2 / 2
half 0 / 0 (n/a)
float 4 / 4
double 2 / 2 (n/a)
Half-precision Floating-point support (n/a)
Single-precision Floating-point support (core)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero No
Round to infinity No
IEEE754-2008 fused multiply-add No
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Double-precision Floating-point support (n/a)
Address bits 32, Little-Endian
Global memory size 1073741824 (1024MiB)
Error Correction support No
Max memory allocation 268435456 (256MiB)
Unified memory for Host and Device Yes
Minimum alignment for any data type 128 bytes
Alignment of base address 128 bits (16 bytes)
Global Memory cache type None
Image support No
Local memory type Local
Local memory size 32768 (32KiB)
Max constant buffer size 268435456 (256MiB)
Max number of constant args 13
Max size of kernel argument 1024
Queue properties
Out-of-order execution No
Profiling Yes
Profiling timer resolution 0ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
Device Available Yes
Compiler Available Yes
Device Extensions
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) Clover
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [MESA]
clCreateContext(NULL, ...) [default] Success [MESA]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) Success (1)
Platform Name Clover
Device Name AMD JUNIPER
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) Success (1)
Platform Name Clover
Device Name AMD JUNIPER
ICD loader properties
ICD loader Name OpenCL ICD Loader
ICD loader Vendor OCL Icd free software
ICD loader Version 2.2.3
ICD loader Profile OpenCL 1.2
Intel-Compute-Runtime
# uname -a
FreeBSD 12.2-RELEASE-p1 FreeBSD 12.2-RELEASE-p1 GENERIC amd64
# dmesg | grep drm
drmn0: <drmn> on vgapci0
vgapci0: child drmn0 requested pci_enable_io
vgapci0: child drmn0 requested pci_enable_io
[drm] Unable to create a private tmpfs mount, hugepage support will be disabled(-19).
[drm] Got stolen memory base 0xde000000, size 0x2000000
[drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[drm] Driver supports precise vblank timestamp query.
[drm] Connector eDP-1: get mode from tunables:
[drm] - kern.vt.fb.modes.eDP-1
[drm] - kern.vt.fb.default_mode
[drm] Connector HDMI-A-1: get mode from tunables:
[drm] - kern.vt.fb.modes.HDMI-A-1
[drm] - kern.vt.fb.default_mode
[drm] Connector DP-1: get mode from tunables:
[drm] - kern.vt.fb.modes.DP-1
[drm] - kern.vt.fb.default_mode
[drm] Initialized i915 1.6.0 20171222 for drmn0 on minor 0
name=drmn0 flags=0x0 stride=10240 bpp=32
drmn0: fb0: inteldrmfb frame buffer device
# pkg info -x drm
drm-fbsd12.0-kmod-4.16.g20201016
drm-kmod-g20190710
drm_info-2.2.0_1
libdrm-2.4.103,1
linux-c7-libdrm-2.4.91
# clinfo
Number of platforms 1
Platform Name Intel(R) OpenCL HD Graphics
Platform Vendor Intel(R) Corporation
Platform Version OpenCL 3.0
Platform Numeric Version 0xc00000 (3.0.0)
Platform Profile FULL_PROFILE
Platform Extensions cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_intel_subgroups cl_intel_required_subgroup_size cl_intel_subgroups_short cl_khr_spir cl_intel_accelerator cl_intel_driver_diagnostics cl_khr_priority_hints cl_khr_throttle_hints cl_khr_create_command_queue cl_intel_subgroups_char cl_intel_subgroups_long cl_khr_il_program cl_intel_mem_force_host_memory cl_khr_subgroup_extended_types cl_khr_subgroup_non_uniform_vote cl_khr_subgroup_ballot cl_khr_subgroup_non_uniform_arithmetic cl_khr_subgroup_shuffle cl_khr_subgroup_shuffle_relative cl_khr_subgroup_clustered_reduce cl_khr_fp64 cl_khr_subgroups cl_intel_spirv_media_block_io cl_intel_spirv_subgroups cl_khr_spirv_no_integer_wrap_decoration cl_intel_unified_shared_memory_preview cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_intel_planar_yuv cl_intel_packed_yuv cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_3d_image_writes cl_intel_media_block_io cl_intel_va_api_media_sharing
Platform Extensions with Version cl_khr_byte_addressable_store 0x400000 (1.0.0)
cl_khr_fp16 0x400000 (1.0.0)
cl_khr_global_int32_base_atomics 0x400000 (1.0.0)
cl_khr_global_int32_extended_atomics 0x400000 (1.0.0)
cl_khr_icd 0x400000 (1.0.0)
cl_khr_local_int32_base_atomics 0x400000 (1.0.0)
cl_khr_local_int32_extended_atomics 0x400000 (1.0.0)
cl_intel_subgroups 0x400000 (1.0.0)
cl_intel_required_subgroup_size 0x400000 (1.0.0)
cl_intel_subgroups_short 0x400000 (1.0.0)
cl_khr_spir 0x400000 (1.0.0)
cl_intel_accelerator 0x400000 (1.0.0)
cl_intel_driver_diagnostics 0x400000 (1.0.0)
cl_khr_priority_hints 0x400000 (1.0.0)
cl_khr_throttle_hints 0x400000 (1.0.0)
cl_khr_create_command_queue 0x400000 (1.0.0)
cl_intel_subgroups_char 0x400000 (1.0.0)
cl_intel_subgroups_long 0x400000 (1.0.0)
cl_khr_il_program 0x400000 (1.0.0)
cl_intel_mem_force_host_memory 0x400000 (1.0.0)
cl_khr_subgroup_extended_types 0x400000 (1.0.0)
cl_khr_subgroup_non_uniform_vote 0x400000 (1.0.0)
cl_khr_subgroup_ballot 0x400000 (1.0.0)
cl_khr_subgroup_non_uniform_arithmetic 0x400000 (1.0.0)
cl_khr_subgroup_shuffle 0x400000 (1.0.0)
cl_khr_subgroup_shuffle_relative 0x400000 (1.0.0)
cl_khr_subgroup_clustered_reduce 0x400000 (1.0.0)
cl_khr_fp64 0x400000 (1.0.0)
cl_khr_subgroups 0x400000 (1.0.0)
cl_intel_spirv_media_block_io 0x400000 (1.0.0)
cl_intel_spirv_subgroups 0x400000 (1.0.0)
cl_khr_spirv_no_integer_wrap_decoration 0x400000 (1.0.0)
cl_intel_unified_shared_memory_preview 0x400000 (1.0.0)
cl_khr_mipmap_image 0x400000 (1.0.0)
cl_khr_mipmap_image_writes 0x400000 (1.0.0)
cl_intel_planar_yuv 0x400000 (1.0.0)
cl_intel_packed_yuv 0x400000 (1.0.0)
cl_khr_int64_base_atomics 0x400000 (1.0.0)
cl_khr_int64_extended_atomics 0x400000 (1.0.0)
cl_khr_image2d_from_buffer 0x400000 (1.0.0)
cl_khr_depth_images 0x400000 (1.0.0)
cl_khr_3d_image_writes 0x400000 (1.0.0)
cl_intel_media_block_io 0x400000 (1.0.0)
cl_intel_va_api_media_sharing 0x400000 (1.0.0)
Platform Host timer resolution 1ns
Platform Extensions function suffix INTEL
Platform Name Intel(R) OpenCL HD Graphics
Number of devices 1
Device Name Intel(R) Graphics [0x1616]
Device Vendor Intel(R) Corporation
Device Vendor ID 0x8086
Device Version OpenCL 3.0 NEO
Device Numeric Version 0xc00000 (3.0.0)
Driver Version 21.02.18820
Device OpenCL C Version OpenCL C 3.0
Device OpenCL C all versions OpenCL C 0x400000 (1.0.0)
OpenCL C 0x401000 (1.1.0)
OpenCL C 0x402000 (1.2.0)
OpenCL C 0x800000 (2.0.0)
OpenCL C 0xc00000 (3.0.0)
Device OpenCL C features __opencl_c_int64 0xc00000 (3.0.0)
__opencl_c_3d_image_writes 0xc00000 (3.0.0)
__opencl_c_images 0xc00000 (3.0.0)
__opencl_c_read_write_images 0xc00000 (3.0.0)
__opencl_c_atomic_order_acq_rel 0xc00000 (3.0.0)
__opencl_c_atomic_order_seq_cst 0xc00000 (3.0.0)
__opencl_c_atomic_scope_all_devices 0xc00000 (3.0.0)
__opencl_c_atomic_scope_device 0xc00000 (3.0.0)
__opencl_c_generic_address_space 0xc00000 (3.0.0)
__opencl_c_program_scope_global_variables 0xc00000 (3.0.0)
__opencl_c_work_group_collective_functions 0xc00000 (3.0.0)
__opencl_c_subgroups 0xc00000 (3.0.0)
__opencl_c_device_enqueue 0xc00000 (3.0.0)
__opencl_c_pipes 0xc00000 (3.0.0)
__opencl_c_fp64 0xc00000 (3.0.0)
Latest comfornace test passed v2020-11-23-00
Device Type GPU
Device Profile FULL_PROFILE
Device Available Yes
Compiler Available Yes
Linker Available Yes
Max compute units 24
Max clock frequency 0MHz
Device Partition (core)
Max number of sub-devices 0
Supported partition types None
Supported affinity domains (n/a)
Max work item dimensions 3
Max work item sizes 256x256x256
Max work group size 256
Preferred work group size multiple (device) 32
Preferred work group size multiple (kernel) 32
Max sub-groups per work group 32
Sub-group sizes (Intel) 8, 16, 32
Preferred / native vector sizes
char 16 / 16
short 8 / 8
int 4 / 4
long 1 / 1
half 8 / 8 (cl_khr_fp16)
float 1 / 1
double 1 / 1 (cl_khr_fp64)
Half-precision Floating-point support (cl_khr_fp16)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Single-precision Floating-point support (core)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Correctly-rounded divide and sqrt operations Yes
Double-precision Floating-point support (cl_khr_fp64)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Address bits 64, Little-Endian
Global memory size 6747164672 (6.284GiB)
Error Correction support No
Max memory allocation 2139095040 (1.992GiB)
Unified memory for Host and Device Yes
Shared Virtual Memory (SVM) capabilities (core)
Coarse-grained buffer sharing Yes
Fine-grained buffer sharing No
Fine-grained system sharing No
Atomics No
Minimum alignment for any data type 128 bytes
Alignment of base address 1024 bits (128 bytes)
Preferred alignment for atomics
SVM 64 bytes
Global 64 bytes
Local 64 bytes
Atomic memory capabilities relaxed, acquire/release, sequentially-consistent, work-group scope, device scope, all-devices scope
Atomic fence capabilities relaxed, acquire/release, sequentially-consistent, work-item scope, work-group scope, device scope, all-devices scope
Max size for global variable 65536 (64KiB)
Preferred total size of global vars 2139095040 (1.992GiB)
Global Memory cache type Read/Write
Global Memory cache size 262144 (256KiB)
Global Memory cache line size 64 bytes
Image support Yes
Max number of samplers per kernel 16
Max size for 1D images from buffer 133693440 pixels
Max 1D or 2D image array size 2048 images
Base address alignment for 2D image buffers 4 bytes
Pitch alignment for 2D image buffers 4 pixels
Max 2D image size 16384x16384 pixels
Max planar YUV image size 16384x16352 pixels
Max 3D image size 2048x2048x2048 pixels
Max number of read image args 128
Max number of write image args 128
Max number of read/write image args 128
Pipe support Yes
Max number of pipe args 16
Max active pipe reservations 1
Max pipe packet size 1024
Local memory type Local
Local memory size 65536 (64KiB)
Max number of constant args 8
Max constant buffer size 2139095040 (1.992GiB)
Generic address space support Yes
Max size of kernel argument 2048 (2KiB)
Queue properties (on host)
Out-of-order execution Yes
Profiling Yes
Device enqueue capabilities supported, replaceable default queue
Queue properties (on device)
Out-of-order execution Yes
Profiling Yes
Preferred size 131072 (128KiB)
Max size 67108864 (64MiB)
Max queues on device 1
Max events on device 1024
Prefer user sync for interop Yes
Profiling timer resolution 80ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
Non-uniform work-groups Yes
Work-group collective functions Yes
Sub-group independent forward progress Yes
IL version SPIR-V_1.2
ILs with version SPIR-V 0x402000 (1.2.0)
SPIR versions 1.2
printf() buffer size 4194304 (4MiB)
Built-in kernels (n/a)
Built-in kernels with version (n/a)
Device Extensions cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_intel_subgroups cl_intel_required_subgroup_size cl_intel_subgroups_short cl_khr_spir cl_intel_accelerator cl_intel_driver_diagnostics cl_khr_priority_hints cl_khr_throttle_hints cl_khr_create_command_queue cl_intel_subgroups_char cl_intel_subgroups_long cl_khr_il_program cl_intel_mem_force_host_memory cl_khr_subgroup_extended_types cl_khr_subgroup_non_uniform_vote cl_khr_subgroup_ballot cl_khr_subgroup_non_uniform_arithmetic cl_khr_subgroup_shuffle cl_khr_subgroup_shuffle_relative cl_khr_subgroup_clustered_reduce cl_khr_fp64 cl_khr_subgroups cl_intel_spirv_media_block_io cl_intel_spirv_subgroups cl_khr_spirv_no_integer_wrap_decoration cl_intel_unified_shared_memory_preview cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_intel_planar_yuv cl_intel_packed_yuv cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_3d_image_writes cl_intel_media_block_io cl_intel_va_api_media_sharing
Device Extensions with Version cl_khr_byte_addressable_store 0x400000 (1.0.0)
cl_khr_fp16 0x400000 (1.0.0)
cl_khr_global_int32_base_atomics 0x400000 (1.0.0)
cl_khr_global_int32_extended_atomics 0x400000 (1.0.0)
cl_khr_icd 0x400000 (1.0.0)
cl_khr_local_int32_base_atomics 0x400000 (1.0.0)
cl_khr_local_int32_extended_atomics 0x400000 (1.0.0)
cl_intel_subgroups 0x400000 (1.0.0)
cl_intel_required_subgroup_size 0x400000 (1.0.0)
cl_intel_subgroups_short 0x400000 (1.0.0)
cl_khr_spir 0x400000 (1.0.0)
cl_intel_accelerator 0x400000 (1.0.0)
cl_intel_driver_diagnostics 0x400000 (1.0.0)
cl_khr_priority_hints 0x400000 (1.0.0)
cl_khr_throttle_hints 0x400000 (1.0.0)
cl_khr_create_command_queue 0x400000 (1.0.0)
cl_intel_subgroups_char 0x400000 (1.0.0)
cl_intel_subgroups_long 0x400000 (1.0.0)
cl_khr_il_program 0x400000 (1.0.0)
cl_intel_mem_force_host_memory 0x400000 (1.0.0)
cl_khr_subgroup_extended_types 0x400000 (1.0.0)
cl_khr_subgroup_non_uniform_vote 0x400000 (1.0.0)
cl_khr_subgroup_ballot 0x400000 (1.0.0)
cl_khr_subgroup_non_uniform_arithmetic 0x400000 (1.0.0)
cl_khr_subgroup_shuffle 0x400000 (1.0.0)
cl_khr_subgroup_shuffle_relative 0x400000 (1.0.0)
cl_khr_subgroup_clustered_reduce 0x400000 (1.0.0)
cl_khr_fp64 0x400000 (1.0.0)
cl_khr_subgroups 0x400000 (1.0.0)
cl_intel_spirv_media_block_io 0x400000 (1.0.0)
cl_intel_spirv_subgroups 0x400000 (1.0.0)
cl_khr_spirv_no_integer_wrap_decoration 0x400000 (1.0.0)
cl_intel_unified_shared_memory_preview 0x400000 (1.0.0)
cl_khr_mipmap_image 0x400000 (1.0.0)
cl_khr_mipmap_image_writes 0x400000 (1.0.0)
cl_intel_planar_yuv 0x400000 (1.0.0)
cl_intel_packed_yuv 0x400000 (1.0.0)
cl_khr_int64_base_atomics 0x400000 (1.0.0)
cl_khr_int64_extended_atomics 0x400000 (1.0.0)
cl_khr_image2d_from_buffer 0x400000 (1.0.0)
cl_khr_depth_images 0x400000 (1.0.0)
cl_khr_3d_image_writes 0x400000 (1.0.0)
cl_intel_media_block_io 0x400000 (1.0.0)
cl_intel_va_api_media_sharing 0x400000 (1.0.0)
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) Intel(R) OpenCL HD Graphics
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [INTEL]
clCreateContext(NULL, ...) [default] Success [INTEL]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) Success (1)
Platform Name Intel(R) OpenCL HD Graphics
Device Name Intel(R) Graphics [0x1616]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) Success (1)
Platform Name Intel(R) OpenCL HD Graphics
Device Name Intel(R) Graphics [0x1616]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) Success (1)
Platform Name Intel(R) OpenCL HD Graphics
Device Name Intel(R) Graphics [0x1616]
ICD loader properties
ICD loader Name OpenCL ICD Loader
ICD loader Vendor OCL Icd free software
ICD loader Version 2.2.13
ICD loader Profile OpenCL 3.0
# clpeak
Platform: Intel(R) OpenCL HD Graphics
Device: Intel(R) Graphics [0x1616]
Driver version : 21.02.18820 (FreeBSD)
Compute units : 24
Clock frequency : 0 MHz
Global memory bandwidth (GBPS)
float : 12.23
float2 : 15.00
float4 : 15.58
float8 : 15.34
float16 : 14.01
Single-precision compute (GFLOPS)
float : 257.06
float2 : 258.25
float4 : 259.26
float8 : 240.18
float16 : 252.49
Half-precision compute (GFLOPS)
half : 259.30
half2 : 251.59
half4 : 257.62
half8 : 239.43
half16 : 251.78
Double-precision compute (GFLOPS)
double : 69.22
double2 : 68.70
double4 : 67.88
double8 : 68.08
double16 : 64.21
Integer compute (GIOPS)
int : 73.58
int2 : 74.52
int4 : 80.22
int8 : 78.41
int16 : 81.11
Transfer bandwidth (GBPS)
enqueueWriteBuffer : 4.24
enqueueReadBuffer : 5.36
enqueueMapBuffer(for read) : 21738.77
memcpy from mapped ptr : 4.70
enqueueUnmap(after write) : 1069547.50
memcpy to mapped ptr : 5.41
Kernel launch latency : 73.42 us