Well, that depends on the OpenCL platforms made available by the various "vendors".

On my 2017 15-inch MacBook Pro QC i7 the only available platform is by Apple, counting 4 devices:

- the Intel QC i7 CPU

- the integrated Intel HD Graphics 630

- the additional AMD Radeon Pro 630

- and the eGPU NVidia GeForce GTX 1080Ti

This info comes from running the OpenCLInfo example from the OpenCL Programming Guide. So, it appears that there are no separate platforms by AMD and NVidia.

It's barely work with my MBP 2017 15" touchbar + aorus gaming box 1080

I installed nvidia web driver, and I installed NVDEGsupport
but I don't see NVDEGsupport.kext in extension.

however I put fitHeadlessHDMI and use internal MacBook screen as using mirror,
Unigine Valley benchmark shows that only got 30fps avg on 1920x1080x8AA ultra, Score about 2xxx, which is 40-50% dropped compare to regular windows computer that show on another review of gaming box 1080.

what can I do to boost performance of gaming box?

Posted by: hkoster1

Well, that depends on the OpenCL platforms made available by the various "vendors".

On my 2017 15-inch MacBook Pro QC i7 the only available platform is by Apple, counting 4 devices:

- the Intel QC i7 CPU

- the integrated Intel HD Graphics 630

- the additional AMD Radeon Pro 630

- and the eGPU NVidia GeForce GTX 1080Ti

This info comes from running the OpenCLInfo example from the OpenCL Programming Guide. So, it appears that there are no separate platforms by AMD and NVidia.

Thank you, @hkoster1. Could I bother you by asking what eGPU enclosure you are using, and what your config is like (MacOS, nvidia drivers, eGPU setup)? I am trying to figure out if my problem is one of configuration (it probably is) or if it is my older MBP (doubtful). I really want to get to the bottom of this... and get cracking with OpenCL by getting a GTX 1080 Ti too, but I won't buy it until it works with the GTX 960.

If you want to discuss this off-forum until we find a solution for my setup, my email address is j.navarro@ieee.org

Posted by: j.navarro

So let me ask it differently: has anyone managed to run OpenCL on an eGPU?

I mean from a Mac over Thunderbolt.

And, if so, what does clinfo say? 

If you can give me some details on how/what to use to test OpenCL I'd be more than happy to play around and see if i can get things working reliably.

Posted by: ricosuave0922

Posted by: j.navarro

So let me ask it differently: has anyone managed to run OpenCL on an eGPU?

I mean from a Mac over Thunderbolt.

And, if so, what does clinfo say? 

If you can give me some details on how/what to use to test OpenCL I'd be more than happy to play around and see if i can get things working reliably.

That would be great, @ricosuave0922, thank you. clinfo can be obtained from

https://github.com/Oblomov/clinfo/archive/master.zip

and it compiles and runs right away for me.

CLInfo

Number of platforms                               1

  Platform Name                                   Apple

  Platform Vendor                                 Apple

  Platform Version                                OpenCL 1.2 (Oct 31 2017 18:19:43)

  Platform Profile                                FULL_PROFILE

  Platform Extensions                             cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event

  Platform Name                                   Apple

Number of devices                                 4

  Device Name                                     Intel(R) Core(TM) i7-7920HQ CPU @ 3.10GHz

  Device Vendor                                   Intel

  Device Vendor ID                                0xffffffff

  Device Version                                  OpenCL 1.2 

  Driver Version                                  1.1

  Device OpenCL C Version                         OpenCL C 1.2 

  Device Type                                     CPU

  Device Available                                Yes

  Device Profile                                  FULL_PROFILE

  Max compute units                               8

  Max clock frequency                             3100MHz

  Device Partition                                (core)

    Max number of sub-devices                     0

    Supported partition types                     None

  Max work item dimensions                        3

  Max work item sizes                             1024x1x1

  Max work group size                             1024

  Compiler Available                              Yes

  Linker Available                                Yes

  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        (cl_khr_fp64)

  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                                 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                              17179869184 (16GiB)

  Error Correction support                        No

  Max memory allocation                           4294967296 (4GiB)

  Unified memory for Host and Device              Yes

  Minimum alignment for any data type             128 bytes

  Alignment of base address                       1024 bits (128 bytes)

  Global Memory cache type                        Read/Write

  Global Memory cache size                        64

  Global Memory cache line size                   8388608 bytes

  Image support                                   Yes

    Max number of samplers per kernel             16

    Max size for 1D images from buffer            65536 pixels

    Max 1D or 2D image array size                 2048 images

    Base address alignment for 2D image buffers   1 bytes

    Pitch alignment for 2D image buffers          1 pixels

    Max 2D image size                             8192x8192 pixels

    Max 3D image size                             2048x2048x2048 pixels

    Max number of read image args                 128

    Max number of write image args                8

  Local memory type                               Global

  Local memory size                               32768 (32KiB)

  Max constant buffer size                        65536 (64KiB)

  Max number of constant args                     8

  Max size of kernel argument                     4096 (4KiB)

  Queue properties                                

    Out-of-order execution                        No

    Profiling                                     Yes

  Prefer user sync for interop                    Yes

  Profiling timer resolution                      1ns

  Execution capabilities                          

    Run OpenCL kernels                            Yes

    Run native kernels                            Yes

  printf() buffer size                            1048576 (1024KiB)

  Built-in kernels                                

  Device Extensions                               cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_image2d_from_buffer cl_APPLE_fp64_basic_ops cl_APPLE_fixed_alpha_channel_orders cl_APPLE_biased_fixed_point_image_formats cl_APPLE_command_queue_priority

  Device Name                                     Intel(R) HD Graphics 630

  Device Vendor                                   Intel Inc.

  Device Vendor ID                                0x1024500

  Device Version                                  OpenCL 1.2 

  Driver Version                                  1.2(Nov  9 2017 18:56:54)

  Device OpenCL C Version                         OpenCL C 1.2 

  Device Type                                     GPU

  Device Available                                Yes

  Device Profile                                  FULL_PROFILE

  Max compute units                               24

  Max clock frequency                             1100MHz

  Device Partition                                (core)

    Max number of sub-devices                     0

    Supported partition types                     None

  Max work item dimensions                        3

  Max work item sizes                             256x256x256

  Max work group size                             256

  Compiler Available                              Yes

  Linker Available                                Yes

  Preferred work group size multiple              32

  Preferred / native vector sizes                 

    char                                                 1 / 1       

    short                                                1 / 1       

    int                                                  1 / 1       

    long                                                 1 / 1       

    half                                                 0 / 0        (n/a)

    float                                                1 / 1       

    double                                               0 / 0        (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                                 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         (n/a)

  Address bits                                    64, Little-Endian

  Global memory size                              1610612736 (1.5GiB)

  Error Correction support                        No

  Max memory allocation                           402653184 (384MiB)

  Unified memory for Host and Device              Yes

  Minimum alignment for any data type             128 bytes

  Alignment of base address                       1024 bits (128 bytes)

  Global Memory cache type                        None

  Image support                                   Yes

    Max number of samplers per kernel             16

    Max size for 1D images from buffer            25165824 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          32 pixels

    Max 2D image size                             16384x16384 pixels

    Max 3D image size                             2048x2048x2048 pixels

    Max number of read image args                 128

    Max number of write image args                8

  Local memory type                               Local

  Local memory size                               65536 (64KiB)

  Max constant buffer size                        65536 (64KiB)

  Max number of constant args                     8

  Max size of kernel argument                     1024

  Queue properties                                

    Out-of-order execution                        No

    Profiling                                     Yes

  Prefer user sync for interop                    Yes

  Profiling timer resolution                      80ns

  Execution capabilities                          

    Run OpenCL kernels                            Yes

    Run native kernels                            No

  printf() buffer size                            1048576 (1024KiB)

  Built-in kernels                                

  Device Extensions                               cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_image2d_from_buffer cl_khr_gl_depth_images cl_khr_depth_images cl_khr_3d_image_writes 

  Device Name                                     AMD Radeon Pro 560 Compute Engine

  Device Vendor                                   AMD

  Device Vendor ID                                0x1021c00

  Device Version                                  OpenCL 1.2 

  Driver Version                                  1.2 (Nov  9 2017 18:48:40)

  Device OpenCL C Version                         OpenCL C 1.2 

  Device Type                                     GPU

  Device Available                                Yes

  Device Profile                                  FULL_PROFILE

  Max compute units                               16

  Max clock frequency                             907MHz

  Device Partition                                (core)

    Max number of sub-devices                     0

    Supported partition types                     None

  Max work item dimensions                        3

  Max work item sizes                             256x256x256

  Max work group size                             256

  Compiler Available                              Yes

  Linker Available                                Yes

  Preferred work group size multiple              64

  Preferred / native vector sizes                 

    char                                                 4 / 4       

    short                                                2 / 2       

    int                                                  1 / 1       

    long                                                 1 / 1       

    half                                                 0 / 0        (n/a)

    float                                                1 / 1       

    double                                               1 / 1        (cl_khr_fp64)

  Half-precision Floating-point support           (n/a)

  Single-precision Floating-point support         (core)

    Denormals                                     No

    Infinity and NANs                             Yes

    Round to nearest                              Yes

    Round to zero                                 Yes

    Round to infinity                             Yes

    IEEE754-2008 fused multiply-add               No

    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                                    32, Little-Endian

  Global memory size                              4294967296 (4GiB)

  Error Correction support                        No

  Max memory allocation                           1073741824 (1024MiB)

  Unified memory for Host and Device              No

  Minimum alignment for any data type             128 bytes

  Alignment of base address                       32768 bits (4096 bytes)

  Global Memory cache type                        None

  Image support                                   Yes

    Max number of samplers per kernel             16

    Max size for 1D images from buffer            65536 pixels

    Max 1D or 2D image array size                 2048 images

    Base address alignment for 2D image buffers   256 bytes

    Pitch alignment for 2D image buffers          256 pixels

    Max 2D image size                             16384x16384 pixels

    Max 3D image size                             2048x2048x2048 pixels

    Max number of read image args                 128

    Max number of write image args                8

  Local memory type                               Local

  Local memory size                               32768 (32KiB)

  Max constant buffer size                        65536 (64KiB)

  Max number of constant args                     8

  Max size of kernel argument                     1024

  Queue properties                                

    Out-of-order execution                        No

    Profiling                                     Yes

  Prefer user sync for interop                    Yes

  Profiling timer resolution                      40ns

  Execution capabilities                          

    Run OpenCL kernels                            Yes

    Run native kernels                            No

  printf() buffer size                            134217728 (128MiB)

  Built-in kernels                                

  Device Extensions                               cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_image2d_from_buffer cl_khr_depth_images cl_APPLE_command_queue_priority cl_APPLE_command_queue_select_compute_units cl_khr_fp64

  Device Name                                     GeForce GTX 1080 Ti

  Device Vendor                                   NVIDIA

  Device Vendor ID                                0x1022700

  Device Version                                  OpenCL 1.2 

  Driver Version                                  10.29.10 378.10.10.10.25.102

  Device OpenCL C Version                         OpenCL C 1.2 

  Device Type                                     GPU

  Device Available                                Yes

  Device Profile                                  FULL_PROFILE

  Max compute units                               28

  Max clock frequency                             1316MHz

  Device Partition                                (core)

    Max number of sub-devices                     0

    Supported partition types                     None

  Max work item dimensions                        3

  Max work item sizes                             1024x1024x64

  Max work group size                             1024

  Compiler Available                              Yes

  Linker Available                                Yes

  Preferred work group size multiple              32

  Preferred / native vector sizes                 

    char                                                 1 / 1       

    short                                                1 / 1       

    int                                                  1 / 1       

    long                                                 1 / 1       

    half                                                 0 / 0        (n/a)

    float                                                1 / 1       

    double                                               1 / 1        (cl_khr_fp64)

  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                                 Yes

    Round to infinity                             Yes

    IEEE754-2008 fused multiply-add               No

    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                              11811160064 (11GiB)

  Error Correction support                        No

  Max memory allocation                           2952790016 (2.75GiB)

  Unified memory for Host and Device              No

  Minimum alignment for any data type             128 bytes

  Alignment of base address                       1024 bits (128 bytes)

  Global Memory cache type                        None

  Image support                                   Yes

    Max number of samplers per kernel             32

    Max size for 1D images from buffer            134217728 pixels

    Max 1D or 2D image array size                 2048 images

    Base address alignment for 2D image buffers   0 bytes

    Pitch alignment for 2D image buffers          32 pixels

    Max 2D image size                             32768x32768 pixels

    Max 3D image size                             16384x16384x2048 pixels

    Max number of read image args                 256

    Max number of write image args                16

  Local memory type                               Local

  Local memory size                               49152 (48KiB)

  Max constant buffer size                        65536 (64KiB)

  Max number of constant args                     8

  Max size of kernel argument                     4352 (4.25KiB)

  Queue properties                                

    Out-of-order execution                        No

    Profiling                                     Yes

  Prefer user sync for interop                    Yes

  Profiling timer resolution                      1000ns

  Execution capabilities                          

    Run OpenCL kernels                            Yes

    Run native kernels                            No

  printf() buffer size                            134217728 (128MiB)

  Built-in kernels                                

  Device Extensions                               cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_byte_addressable_store cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_APPLE_fp64_basic_ops cl_khr_fp64 cl_khr_3d_image_writes cl_khr_depth_images cl_khr_gl_depth_images cl_khr_gl_msaa_sharing cl_khr_image2d_from_buffer cl_APPLE_ycbcr_422 cl_APPLE_rgb_422 

NULL platform behavior

  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  Apple

  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   Success [P0]

  clCreateContext(NULL, ...) [default]            Success [P0]

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (3)

    Platform Name                                 Apple

    Device Name                                   Intel(R) HD Graphics 630

    Device Name                                   AMD Radeon Pro 560 Compute Engine

    Device Name                                   GeForce GTX 1080 Ti

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  Success (1)

    Platform Name                                 Apple

    Device Name                                   Intel(R) Core(TM) i7-7920HQ CPU @ 3.10GHz

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  Success (3)

    Platform Name                                 Apple

    Device Name                                   Intel(R) HD Graphics 630

    Device Name                                   AMD Radeon Pro 560 Compute Engine

    Device Name                                   GeForce GTX 1080 Ti

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  <checkNullCtxFromType:2444: create context from type CL_DEVICE_TYPE_CUSTOM : error -30>

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  Success (4)

    Platform Name                                 Apple

    Device Name                                   Intel(R) HD Graphics 630

    Device Name                                   AMD Radeon Pro 560 Compute Engine

    Device Name                                   GeForce GTX 1080 Ti

    Device Name                                   Intel(R) Core(TM) i7-7920HQ CPU @ 3.10GHz

NOTE:your OpenCL library only supports OpenCL 1.0,

but some installed platforms support OpenCL 1.2.

Programs using 1.2 features may crash

or behave unexepectedly

logout

Dang! Thanks, now I have to figure out why my clinfo does not detect my eGPU!

This is what I get:

% clinfo
Number of platforms 1
Platform Name Apple
Platform Vendor Apple
Platform Version OpenCL 1.2 (Oct 31 2017 18:19:43)
Platform Profile FULL_PROFILE
Platform Extensions cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event

Platform Name Apple
Number of devices 2
Device Name Intel(R) Core(TM) i7-2720QM CPU @ 2.20GHz
Device Vendor Intel
Device Vendor ID 0xffffffff
Device Version OpenCL 1.2
Driver Version 1.1
Device OpenCL C Version OpenCL C 1.2
Device Type CPU
Device Available Yes
Device Profile FULL_PROFILE
Max compute units 8
Max clock frequency 2200MHz
Device Partition (core)
Max number of sub-devices 0
Supported partition types None
Max work item dimensions 3
Max work item sizes 1024x1x1
Max work group size 1024
Compiler Available Yes
Linker Available Yes
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 (cl_khr_fp64)
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 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 17179869184 (16GiB)
Error Correction support No
Max memory allocation 4294967296 (4GiB)
Unified memory for Host and Device Yes
Minimum alignment for any data type 128 bytes
Alignment of base address 1024 bits (128 bytes)
Global Memory cache type Read/Write
Global Memory cache size 64
Global Memory cache line size 6291456 bytes
Image support Yes
Max number of samplers per kernel 16
Max size for 1D images from buffer 65536 pixels
Max 1D or 2D image array size 2048 images
Base address alignment for 2D image buffers 1 bytes
Pitch alignment for 2D image buffers 1 pixels
Max 2D image size 8192x8192 pixels
Max 3D image size 2048x2048x2048 pixels
Max number of read image args 128
Max number of write image args 8
Local memory type Global
Local memory size 32768 (32KiB)
Max constant buffer size 65536 (64KiB)
Max number of constant args 8
Max size of kernel argument 4096 (4KiB)
Queue properties
Out-of-order execution No
Profiling Yes
Prefer user sync for interop Yes
Profiling timer resolution 1ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels Yes
printf() buffer size 1048576 (1024KiB)
Built-in kernels
Device Extensions cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_image2d_from_buffer cl_APPLE_fp64_basic_ops cl_APPLE_fixed_alpha_channel_orders cl_APPLE_biased_fixed_point_image_formats cl_APPLE_command_queue_priority

Device Name ATI Radeon HD 6750M
Device Vendor AMD
Device Vendor ID 0x1021b00
Device Version OpenCL 1.2
Driver Version 1.2 (Nov 9 2017 18:48:37)
Device OpenCL C Version OpenCL C 1.2
Device Type GPU
Device Available Yes
Device Profile FULL_PROFILE
Max compute units 6
Max clock frequency 600MHz
Device Partition (core)
Max number of sub-devices 0
Supported partition types None
Max work item dimensions 3
Max work item sizes 256x256x256
Max work group size 256
Compiler Available Yes
Linker Available Yes
Preferred work group size multiple 64
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 0 / 0 (n/a)
Half-precision Floating-point support (n/a)
Single-precision Floating-point support (core)
Denormals No
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
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 No
Minimum alignment for any data type 128 bytes
Alignment of base address 32768 bits (4096 bytes)
Global Memory cache type None
Image support Yes
Max number of samplers per kernel 16
Max size for 1D images from buffer 65536 pixels
Max 1D or 2D image array size 2048 images
Base address alignment for 2D image buffers 256 bytes
Pitch alignment for 2D image buffers 256 pixels
Max 2D image size 16384x16384 pixels
Max 3D image size 2048x2048x2048 pixels
Max number of read image args 128
Max number of write image args 8
Local memory type Local
Local memory size 32768 (32KiB)
Max constant buffer size 65536 (64KiB)
Max number of constant args 8
Max size of kernel argument 1024
Queue properties
Out-of-order execution No
Profiling Yes
Prefer user sync for interop Yes
Profiling timer resolution 37ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
printf() buffer size 134217728 (128MiB)
Built-in kernels
Device Extensions cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_image2d_from_buffer cl_khr_depth_images

NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) Apple
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [P0]
clCreateContext(NULL, ...) [default] Success [P0]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) Success (1)
Platform Name Apple
Device Name ATI Radeon HD 6750M
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) Success (1)
Platform Name Apple
Device Name Intel(R) Core(TM) i7-2720QM CPU @ 2.20GHz
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) Success (1)
Platform Name Apple
Device Name ATI Radeon HD 6750M
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) <checkNullCtxFromType:2444: create context from type CL_DEVICE_TYPE_CUSTOM : error -30>
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) Success (2)
Platform Name Apple
Device Name ATI Radeon HD 6750M
Device Name Intel(R) Core(TM) i7-2720QM CPU @ 2.20GHz
NOTE: your OpenCL library only supports OpenCL 1.0,
but some installed platforms support OpenCL 1.2.
Programs using 1.2 features may crash
or behave unexepectedly

Posted by: j.navarro

Posted by: ricosuave0922

Posted by: j.navarro

So let me ask it differently: has anyone managed to run OpenCL on an eGPU?

I mean from a Mac over Thunderbolt.

And, if so, what does clinfo say? 

If you can give me some details on how/what to use to test OpenCL I'd be more than happy to play around and see if i can get things working reliably.

That would be great, @ricosuave0922, thank you. clinfo can be obtained from

https://github.com/Oblomov/clinfo/archive/master.zip

and it compiles and runs right away for me.

Looks like ive got the same result as @mbnwa. How do you have your monitors configured? Are you using headless adapter to accelerate internal screen or are you using external monitor? Which monitor is set as your default? etc.

Currently have installed:

OSX 10.13.2

Nvidia Driver 378.10.10.10.25.102

CUDA 387.99

Utilizing a Mantiz MZ-02

Acer X34 connected to the 1080TI in addition set as primary / default monitor for OSX

@ricosuave0922: Here is my setup:

• early 2011 17″ MacBook-Pro HD6750M
• MacOS 10.13.2
• Akitio Thunder3 with modified power supply for the GTX 960
• Nvidia GTX 960
• HDMI display

I have run the setup as per the top of the page, and both monitors are shown as running off the GTX960, as in the screenshots above. However, everything including iStat Menus seems to indicate that the internal GPU is the one running the show, not the GTX 960.

Not sure what you mean by "default monitor", but this is what I did:

• power on with the eGPU on
• it is detected and it shows in Settings/Displays
• move the menu bar from the laptop window to the external display
• option-move the laptop window to the external display (with menu bar), which causes mirroring
• and this is where it automatically is at when I reboot.

I don't have a headless display but I could try different combinations without the display etc. Not sure what would be most telling.

What I get is this, just two devices:

% clinfo –list
Platform #0: Apple
+– Device #0: Intel(R) Core(TM) i7-2720QM CPU @ 2.20GHz
`– Device #1: ATI Radeon HD 6750M

Please see the email I sent you outside this thread

Platform #0: Apple
+-- Device #0: Intel(R) Core(TM) i7-7920HQ CPU @ 3.10GHz
+-- Device #1: Intel(R) HD Graphics 630
+-- Device #2: AMD Radeon Pro 560 Compute Engine
`-- Device #3: GeForce GTX 1080 Ti

@j.navarro do you have automatic graphics switching checked in Energy Saver? If so uncheck it and try again. Also, is there a reason you are using an external display and mirroring? why not use them as separate displays?

@ricosuave0922 For some reason I did have automatic graphics switching on now, even if I had disabled it earlier. But I had just tried an nvram clear and a bunch of things that forced me to repeat the procedure in this blog from scratch. Now I am back to where I started, with automatic switching off and everything appearing ok but clinfo still not showing the GTX960 eGPU.

What led to this is that I didn't have CUDA installed and thought that maybe it would help. Normally I run OpenCL on the integrated GPU and it works fine, so I now thought that installing CUDA 387.99 perhaps helped Apple OpenCL talk with the Nvidia GPU... It appears like there was no luck.

However, I notice something strange in the console: an endless repetition of this error message for WindowServer (SkyLight) in subsystem com.apple.SkyLight.

           default  WindowServer No matching context for device (0x7f8c47229e10) - disabling acceleration.

Does it ring any bells?

@ricosuave0922: I only had mirroring on because it seemed to be indicated in these instructions, but now it is off. No difference to clinfo, unfortunately.

Posted by: j.navarro

Posted by: hkoster1

Well, that depends on the OpenCL platforms made available by the various "vendors".

On my 2017 15-inch MacBook Pro QC i7 the only available platform is by Apple, counting 4 devices:

- the Intel QC i7 CPU

- the integrated Intel HD Graphics 630

- the additional AMD Radeon Pro 630

- and the eGPU NVidia GeForce GTX 1080Ti

This info comes from running the OpenCLInfo example from the OpenCL Programming Guide. So, it appears that there are no separate platforms by AMD and NVidia.

Thank you, @hkoster1. Could I bother you by asking what eGPU enclosure you are using, and what your config is like (MacOS, nvidia drivers, eGPU setup)? I am trying to figure out if my problem is one of configuration (it probably is) or if it is my older MBP (doubtful). I really want to get to the bottom of this... and get cracking with OpenCL by getting a GTX 1080 Ti too, but I won't buy it until it works with the GTX 960.

I'm using the Sonnet eGFX 550 enclosure, a Dell P2415Q DisplayPort monitor, and further all the latest updates of macOS and the procedure on page 1 of this thread. I hotplug the Sonnet box when logged in as my user self (System Report then recognizes some undefined NVidia object), then I log out and after a momentary black screen the external monitor comes to life. I make that screen the primary one by dragging the menu bar to it, and then I just close the lid of my MacBook Pro (which shuts off the internal screen).

Unigine Heaven or Valley show excellent performance, but my favourite Trainz game (T:ANE) refuses to utilize the eGPU, and instead reverts to the Intel integrated HD 630 graphics (judging by the frame rates) -- why I don't know.

My OpenCLInfo (showing the 4 devices) is the example from The OpenCL Programming Guide). I'll try and run one of the other examples exclusively off the GTX 1080Ti, but it's been awhile since I last looked at OpenCL... 

Hello,

I can report some success (thank you all!):

MBP 2016 15'' Radeon Pro 460 [macOS 10.13.2 (17C88)]
Sonnet eGFX Breakaway Box 550
EVGA SC2 1080ti

My steps:

Install Graphics Driver (WebDriver-378.10.10.10.25.102.pkg)
Disable SIP (only kext)
Install NVDAEGPUSupport-v4.pkg
Reboot
[Steps Below]
Install Cuda Toolkit 9.1.85
Install Cuda Driver 387.99

It works when

-I log in without the GPU connected, then
-connect (NVIDIA Chipset Model detected),
-log out and in again.

If I start up with the eGPU connected then the internal display stays dark after logging in, a then connected external display reveals serious lagginess that only can be resolved by following my steps above (Shutdown, unplug, boot, log in, plug in, log out, log in). Hot unplug leads to total system crash (would be nice if it wouldn't do that anymore).

Blender works fine even on the internal display (native speed), games only work on the external display (Unigine Heaven and Valley as well).
But something that doesn't work and I have no idea how to fix is Wolfram's Mathematica Neural Network Training Function.

IN: (NetTrain[..., TragetDevice->"GPU"]
OUT: NetTrain::badtrgdev: TargetDevice -> GPU could not be used, please ensure that you have a compatible NVIDIA graphics card and have installed the latest operating system drivers.

{"GPU",1}, {"GPU",2} etc. won't help either.

Furthermore:

Needs["CUDALink`"];
Needs["CCompilerDriver`"];
CUDAResourcesInstall[];

returns fine but
CUDAQ[] then crashes the Mathematica Kernel, so does SystemInformation[].

Any ideas on how to fix this? (EDIT!)

Gaming is not my priority (though it sad to see that the performance of my 1080ti is diminished so much by macOS, I'll post my Heaven and Valley benchmarks soon), getting CUDA to work is. NN-Taining without CUDA these days is horrible...

-------- (EDIT) -----------

I found a patch to get it to work with Mathematica:

- Start Mathematica
- Open a new Document
- Needs["CUDALink`"]
- CUDAQ[] should now return true
- Open Document you want to execute

Macbook Pro Retina 15inch. Early 2013.
Aorus Box GTX 1080
High Sierra 10.13.2 

Not working unfortunately. EGPU not detected. Did all steps 100% correct.

MBP 2016 15” Radeon Pro 460 [macOS 10.13.2 (17C88)]
Sonnet eGFX Breakaway Box 550
EVGA SC2 1080ti

Unigine Heaven 4.0
OpenGL, 1920x1080, 8xAA, Fullscreen, Ultra, Tessellation Extreme:
102.7 fps (11.2, 199.5), 2586 Score

Unigine Valley 1.0
OpenGL, 1920x1080, 8xAA, Fullscreen, Extreme HD Preset:
80.7 fps (27.6, 114.6), 3375 Score

Cycles Benchmark, Blender 2.79:
BMW: 2:15.66 min
Pavillion: 5:33.22 min

I've also encountered a problem: 4K@60Hz does not work.

I'm having trouble with a setup that was working fine up until right before I upgraded to High Sierra 10.13.2:

* Mid-2012 MacBook Pro (Nvidia GT 650 GPU)
* High Sierra 10.13.2 (17C88)
* Aorus GTX 1080 Gaming Box
* Apple TB2->TB3 cable
* Two Dell 2415Q displays

I've installed the correct versions of the Nvidia Web Driver and the NVDAEGPUSupportDriver. When I connect the Aorus box, its fans start spinning, but I don't see anything on the external displays. 'System Information' correctly reports the USB devices that are connected to the Aorus box, and also correctly lists 'GV-N1080IXEB-8GD' in the Thunderbolt Device Tree. However, when I try to list the installed 'Graphics/Displays', 'System Information' becomes unresponsive, and does not even list the built in GPU's.

If I logout of the system with the Aorus box still connected, the MBP's main LCD display goes black, but nothing appears on the external displays. When I then disconnect the Aorus box, the LCD display comes back reporting a kernel panic.

One thing I've noticed, which I never bothered to observe before, is that the system never remembers my attempts to disable 'Automatic graphics switching'; I can turn it off, but when the system reboots, the setting is back on again.

Any help with this is much appreciated. Like I said, this setup worked last week when I was still using High Sierra 10.13.1

And lo, now the thing works again, after a week of trying. Here's what was different this time:

* I re-enabled SIP with 'csrutil enable --without kext'
* I connected the Aorus box to the MacBook Pro before powering on the computer
* I disconnected the USB devices from the Aorus box while it booted (those devices are a Das keyboard which has a dongle for a Logitech wireless mouse connected to it).

No idea if or how any of that made it work, but I'm glad it did

This is the first time I make some progress. Now on 10.13.2 with the latest v.4 I can get the GPU at least partly to be recognized.

However, the (internal) screen turns black. When having an external screen connected I somehow managed to click the shutdown button - but cancel out the window where you are asked - do you really want to shut down. Then like magic the internal display turned on again.

How can I use the internal display, i.e. have it turned on again (only interested in CUDA) when no external monitor is plugged in to press the shut down button?

Need Help...

Specs:

2017 MBP, running OSX 10.13.2.

Aorus Gaming box 1080 via tb3

Cinema4d/ Octane Render V3

So First. Thank you for any suggestions you can help me with at this point,  

Ive installed Goalque's script and followed the instructions. Great, computer recognizing ...but as NVIDIA chipset. Ive downloaded the latest CUDA drivers and it registers having them, but no applications seem to, (adobe, cinema4d, Octane Render, etc) So I should have followed your initial instructions: not to run Rastafabi's script with Goalques....and alas, I have done just that. I did the install without the Box attached. Then the computer rebooted and wouldnt boot up. Now it only seems to start in Safe mode or recovery. 1 thing I noticed is that after I did open the computer in Safe mode, I saw that my laptop read the nvidia 1080 chip, no longer just chipset. 

So, Im a bit lost with all of this tbh. Any help yall can give would be greatly appreciated. I was able to boot the computer up finally, since I had a 10.13.2 update still and just installed it again. The computer still has issues at the initial boot. I have deleted the NVAEDAE extension, Im just a bit nervous to restart my computer now...Help?! please. 

@dmiller, Goalques script does not work for high Sierra - use the instructions on page 1 of this posting.

Hi guys,

I just got the AORUS gaming box with GTX 1080 today and tried the steps in the first page of this thread but I cannot seem to get the egpu to work.

I have a 2017 13" Macbook Pro with Touchbar and I'm connecting it to the AORUS gaming box with GTX 1080. 

When I go into system report and go to graphics card, it only shows the built in Iris Plus 650, the egpu isn't listed at all under the graphics card section. But it is recognised in the Thunderbolt section though.

Not sure what I did incorrectly, can someone please help and/or point me in the right direction?

Thank you

Was able to get the system to recognise the egpu under graphics card but it's not actually being utilised in any programs, nor selected in the display menu.

I'm using a LG ultrafine display connected to the Macbook, while the egpu is also connected to the Macbook. I know it's not ideal but I've seen it work from other people's set ups.

How do I get the computer to use the egpu instead of the internal graphics chip?

Posted by: Bryan Lee

I'm using a LG ultrafine display connected to the Macbook, while the egpu is also connected to the Macbook. I know it's not ideal but I've seen it work from other people's set ups.

 

Why aren't you connecting the screen to the eGPU?

Because he is using an LG UltraFine which only connects via TB3 and no GPU's come with TB3 connection.

@bryan_lee I have the same, never managed to accelerate the display, accelerated internal using a headless adapter connected to GPU.

Ah, sorry, expected the Ultrafine to have the same number of connectors like the 4K LGs with USB-C.

It actually doesn't do DP over TB3 (so one could use a converter cable)?

EDIT: just checked, it doesnt.

Sadly not... its not even USB 3 its clever TB 3 combining 2 streams into one cable... the 4k Ultrafine is USB C but haven't been able to test as I have the 5k.

This gentleman seemed to have gotten it to work by routing from the egpu, through the Macbook, and to the LG display. Any clue as how his setup was successful? His post says it was pretty much plug and play, I've sent him a PM so hopefully he can enlighten us. 

Were you able to get your apps to recognize the eGPU? Im having the same issue. Cinema4d and Octane doesnt recognize it. Ae doesnt, Cuda seems to work, on the display, but not for the programs. Did you find a work around?

Posted by: Bryan Lee

Hi guys,

I just got the AORUS gaming box with GTX 1080 today and tried the steps in the first page of this thread but I cannot seem to get the egpu to work.

I have a 2017 13" Macbook Pro with Touchbar and I'm connecting it to the AORUS gaming box with GTX 1080. 

When I go into system report and go to graphics card, it only shows the built in Iris Plus 650, the egpu isn't listed at all under the graphics card section. But it is recognised in the Thunderbolt section though.

Not sure what I did incorrectly, can someone please help and/or point me in the right direction?

Thank you

Bryan, I have exactly the same setup as yours. I initially tried to stay on Sierra (10.12) and run the egpu automate script, but while the Gaming Box was recognized and an external GPU was also recognized, the extrernal display was not. So, I decided to take the plunge and upgrade to High Sierra (10.13.2) and follow the instructions in the beginning of this thread.

I think I actually faced the same problem as you - the Gaming Box was recognized, but the GPU was not. Here's what helped me - after following all the steps in the first post, I rebooted to Revocery Mode > opened Terminal > typed: csrutil disable > restarted. After the restart, the GPU was recognized, and it's also working - was able to run some 4K games.

Of course, the issue there is that SIP is still disabled. I have not yet tried re-enabling it to see if maybe that was just a one-time requirement and I don't need to leave it disabled...

With that said, there are some issues:

• Hot unplugging crashes the Mac
• Clamshell mode does not work - if I close the laptop lid, the internal screen is turned off, but it's still attached as a desktop - so if you have some apps on your internal screen desktop, they will still stay there, and you can accidentally "lose" your mouse pointer by moving it to the internal screen, which you can't see anymore. I am wondering if anybody else is also experiencing this and whether this is an issue with AMD cards as well...
• Putting the Mac to sleep does not put the Aorus Gaming Box to sleep - it just keeps on running and spinning the vents... which is pretty annoying considering that the only way to turn it off is to disconnect the TB3 cable from Mac, which will crash it...
• When waking the Mac from sleep, the external display is not instantly recognized. I have found that when I press a key to wake the map from sleep, I will have to click Cancel or press Esc and the press any key again - only then is the external display recognized and turned on.
• After shutting the Mac down and starting it up again, the external display isn't recognized - I need to unplug the DP cable and plug it back in for it to be recognized.
• HDMI does not seem to work - the display is kind of recognized, but the screen just stays black
• Audio over DP to monitor is not supported (I’ve read it works fine on Windows/bootcamp)

If anybody knows fixes to any of these issues, I'm all ears. My biggest annoyances are the hot unplug crash and Gaming Box not shutting off going to sleep when the Mac is.

UUuugh.  So close!

MBP 13" (2016 - four thunderbolt 3 ports), 10.13.2, using 4035-nvidia-egpu-v4 installer.

Using Mantiz Venus with a GTX 1080 stock ref. card from NVIDIA.

System identifies the secondary display, but its black and only displays mouse.  Tried several recommendations having egpu not plugged in while starting up, then plugging in and logging in, etc.  

Have you tried disabling SIP ? It worked for me.