Ticket #3156: clinfo_200507-01.txt

File clinfo_200507-01.txt, 12.5 KB (added by gokhan@…, 6 years ago)

Added by email2trac

Line 
1Number of platforms 1
2 Platform Name NVIDIA CUDA
3 Platform Vendor NVIDIA Corporation
4 Platform Version OpenCL 1.2 CUDA 10.1.0
5 Platform Profile FULL_PROFILE
6 Platform Extensions 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_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer
7 Platform Extensions function suffix NV
8
9 Platform Name NVIDIA CUDA
10Number of devices 2
11 Device Name Tesla V100-PCIE-32GB
12 Device Vendor NVIDIA Corporation
13 Device Vendor ID 0x10de
14 Device Version OpenCL 1.2 CUDA
15 Driver Version 435.21
16 Device OpenCL C Version OpenCL C 1.2
17 Device Type GPU
18 Device Topology (NV) PCI-E, 3b:00.0
19 Device Profile FULL_PROFILE
20 Device Available Yes
21 Compiler Available Yes
22 Linker Available Yes
23 Max compute units 80
24 Max clock frequency 1380MHz
25 Compute Capability (NV) 7.0
26 Device Partition (core)
27 Max number of sub-devices 1
28 Supported partition types None
29 Max work item dimensions 3
30 Max work item sizes 1024x1024x64
31 Max work group size 1024
32 Preferred work group size multiple 32
33 Warp size (NV) 32
34 Preferred / native vector sizes
35 char 1 / 1
36 short 1 / 1
37 int 1 / 1
38 long 1 / 1
39 half 0 / 0 (n/a)
40 float 1 / 1
41 double 1 / 1 (cl_khr_fp64)
42 Half-precision Floating-point support (n/a)
43 Single-precision Floating-point support (core)
44 Denormals Yes
45 Infinity and NANs Yes
46 Round to nearest Yes
47 Round to zero Yes
48 Round to infinity Yes
49 IEEE754-2008 fused multiply-add Yes
50 Support is emulated in software No
51 Correctly-rounded divide and sqrt operations Yes
52 Double-precision Floating-point support (cl_khr_fp64)
53 Denormals Yes
54 Infinity and NANs Yes
55 Round to nearest Yes
56 Round to zero Yes
57 Round to infinity Yes
58 IEEE754-2008 fused multiply-add Yes
59 Support is emulated in software No
60 Address bits 64, Little-Endian
61 Global memory size 34089730048 (31.75GiB)
62 Error Correction support Yes
63 Max memory allocation 8522432512 (7.937GiB)
64 Unified memory for Host and Device No
65 Integrated memory (NV) No
66 Minimum alignment for any data type 128 bytes
67 Alignment of base address 4096 bits (512 bytes)
68 Global Memory cache type Read/Write
69 Global Memory cache size 2621440 (2.5MiB)
70 Global Memory cache line size 128 bytes
71 Image support Yes
72 Max number of samplers per kernel 32
73 Max size for 1D images from buffer 134217728 pixels
74 Max 1D or 2D image array size 2048 images
75 Max 2D image size 32768x32768 pixels
76 Max 3D image size 16384x16384x16384 pixels
77 Max number of read image args 256
78 Max number of write image args 32
79 Local memory type Local
80 Local memory size 49152 (48KiB)
81 Registers per block (NV) 65536
82 Max number of constant args 9
83 Max constant buffer size 65536 (64KiB)
84 Max size of kernel argument 4352 (4.25KiB)
85 Queue properties
86 Out-of-order execution Yes
87 Profiling Yes
88 Prefer user sync for interop No
89 Profiling timer resolution 1000ns
90 Execution capabilities
91 Run OpenCL kernels Yes
92 Run native kernels No
93 Kernel execution timeout (NV) Yes
94 Concurrent copy and kernel execution (NV) Yes
95 Number of async copy engines 7
96 printf() buffer size 1048576 (1024KiB)
97 Built-in kernels
98 Device Extensions 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_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer
99
100 Device Name Tesla V100-PCIE-32GB
101 Device Vendor NVIDIA Corporation
102 Device Vendor ID 0x10de
103 Device Version OpenCL 1.2 CUDA
104 Driver Version 435.21
105 Device OpenCL C Version OpenCL C 1.2
106 Device Type GPU
107 Device Topology (NV) PCI-E, d8:00.0
108 Device Profile FULL_PROFILE
109 Device Available Yes
110 Compiler Available Yes
111 Linker Available Yes
112 Max compute units 80
113 Max clock frequency 1380MHz
114 Compute Capability (NV) 7.0
115 Device Partition (core)
116 Max number of sub-devices 1
117 Supported partition types None
118 Max work item dimensions 3
119 Max work item sizes 1024x1024x64
120 Max work group size 1024
121 Preferred work group size multiple 32
122 Warp size (NV) 32
123 Preferred / native vector sizes
124 char 1 / 1
125 short 1 / 1
126 int 1 / 1
127 long 1 / 1
128 half 0 / 0 (n/a)
129 float 1 / 1
130 double 1 / 1 (cl_khr_fp64)
131 Half-precision Floating-point support (n/a)
132 Single-precision Floating-point support (core)
133 Denormals Yes
134 Infinity and NANs Yes
135 Round to nearest Yes
136 Round to zero Yes
137 Round to infinity Yes
138 IEEE754-2008 fused multiply-add Yes
139 Support is emulated in software No
140 Correctly-rounded divide and sqrt operations Yes
141 Double-precision Floating-point support (cl_khr_fp64)
142 Denormals Yes
143 Infinity and NANs Yes
144 Round to nearest Yes
145 Round to zero Yes
146 Round to infinity Yes
147 IEEE754-2008 fused multiply-add Yes
148 Support is emulated in software No
149 Address bits 64, Little-Endian
150 Global memory size 34089730048 (31.75GiB)
151 Error Correction support Yes
152 Max memory allocation 8522432512 (7.937GiB)
153 Unified memory for Host and Device No
154 Integrated memory (NV) No
155 Minimum alignment for any data type 128 bytes
156 Alignment of base address 4096 bits (512 bytes)
157 Global Memory cache type Read/Write
158 Global Memory cache size 2621440 (2.5MiB)
159 Global Memory cache line size 128 bytes
160 Image support Yes
161 Max number of samplers per kernel 32
162 Max size for 1D images from buffer 134217728 pixels
163 Max 1D or 2D image array size 2048 images
164 Max 2D image size 32768x32768 pixels
165 Max 3D image size 16384x16384x16384 pixels
166 Max number of read image args 256
167 Max number of write image args 32
168 Local memory type Local
169 Local memory size 49152 (48KiB)
170 Registers per block (NV) 65536
171 Max number of constant args 9
172 Max constant buffer size 65536 (64KiB)
173 Max size of kernel argument 4352 (4.25KiB)
174 Queue properties
175 Out-of-order execution Yes
176 Profiling Yes
177 Prefer user sync for interop No
178 Profiling timer resolution 1000ns
179 Execution capabilities
180 Run OpenCL kernels Yes
181 Run native kernels No
182 Kernel execution timeout (NV) Yes
183 Concurrent copy and kernel execution (NV) Yes
184 Number of async copy engines 7
185 printf() buffer size 1048576 (1024KiB)
186 Built-in kernels
187 Device Extensions 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_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer
188
189NULL platform behavior
190 clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) NVIDIA CUDA
191 clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [NV]
192 clCreateContext(NULL, ...) [default] Success [NV]
193 clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) No platform
194 clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
195 clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) No platform
196 clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform
197 clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) Invalid device type for platform
198 clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) No platform
199
200ICD loader properties
201 ICD loader Name OpenCL ICD Loader
202 ICD loader Vendor OCL Icd free software
203 ICD loader Version 2.2.11
204 ICD loader Profile OpenCL 2.1