OpenCL

[paul17041993]

I have a 290X and vega64 in the same system, the v64 is the primary card, however in the context of openCL it's device '1', with the 290X being '0'. This is all running windows 10 pro 64 with the latest 18.3.4 drivers.

With the APP SDK 3.0 examples, many will fail to compile or operate on the vega64, but will operate perfectly fine on the 290X (which is the default device, despite being the second card).
A couple I've found so far;
- GlobalMemoryBandwidth returns " clEnqueueSVMMap (inputSVMBuffer) failed! Error code : CL_OUT_OF_RESOURCES "
- FFT returns 31 compile errors, top one being " unknown type name 'class' ", rather obvious that CL 2+ isn't working there...

I haven't tested absolutely all of them, but most of them seem to work just fine, are these issues already known with vega? I haven't been able to find anything about it...

edit; ok so after removing the 290X, GlobalMemoryBandwidth now works correctly, so that could be either a bug with it itself and multiple cards, or a particular driver bug with multiple cards. FFT still returns the same compile errors however...

[dipak]

Hi Paul,

Thank you for reporting this issue. APP SDK 3.0 is quite old, so I don't think APP SDK was tested with the newer cards like vega. Also, it might be a driver issue too. Anyway, I'll report the problem to the concerned team. Please share the clinfo output.

Btw, may I assume that you observed this errors with the pre-built sample binaries available in APP SDK?

To understand the issue better, it would be helpful if you try some older drivers and let us know your observation.

FFT returns 31 compile errors, top one being " unknown type name 'class' ", rather obvious that CL 2+ isn't working there...

FFT kernel contains static c++ code, hence compilation flag "-x clc++" is used to build the kernel. From the above error, it looks like the compiler failed to recognize the static c++ code.

P.S. You have been whitelisted now.

[paul17041993]

This is the clinfo for vega by itself (+CPU), I cant currently provide the clinfo for the vega+hawaii config atm as the hawaii's unplugged...

Number of platforms:			1
Platform Profile:			FULL_PROFILE
Platform Version:			OpenCL 2.1 AMD-APP (2527.10)
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:			2
Device Type:				CL_DEVICE_TYPE_GPU
Vendor ID:				1002h
Board name:				Radeon RX Vega
Device Topology:			PCI[ B#13, D#0, F#0 ]
Max compute units:			64
Max work items dimensions:		3
	Max work items[0]:			1024
	Max work items[1]:			1024
	Max work items[2]:			1024
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:			1630Mhz
Address bits:				64
Max memory allocation:		4244635648
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:			8573157376
Constant buffer size:			4244635648
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:			4244635648
Max global variable size:		3820172032

  Max global variable preferred total size:     8573157376

Max read/write image args:		64
Max on device events:			1024
Queue on device max size:		8388608
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:				00007FF8D2994AB0
Name:					gfx900
Vendor:				Advanced Micro Devices, Inc.
Device OpenCL C version:		OpenCL C 2.0
Driver version:			2527.10 (PAL,HSAIL)
Profile:				FULL_PROFILE
Version:				OpenCL 2.0 AMD-APP (2527.10)
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_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images 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 cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash

Device Type:				CL_DEVICE_TYPE_CPU
Vendor ID:				1002h
Board name:
Max compute units:			12
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:			3593Mhz
Address bits:				64
Max memory allocation:		4276194304
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:				32768
Global memory size:			17104777216
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:			4276194304
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:		284
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:				00007FF8D2994AB0
Name:					AMD Ryzen 5 1600X Six-Core Processor
Vendor:				AuthenticAMD
Device OpenCL C version:		OpenCL C 1.2
Driver version:			2527.10 (sse2,avx)
Profile:				FULL_PROFILE
Version:				OpenCL 1.2 AMD-APP (2527.10)
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_ext_device_fission 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_spir cl_khr_gl_event

[paul17041993]

apparently the forums has some auto-formatting that messes with the info...

[dipak]

Thank you for sharing the clinfo output.

Regarding the compilation error, it looks like the opencl compiler does not support static c++ for Vega devices, hence the compilation is failing with those errors. I'll check with the compiler team and confirm.

apparently the forums has some auto-formatting that messes with the info..

Advanced editor [via "Use Advanced editor" option] can be used for controlling the format. It also provides file attachment option which is useful for large contents.

[paul17041993]

right... for some reason I didn't see the advanced option...

[dipak]

Here is an update. The compiler team has confirmed that static c++ is no longer supported on vega. Actually, OpenCL compiler uses a different toolchain for newer generation GCN devices compared to older cards and hence, some features may no longer be available on those devices. As a result, some of the APP SDK samples (e.g. FFT) may fail on vega.

Regards,