the HW is : A10-7400P + Radeon R9 m270X
the OS is :Windows 8.1 64bit
I have downloaded the opencl 2.0 driver and it says that the driver is compatible with AMD graphics products based on GCN 1.1 Architecture or higher. But i don't know the architecture of my graphics products.
Please help,thanks a lot
The AMD OpenCL 2.0 driver is compatible with AMD graphics products based on GCN 1.1 Architecture or higher.
The m290X is built on GCN 1.1.
The graphics cores in the A10-7400 P are as well.
So the answer is yes.
Perhaps an excess of caution, but keep in mind that while your hardware supports this technology, that doesn't mean there are a lot of apps out there that use it. OpenCL 2.0 is still pretty new. However, any app that does, will work effectively and likely perform better on your hardware than it would on hardware that only supports OpenCL 1.2.
By the way, I have NOT added you to the DevGuru's white list. This is not really a developer question. Happy to answer, this is a good place.
Thank you for your reply.
I am sorry that i didn't explain my question clearly.
I want to use SVM in my OpenCL code which is supported by OpenCL 2.0.
So i must confirm that my devices if support OpenCL 2.0 and SVM or not.
I did some test as follows:
step1:
cout<<"CL_DEVICE_SVM_COARSE_GRAIN_BUFFER: "<<CL_DEVICE_SVM_COARSE_GRAIN_BUFFER<<endl;
cout<<"CL_DEVICE_SVM_FINE_GRAIN_BUFFER: "<<CL_DEVICE_SVM_FINE_GRAIN_BUFFER<<endl;
cout<<"CL_DEVICE_SVM_FINE_GRAIN_SYSTEM: "<<CL_DEVICE_SVM_FINE_GRAIN_SYSTEM<<endl;
cout<<"CL_DEVICE_SVM_ATOMICS: "<<CL_DEVICE_SVM_ATOMICS<<endl;
cl_device_svm_capabilities *cap1,*cap2,*cap3;
size_t cap1_size,cap2_size,cap3_size;
status=clGetDeviceInfo(devices[0],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap1_size);
status=clGetDeviceInfo(devices[1],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap2_size);
status=clGetDeviceInfo(devices[2],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap3_size);
cap1=(cl_device_svm_capabilities *)malloc(cap1_size);
cap2=(cl_device_svm_capabilities *)malloc(cap2_size);
cap3=(cl_device_svm_capabilities *)malloc(cap3_size);
status=clGetDeviceInfo(devices[0],CL_DEVICE_SVM_CAPABILITIES,cap1_size,cap1,NULL);
cout<<"cap1:"<<*cap1<<endl;
status=clGetDeviceInfo(devices[1],CL_DEVICE_SVM_CAPABILITIES,cap2_size,cap2,NULL);
cout<<"cap2:"<<*cap2<<endl;
status=clGetDeviceInfo(devices[2],CL_DEVICE_SVM_CAPABILITIES,cap3_size,cap3,NULL);
cout<<"cap3:"<<*cap3<<endl;
The result:
CL_DEVICE_SVM_COARSE_GRAIN_BUFFER: 1
CL_DEVICE_SVM_FINE_GRAIN_BUFFER: 2
CL_DEVICE_SVM_FINE_GRAIN_SYSTEM:4
CL_DEVICE_SVM_ATOMICS: 8
cap1:0
cap2:0
cap3:15
PS: devices[0] (name):Spectre, devices[1](name):Bonaire, devices[2](name):AMD A10-7400p Radeon R6, 10 Computer Cores 4C+6G
step2:
bufferSVM=clSVMAlloc(context,CL_MEM_READ_WRITE,datasize,0); |
if(bufferSVM==NULL)
cout<<"clSVMAlloc() failure!\n"; |
status=clEnqueueSVMMap(cmdQueue1,CL_TRUE,CL_MAP_READ,bufferSVM,datasize,0,NULL,NULL);
if(status!=CL_SUCCESS){
cout<<"clEnqueueSVMMap is failure!"<<endl; | |
exit(1); |
}
The clEnqueueSVMMap function return CL_OUT_OF_RESOURCES.
Thanks for the clarification! I will whitelist you and move this into the OpenCL forum. We are clearly NOT talking to an end-user buying a laptop. Welcome to the community!
Hi,
You can check OpenCL version supported by the device using "clinfo" command. It lists all the information about each of the detected devices.
Please make sure you're using a device that supports SVM capabilities, otherwise SVM functions will fail. Could you please share your "clinfo" output?
Regards,
hi,
Thank you for your attention.
I take two kinds of ways to check OpenCL version:
(1)I excute the "clinfo.exe" located in AMDAPPSDKROOT\bin\x86_64\ and get the following result.
******************************************result1************************************************************
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 654Mhz
Address bits: 64
Max memory allocation: 602406912
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2146959360
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 602406912
Max global variable size: 542166016
Max global variable preferred total size: 2146959360
Max read/write image args: 64
Max on device events: 1024
Queue on device max size: 524288
Max on device queues: 1
Queue on device preferred size: 16384
SVM capabilities:
Coarse grain buffer: Yes
Fine grain buffer: Yes
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: Yes
Profiling : Yes
Platform ID: 00007FFC9C8F5B60
Name: Spectre
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 2.0
Driver version: 1642.5 (VM)
Profile: FULL_PROFILE
Version: OpenCL 2.0 AMD-APP (1642.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd
vec3 clamd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d1
0_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buff
er cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images
Device Type: CL_DEVICE_TYPE_GPU
Vendor ID: 1002h
Board name: AMD Radeon R9 M270X
Device Topology: PCI[ B#1, D#0, F#0 ]
Max compute units: 12
Max work items dimensions: 3
Max work items[0]: 256
Max work items[1]: 256
Max work items[2]: 256
Max work group size: 256
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 654Mhz
Address bits: 32
Max memory allocation: 1408499712
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 8
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2147483648
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 0
Max pipe active reservations: 0
Max pipe packet size: 0
Max global variable size: 0
Max global variable preferred total size: 0
Max read/write image args: 0
Max on device events: 0
Queue on device max size: 0
Max on device queues: 0
Queue on device preferred size: 0
SVM capabilities:
Coarse grain buffer: No
Fine grain buffer: No
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 0
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: No
Profiling : No
Platform ID: 00007FFC9C8F5B60
Name: Bonaire
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 1.2
Driver version: 1642.5 (VM)
Profile: FULL_PROFILE
Version: OpenCL 1.2 AMD-APP (1642.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd
vec3 clamd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d1
0_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buff
er cl_khr_spir cl_khr_gl_event
Device Type: CL_DEVICE_TYPE_CPU
Vendor ID: 1002h
Board name:
Max compute units: 4
Max work items dimensions: 3
Max work items[0]: 1024
Max work items[1]: 1024
Max work items[2]: 1024
Max work group size: 1024
Preferred vector width char: 16
Preferred vector width short: 8
Preferred vector width int: 4
Preferred vector width long: 2
Preferred vector width float: 8
Preferred vector width double: 4
Native vector width char: 16
Native vector width short: 8
Native vector width int: 4
Native vector width long: 2
Native vector width float: 8
Native vector width double: 4
Max clock frequency: 2495Mhz
Address bits: 64
Max memory allocation: 2147483648
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 8192
Max image 2D height: 8192
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 4096
Alignment (bits) of base address: 1024
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: Yes
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 7448162304
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Global
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 2147483648
Max global variable size: 1879048192
Max global variable preferred total size: 1879048192
Max read/write image args: 64
Max on device events: 0
Queue on device max size: 0
Max on device queues: 0
Queue on device preferred size: 0
SVM capabilities:
Coarse grain buffer: Yes
Fine grain buffer: Yes
Fine grain system: Yes
Atomics: Yes
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 1
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 410
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: Yes
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: No
Profiling : No
Platform ID: 00007FFC9C8F5B60
Name: AMD A10-7400P Radeon R6, 10 Com
pute Cores 4C+6G
Vendor: AuthenticAMD
Device OpenCL C version: OpenCL C 1.2
Driver version: 1642.5 (sse2,avx,fma4)
Profile: FULL_PROFILE
Version: OpenCL 1.2 AMD-APP (1642.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec
3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sh
aring cl_khr_spir cl_khr_gl_event
**********************************************************************************************************
(2)I input clinfo in cmd and get the result 2 as follows.
******************************************result 2********************************************************
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 654Mhz
Address bits: 32
Max memory allocation: 603193344
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 8
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2146959360
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 0
Max pipe active reservations: 0
Max pipe packet size: 0
Max global variable size: 0
Max global variable preferred total size: 0
Max read/write image args: 0
Max on device events: 0
Queue on device max size: 0
Max on device queues: 0
Queue on device preferred size: 0
SVM capabilities:
Coarse grain buffer: No
Fine grain buffer: No
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: No
Profiling : No
Platform ID: 00007FFCB1D03FB0
Name: Spectre
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 1.2
Driver version: 1598.5 (VM)
Profile: FULL_PROFILE
Version: OpenCL 1.2 AMD-APP (1598.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd
vec3 clamd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d1
0_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buff
er cl_khr_spir cl_khr_gl_event
Device Type: CL_DEVICE_TYPE_GPU
Vendor ID: 1002h
Board name: AMD Radeon R9 M270X
Device Topology: PCI[ B#1, D#0, F#0 ]
Max compute units: 12
Max work items dimensions: 3
Max work items[0]: 256
Max work items[1]: 256
Max work items[2]: 256
Max work group size: 256
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 654Mhz
Address bits: 32
Max memory allocation: 1408499712
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 8
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2147483648
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 0
Max pipe active reservations: 0
Max pipe packet size: 0
Max global variable size: 0
Max global variable preferred total size: 0
Max read/write image args: 0
Max on device events: 0
Queue on device max size: 0
Max on device queues: 0
Queue on device preferred size: 0
SVM capabilities:
Coarse grain buffer: No
Fine grain buffer: No
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 0
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: No
Profiling : No
Platform ID: 00007FFCB1D03FB0
Name: Bonaire
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 1.2
Driver version: 1598.5 (VM)
Profile: FULL_PROFILE
Version: OpenCL 1.2 AMD-APP (1598.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd
vec3 clamd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d1
0_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buff
er cl_khr_spir cl_khr_gl_event
Device Type: CL_DEVICE_TYPE_CPU
Vendor ID: 1002h
Board name:
Max compute units: 4
Max work items dimensions: 3
Max work items[0]: 1024
Max work items[1]: 1024
Max work items[2]: 1024
Max work group size: 1024
Preferred vector width char: 16
Preferred vector width short: 8
Preferred vector width int: 4
Preferred vector width long: 2
Preferred vector width float: 8
Preferred vector width double: 4
Native vector width char: 16
Native vector width short: 8
Native vector width int: 4
Native vector width long: 2
Native vector width float: 8
Native vector width double: 4
Max clock frequency: 2495Mhz
Address bits: 64
Max memory allocation: 2147483648
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 8192
Max image 2D height: 8192
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 4096
Alignment (bits) of base address: 1024
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: Yes
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 7448162304
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Global
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 2147483648
Max global variable size: 1879048192
Max global variable preferred total size: 1879048192
Max read/write image args: 64
Max on device events: 0
Queue on device max size: 0
Max on device queues: 0
Queue on device preferred size: 0
SVM capabilities:
Coarse grain buffer: Yes
Fine grain buffer: Yes
Fine grain system: Yes
Atomics: Yes
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 1
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 410
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: Yes
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: No
Profiling : No
Platform ID: 00007FFCB1D03FB0
Name: AMD A10-7400P Radeon R6, 10 Com
pute Cores 4C+6G
Vendor: AuthenticAMD
Device OpenCL C version: OpenCL C 1.2
Driver version: 1598.5 (sse2,avx,fma4)
Profile: FULL_PROFILE
Version: OpenCL 1.2 AMD-APP (1598.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec
3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sh
aring cl_khr_spir cl_khr_gl_event
***********************************************************************************************************
I have 2 questions.
(1) I am sorry that i don't know how to get the full output in dialog box.
(2) Why are the two results different ? Which one should i beileve?
Look forward to your reply.
Regards
1) Open a command prompt window and type: "clinfo > <file_name>.txt" ---> it will redirect the clinfo output to the file named <file_name>.txt under current directory.
2) clinfo and some other libraries come with catalyst driver as well as APP SDK. For OpenCL development targeted to CPU only m/c, one can use the APP SDK without installing the catalyst driver. But, if one wants to access the GPUs, he/she must install the appropriate catalyst driver. Usually as driver is updated more frequently than APP SDK, it is expected that a more recent version of clinfo and other packages would come with catalyst driver. That's why, driver version packages are set as default one. If possible, its always better to have the latest driver installed on the m/c.
In your case, it seems that the driver is little bit older (please check the driver version for both the cases). I would suggest you to install the latest catalyst driver 15.7 available right now and then check the clinfo output as well as your program.
Regards,
hi,
I downloaded the 15.7 driver and installed successfully.
After using clinfo command,I get output as follows:
**************************************************************************
Number of platforms: | 1 | ||
Platform Profile: | FULL_PROFILE | ||
Platform Version: | OpenCL 2.0 AMD-APP (1800.5) | ||
Platform Name: | AMD Accelerated Parallel Processing | ||
Platform Vendor: | Advanced Micro Devices, Inc. | ||
Platform Extensions: | cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices |
Platform Name: AMD Accelerated Parallel Processing
Number of devices: 3
Device Type: CL_DEVICE_TYPE_GPU
Vendor ID: 1002h
Board name: AMD Radeon(TM) R6 Graphics
Device Topology: PCI[ B#0, D#1, F#0 ]
Max compute units: 6
Max work items dimensions: 3
Max work items[0]: 256
Max work items[1]: 256
Max work items[2]: 256
Max work group size: 256
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 654Mhz
Address bits: 64
Max memory allocation: 1207369728
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2683568128
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 1207369728
Max global variable size: 1086632704
Max global variable preferred total size: 2683568128
Max read/write image args: 64
Max on device events: 1024
Queue on device max size: 524288
Max on device queues: 1
Queue on device preferred size: 262144
SVM capabilities:
Coarse grain buffer: Yes
Fine grain buffer: Yes
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: Yes
Profiling : Yes
Platform ID: 00007FFEE2B8F180
Name: Spectre
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 2.0
Driver version: 1800.5 (VM)
Profile: FULL_PROFILE
Version: OpenCL 2.0 AMD-APP (1800.5)
Extensions: | cl_khr_fp64 cl_amd_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_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_khr_gl_depth_images cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images |
Device Type: CL_DEVICE_TYPE_GPU
Vendor ID: 1002h
Board name: AMD Radeon R9 M280X
Device Topology: PCI[ B#1, D#0, F#0 ]
Max compute units: 12
Max work items dimensions: 3
Max work items[0]: 256
Max work items[1]: 256
Max work items[2]: 256
Max work group size: 256
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 654Mhz
Address bits: 64
Max memory allocation: 1409286144
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2147483648
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 1409286144
Max global variable size: 1268357376
Max global variable preferred total size: 2147483648
Max read/write image args: 64
Max on device events: 1024
Queue on device max size: 524288
Max on device queues: 1
Queue on device preferred size: 262144
SVM capabilities:
Coarse grain buffer: Yes
Fine grain buffer: Yes
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 0
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: Yes
Profiling : Yes
Platform ID: 00007FFEE2B8F180
Name: Bonaire
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 2.0
Driver version: 1800.5 (VM)
Profile: FULL_PROFILE
Version: OpenCL 2.0 AMD-APP (1800.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_khr_gl_depth_images cl_ext_atomic_counters_32 cl_amd_device
_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl
_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing c
l_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_de
pth_images
Device Type: CL_DEVICE_TYPE_CPU
Vendor ID: 1002h
Board name:
Max compute units: 4
Max work items dimensions: 3
Max work items[0]: 1024
Max work items[1]: 1024
Max work items[2]: 1024
Max work group size: 1024
Preferred vector width char: 16
Preferred vector width short: 8
Preferred vector width int: 4
Preferred vector width long: 2
Preferred vector width float: 8
Preferred vector width double: 4
Native vector width char: 16
Native vector width short: 8
Native vector width int: 4
Native vector width long: 2
Native vector width float: 8
Native vector width double: 4
Max clock frequency: 2495Mhz
Address bits: 64
Max memory allocation: 2147483648
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 8192
Max image 2D height: 8192
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 4096
Alignment (bits) of base address: 1024
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: Yes
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 7448162304
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Global
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 2147483648
Max global variable size: 1879048192
Max global variable preferred total size: 1879048192
Max read/write image args: 64
Max on device events: 0
Queue on device max size: 0
Max on device queues: 0
Queue on device preferred size: 0
SVM capabilities:
Coarse grain buffer: No
Fine grain buffer: No
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 1
Error correction support: 0
Unified memory for Host and Device: 1
Profiling timer resolution: 410
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: Yes
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: No
Profiling : No
Platform ID: 00007FFEE2B8F180
Name: AMD A10-7400P Radeon R6, 10 Com
pute Cores 4C+6G
Vendor: AuthenticAMD
Device OpenCL C version: OpenCL C 1.2
Driver version: 1800.5 (sse2,avx,fma4)
Profile: FULL_PROFILE
Version: OpenCL 1.2 AMD-APP (1800.5)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_
global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int3
2_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_
khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store
cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec
3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sh
aring cl_khr_spir cl_khr_gl_event
**************************************************************************************************************
Comparing with yesterday's results, GPU R6 and GPU R9 seems to support SVM.
Does it mean that i can use SVM in my OpenCL code?
I am not fammiliar with SVM and the Demo which downloaded from http://developer.amd.com/tools-and-sdks/opencl-zone/amd-accelerated-parallel-processing-app-sdk/open... are little bit compliacted.
Can you show me some simple demos to test SVM?
Many thanks.
Now, you've SVM support for the GPUs and you can use it. Please run the SVM related samples from APP SDK to explore it.
Sure, I'll try to write a simple program to showcase usage of SVM and share with you. However, I think, SVMBinarySearchTree sample is a good one for this purpose. You don't need to understand the code or logic in detail. Just check the SVM related APIs to understand its usage. Also check how kernel code can directly access the pointer data structure which was not possible until OpenCL 2.0. For better understanding, you may refer this blog http://developer.amd.com/community/blog/2014/10/24/opencl-2-shared-virtual-memory/ .
Regards,
Hi,
I followed your advices and debuged the SVMBinarySearchTree. However,i got a error report"Unsupported device! Required CL_DEVICE_OPENCL_C_VERSION 2.0 or higher".
I checked the code and found that when i using clGetDeviceInfo(device[0],CL_DEVICE_VERSION,sizeof(char)*(tempsize),deviceversion,NULL) the API returns "OpenCL 1.2 AMD-APP (1800.5)"
Why is the device version return from clGetDeviceInfo different from the clinfo output?
Looking forward to your reply.
Are you building the program for x86 version?
Because, the program/application needs to be built as x64 version in order to access the OpenCL 2.0 support. Please check and confirm.
Regards,
Hi,
I change the solution platforms to x64 and it works.
Many thanks.
hi,
While the SVMBinarySearchTree debug with no error,I insert some code in my OpenCL project and the clEnqueueSVMMap() still returns CL_OUT_OF_RESOURCES.I feel so confused.
I checked the device version and SVM capabilities using clGetDeviceInfo().
code:
*************************************************************************************************************
cl_device_svm_capabilities *cap1,*cap2,*cap3; |
size_t cap1_size,cap2_size,cap3_size;
status=clGetDeviceInfo(devices[0],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap1_size);
status=clGetDeviceInfo(devices[1],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap2_size);
status=clGetDeviceInfo(devices[2],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap3_size);
cap1=(cl_device_svm_capabilities *)malloc(cap1_size);
cap2=(cl_device_svm_capabilities *)malloc(cap2_size);
cap3=(cl_device_svm_capabilities *)malloc(cap3_size);
status=clGetDeviceInfo(devices[0],CL_DEVICE_SVM_CAPABILITIES,cap1_size,cap1,NULL);
cout<<"cap1:"<<*cap1<<endl;
status=clGetDeviceInfo(devices[1],CL_DEVICE_SVM_CAPABILITIES,cap2_size,cap2,NULL);
cout<<"cap2:"<<*cap2<<endl;
status=clGetDeviceInfo(devices[2],CL_DEVICE_SVM_CAPABILITIES,cap3_size,cap3,NULL);
cout<<"cap3:"<<*cap3<<endl;
char devicename[3][128];
for(int i=0;i<numDevices;i++){
status=clGetDeviceInfo(devices,CL_DEVICE_NAME,128*sizeof(char),devicename,NULL);
if (status != CL_SUCCESS) {
printf("Error: Getting devices Info (device name, clGetDeviceInfo)\n"); | |
} |
printf("device[%d] name:\t%s\n",i,devicename);
}
char * deviceversion[3][128];
for(int i=0;i<numDevices;i++)
{
status=clGetDeviceInfo(devices,CL_DEVICE_VERSION,sizeof(deviceversion),&deviceversion,NULL); | |
printf("device %s 's version: %s\n",devicename,deviceversion); |
}
*********************************************************************************************************************************
output:
*********************************************************************************************************************************
CL_DEVICE_SVM_COARSE_GRAIN_BUFFER: 1
CL_DEVICE_SVM_FINE_GRAIN_BUFFER: 2
CL_DEVICE_SVM_FINE_GRAIN_SYSTEM: 4
CL_DEVICE_SVM_ATOMICS: 8
cap1:3 //return from clGetDeviceInfo(devices[0],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap1_size);
cap2:3 //return from clGetDeviceInfo(devices[1],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap2_size);
cap3:0 //return from clGetDeviceInfo(devices[2],CL_DEVICE_SVM_CAPABILITIES,0,NULL,&cap3_size);
device[0] name: Spectre
device[1] name: Bonaire
device[2] name: AMD A10-7400P Radeon R6, 10 Compute Cores 4C+6G
device Spectre 's version: OpenCL 2.0 AMD-APP (1800.5)
device Bonaire 's version: OpenCL 2.0 AMD-APP (1800.5)
device AMD A10-7400P Radeon R6, 10 Compute Cores 4C+6G 's version: OpenCL 1.2 AM
D-APP (1800.5)
***********************************************************************************************************************************
code1:
***********************************************************************************************************************************
cmdQueue1 = clCreateCommandQueue(
context,
devices[0],
0,
&status);
if(bufferSVM==NULL)
} status=clEnqueueSVMMap(cmdQueue1,CL_TRUE,CL_MAP_WRITE_INVALIDATE_REGION,bufferSVM,datasize,0,NULL,NULL); if(status!=CL_SUCCESS){
} |
************************************************************************************************************************************
The clEnqueueSVMMap() returns CL_OUT_OF_RESOURCES.
Looking forward to your reply.
Regards,
As you've mentioned that the SVMBinarySearchTree sample is running fine, I don't see any reason not to work your own SVM code. The above code portion seems okay to me. Could you please share the complete project so the I can check it once here?
Regards
ok,no problem.
"vecadd.cl" is the kernel.
"vector add.cpp" is the host.
"chapter_2" has the complete project.
If you can't access to the link and download the project, you can send me an email and i will reply with the whole project.
My address:wangfangl@nuaa.edu.cn
Many thanks
I'm unable to download the project from that link. I've sent you an email. Please check.
hi,
I still don't receive your message.Could you post your email address?I will send my project to you.
Thanks a lot
I've sent you a private message.
Yes,i have received your message and sent the project to your email.Please check.
Regards
I haven't received your email yet. Don't know why both of us not getting each others email. May be any firewall issue. If the project is not large, could you please attach the project on that thread itself?
Regards,
Try to create the context with GPU devices only.
hi,
Thank you for your advice.
I create the context with GPU and clEnqueueSVMMap() returns CL_SUCCESS.
But when I build program with options "-cl-std=CL2.0" ,it returns some error.
*************************************************
code:
string buildflag="-cl-std=CL2.0"; |
status = clBuildProgram( | |
program1, | |
numDevices, | |
devices, | |
buildflag.c_str(), | |
NULL, | |
NULL); |
buildlog:
Internal error: Link failed.
Make sure the system setup is correct.
**************************************************
I haven't met this error before.
Look forward to your reply.
Regards
Last time when I was testing the above attached code, I didn't find any such error. Please verify the build settings and code carefully.
Just one point. Do you observe same problem if you select one GPU at a time?
Regards,
I attach the newest project.The old code don't use build options "-cl-std=CL2.0".
I see your reply in OpenCL 2.0 - Shared Virtual Memory Programming and insert some code to add build options.
Aditionally,i use SVM in my code refer to SVMBinarySearchTree .But i don't know if the way is correct or not.
I follow your suggestion and choose one GPU to run my kernel. However, when i debug the project,it failed in clBuildProgram() and the clGetProgramBuildInfo() returns the build log as mentioned above.
If you need know more programming environment information to solve the problem,please tell me.
Thanks for your patience and help.
Regards
PS: If i don't use build options(clBuildProgram( program1,numDevices,devices,NULL,NULL,NULL);),clBuildProgram return CL_SUCCESS.
Please don't use any printf statement inside the kernel code when building for OCL 2.0. AFAIK, kernel compilation for OpenCL 2.0 fails if kernel has any printf statement. Regarding this, I already filed a bug some time ago.
Just to point out:
You've passed two SVM pointers to "vecsvm" kernel (kernel3) though it takes only one argument.
Regards,
Excellent! You are right.I delete printf statement inside the kernel and clBuildProgram() returns CL_SUCCESS.
It's my mistake that pass two SVM pointer to kernel3.Before i use build options of "-cl-std=CL2.0",I try to debug the program which has SVM statement.It returns a error as follows:
Access conflict occurs when 0x00000197000078E0 is read from the 0xC0000005
At that time,kernel3 is same as kernel2.Just using 2 SVM buffer replace the traditional buffers.I thought maybe 2 SVM buffers caused access confliction.So i delete one and just using one argument in kernel3.But i forgot to change the host code.After that ,you know, I found that I should use build options when building for OCL2.0. And met the internel error as mentioned when i add build options.
Now, I have fixed the mistake in host code,just passing one SVM pointer to kernel3. What's more,there is no internel error and clBuildProgram() returns CL_SUCCESS. It should owe to your advices.
However,I met the old access conflict error again when the program running to clEnqueueNDRangeKernel(...,kernel3,..).
Do i use SVM in an incorrect way? Or some other reasons?
You didn't use the build flag "-cl-std=CL2.0" during the kernel compilation.
Few other points:
1) Please check the map operations of buffer D and E. I guess, the buffers should be mapped as int pointer, not as float pointer. You may get different result for that.
2) You've interchanged the datasize and datasize2 variables some places. Though datasize and datasize2 values are expected to be same, still better to use them as per their meaning.
Regards,
I'm sorry for my carelessness.
Thank you for your suggestions.It's really helpful.
@ dipak
I followed the same steps, I installed the latest Catalyst driver and AMD APP SDK 3.0 Beta. My platform is a Windows7-64bit machine with an AMD Radeon HD 8970M. Altough after the installation the platform seems to support OpenCL 2.0, the GPU and CPU do not. I checked also in C:\Windows\System32 folder and the amdocl.dll and OpenCL.dll files have been updated both to OpenCL 2.0. I rebooted the PC, nothing changes. Here the .txt clinfo resume.
Why it does not upgrade to OpenCL 2.0? Landlord has a less powerful !@ card (R9M280X), so I thought on mine OpenCL 2.0 was supported without problems.
Regards,
Marco
I think you need install the OpeCL 2.0 Driver AMD OpenCL™ 2.0 Driver .
I find out that device opencl c version is "OpenCL 1.2" in clinfo output.
However,the AMD OpenCL™ 2.0 driver is designed to support Windows 8.1(64-bit).You can try to install on Windows 7 and check clinfo again because i don't know if it will work or not.
Regards
Thank you for the suggestion. I thought that AMD APP SDK + Catalyst driver updated was the right path. I am gonna try installing it, even if I think it'll not work on Windows7.
dipak, is it correct?
The latest version is available here Desktop . Did you install this driver?
yes, I installed that version of the driver and AMD APP SDK 3.0 Beta.
My Mobile PC setup is Windows 7-64 bit, AMD Radeon HD 8970M, Intel Xeon CPU octa core.
As landlord suggested, do I need to install the AMD OpenCL 2.0 Driver even if that package is written to support Windows 8 and I've got Windows 7? In that case, does a Windows 7 version exist?
As reported before (in the .txt file), it seems strange to me that after the installation the platform seems to support OpenCL 2.0 and the GPU does not. Also the .dll files in C\Windows\System32 are configured for supporting OpenCL 2.0.
The catalyst 14.41 version, which landlord has suggested, is an old one and one of the first drivers to support OpenCL 2.0. Catalyst 15.7.1 is the latest version and if you've already installed it, then you are on the right track.
As per this page http://developer.amd.com/tools-and-sdks/opencl-zone/amd-accelerated-parallel-processing-app-sdk/syst... , HD 8970M supports OpenCL 2.0. Its strange, however, that clinfo detects the card as an OpenCL 1.2 device and recognizes it as a Pitcairn series with only 32-bit addressing. [Note: OpenCL 2.0 requires 64-bit addressing]
Can you try some earlier versions say 15.7 or 15.6 to see whether you get different observation or not?
Regards,
Hi dipak, I've tried with the driver version 15.6 and the result is the same. The only thing that it seemed different to me was during the installation process of the Catalyst driver, since at a certain point the installer asked me to install the device software, and honestly I don't remember the same window/question installing 15.7 version. By the way, as before the platform recognizes OpenCL 2.0 while the GPU supports OpenCL 1.2. Attached there is the clinfo resume.
Exactly, shouldn't be the Neptune chip to be recognized instead of the Pitcairn one? (AMD Radeon HD 8970M - NotebookCheck.net Tech)