HW acceleration doesn't work in RTX4080
I am using Capture One 20. (13.1.4)
I replaced the graphics card with RTX4080.
Then hardware acceleration doesn't work.
Anyone having the same problem?
This card nvidia driver is 527.56
I tried rebuilding the image core by referring to the article in the FAQ, but the same problem occurs.
If anyone has solved this problem, please help.
ImageCore.log
2022-12-23 20:06:31.887> Logging is now active.
2022-12-23 20:06:31.887> CPU: AuthenticAMD [AMD Ryzen 9 5900X 12-Core Processor]
2022-12-23 20:06:31.887> CPU features: MMX, SSE, SSE2, SSE3, SSSE3, SSE41, SSE42, AVX, AVX2, FMA, F16C, MOVBE, CX8, RDRAND, RDSEED
2022-12-23 20:06:31.887> CPU features: ADX, RDTSCP, POPCNT, BMI1, BMI2, LZCNT
2022-12-23 20:06:31.905> First chance exception (thread 9128): 0x406D1388 - Set thread name
2022-12-23 20:06:31.984> First chance exception (thread 4852): 0xE06D7363 - C++ exception
2022-12-23 20:06:32.017> First chance exception (thread 15192): 0xE06D7363 - C++ exception
2022-12-23 20:06:32.028> First chance exception (thread 16136): 0xE06D7363 - C++ exception
2022-12-23 20:06:32.044> First chance exception (thread 16116): 0xE06D7363 - C++ exception
2022-12-23 20:06:32.300> First chance exception (thread 11916): 0xE06D7363 - C++ exception
2022-12-23 20:06:32.709> First chance exception (thread 3776): 0xE06D7363 - C++ exception
2022-12-23 20:06:42.531> First chance exception (thread 4852): 0xE06D7363 - C++ exception
2022-12-23 20:06:42.531> (4 identical messages logged; delayed 10.547s .. 10.547s.)
2022-12-23 20:06:42.531> First chance exception (thread 15192): 0xE06D7363 - C++ exception
2022-12-23 20:06:42.531> (14 identical messages logged; delayed 10.514s .. 10.506s.)
2022-12-23 20:06:42.531> First chance exception (thread 16136): 0xE06D7363 - C++ exception
2022-12-23 20:06:42.531> (9 identical messages logged; delayed 10.503s .. 10.500s.)
2022-12-23 20:06:42.531> First chance exception (thread 16116): 0xE06D7363 - C++ exception
2022-12-23 20:06:42.531> (4 identical messages logged; delayed 10.486s .. 10.486s.)
2022-12-23 20:06:42.531> OpenCL initialization...
2022-12-23 20:06:42.542> Found 1 OpenCL platforms
2022-12-23 20:06:42.542> Found 1 OpenCL devices on platform 0 (NVIDIA CUDA)
2022-12-23 20:06:42.542> OpenCL Device 0 : NVIDIA GeForce RTX 4080
2022-12-23 20:06:42.542> OpenCL Driver Version : 527.56
2022-12-23 20:06:42.542> OpenCL Compute Units : 76
2022-12-23 20:06:42.542> OpenCL Nvidia Compute Capability : 8.9
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_HOST_UNIFIED_MEMORY : 0
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_GLOBAL_MEM_CACHE_SIZE : 2179072
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE : 128
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_GLOBAL_MEM_CACHE_TYPE : 2
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_GLOBAL_MEM_SIZE : 16375 mb
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_MAX_MEM_ALLOC_SIZE : 4093 mb
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_LOCAL_MEM_SIZE : 49152
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE : 65536
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_VENDOR_ID : 4318
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_VENDOR : NVIDIA Corporation
2022-12-23 20:06:42.542> OpenCL CL_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_3d_image_writes 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_d3d10_sharing cl_khr_d3d10_sharing cl_nv_d3d11_sharing cl_nv_copy_opts cl_nv_create_buffer cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_device_uuid cl_khr_pci_bus_info cl_khr_external_semaphore cl_khr_external_memory cl_khr_external_semaphore_win32 cl_khr_external_memory_win32
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_MAX_WORK_GROUP_SIZE : 1024
2022-12-23 20:06:42.542> OpenCL CL_DEVICE_ADDRESS_BITS : 64
2022-12-23 20:06:42.599> OpenCL : Loading [C:\Users\hp870\AppData\Local\CaptureOne\ImageCore\13.1.4.3\ICOCL.bin] started
2022-12-23 20:06:42.599> OpenCL : Loading [C:\Users\hp870\AppData\Local\CaptureOne\ImageCore\13.1.4.3\ICOCL.bin] OK
2022-12-23 20:06:42.599> (ERROR) bin file failed parse [C:\Users\hp870\AppData\Local\CaptureOne\ImageCore\13.1.4.3\ICOCL.bin] (verificationCode=7)
2022-12-23 20:07:01.276> First chance exception (thread 3776): 0xE06D7363 - C++ exception
2022-12-23 20:07:01.276> (9 identical messages logged; delayed 28.566s .. 27.294s.)
2022-12-23 20:07:01.276> First chance exception (thread 11916): 0xE06D7363 - C++ exception
2022-12-23 20:07:01.276> (24 identical messages logged; delayed 28.975s .. 25.571s.)
2022-12-23 20:07:01.276> First chance exception (thread 11916): 0xE06D7363 - C++ exception
-
I deleted imagecore and rerun, but it still didn't work. I have sent the log file to support and am waiting for a reply.
1 -
An important issue, the RTX40 series is not supported in version 517.40. For RTX4080, the first supported version is after 522 version. Version 528 came out today, but the problem is still not resolved.
1 -
Interested in the outcome of this also - Any word?
1 -
Same issue with my new RTX 4060. Hardware acceleration is disabled. I would have to enabled my integrated Intel GPU in order for HA to work again....but I clearly don't want to do that.
Someone mentioned downgrading the driver version, did that work for anyone?
1 -
Let me refresh this (8 months old!) thread.
Recently I bought new computer, one of the main purpose was to use Capture One more effectively.
My specification is GF RTX 4070 Ti, so also 4xxx family. I use the latest Nvidia studio driver 536.99.
My version is exactly the same as OP's - Capture One 20 - 13.1.4.
I have exactly the same issue. No HW acceleration working.
The status in ImageCore/13.1.4.3/ICOCL.xml looks always the same:
<?xml version="1.0" encoding="utf-8"?>
<device>
<Index>0</Index>
<Platform>NVIDIA CUDA</Platform>
<Version>OpenCL 3.0 CUDA 12.2.135</Version>
<Device>NVIDIA GeForce RTX 4070 Ti</Device>
<Driver>536.99</Driver>
<GlobalMemoryMB>12281</GlobalMemoryMB>
<NumberOfKernels>1298</NumberOfKernels>
<NumberOfCompiledKernels>1189</NumberOfCompiledKernels>
<Benchmark>10000.000000</Benchmark>
<Status>Compiled:CL_BUILD_PROGRAM_FAILURE (ResamplePreFilter)</Status>
</device>The NumberOfCompiledKernels is always the same - 1189. The error is compilation error: Compiled:CL_BUILD_PROGRAM_FAILURE (ResamplePreFilter)
My ImageCore.log looks the same as OP's, but I think the most interesting part is in ImageCoreOCL.log, which hasn't been posted yet. This is the content (shrinked the repeting prints):
2023-08-24 22:00:24.448> Logging is now active.
2023-08-24 22:00:24.448> CPU: GenuineIntel [13th Gen Intel(R) Core(TM) i5-13600K]
2023-08-24 22:00:24.448> CPU features: MMX, SSE, SSE2, SSE3, SSSE3, SSE41, SSE42, AVX, AVX2, FMA, F16C, MOVBE, CX8, RDRAND, RDSEED
2023-08-24 22:00:24.448> CPU features: ADX, RDTSCP, POPCNT, BMI1, BMI2, LZCNT, SHA
2023-08-24 22:00:24.448> OpenCL initialization...
2023-08-24 22:00:24.557> Found 1 OpenCL platforms
2023-08-24 22:00:24.557> Found 1 OpenCL devices on platform 0 (NVIDIA CUDA)
2023-08-24 22:00:24.557> OpenCL Device 0 : NVIDIA GeForce RTX 4070 Ti
2023-08-24 22:00:24.557> OpenCL Driver Version : 536.99
2023-08-24 22:00:24.557> OpenCL Compute Units : 60
2023-08-24 22:00:24.557> OpenCL Nvidia Compute Capability : 8.9
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_HOST_UNIFIED_MEMORY : 0
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_GLOBAL_MEM_CACHE_SIZE : 1720320
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE : 128
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_GLOBAL_MEM_CACHE_TYPE : 2
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_GLOBAL_MEM_SIZE : 12281 mb
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_MAX_MEM_ALLOC_SIZE : 3070 mb
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_LOCAL_MEM_SIZE : 49152
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE : 65536
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_VENDOR_ID : 4318
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_VENDOR : NVIDIA Corporation
2023-08-24 22:00:24.557> OpenCL CL_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_3d_image_writes 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_d3d10_sharing cl_khr_d3d10_sharing cl_nv_d3d11_sharing cl_nv_copy_opts cl_nv_create_buffer cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_device_uuid cl_khr_pci_bus_info cl_khr_external_semaphore cl_khr_external_memory cl_khr_external_semaphore_win32 cl_khr_external_memory_win32
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_MAX_WORK_GROUP_SIZE : 1024
2023-08-24 22:00:24.557> OpenCL CL_DEVICE_ADDRESS_BITS : 64
2023-08-24 22:00:24.620> OpenCL : Loading [C:\Users\Wojtek\AppData\Local\CaptureOne\ImageCore\13.1.4.3\ICOCL.bin] started
2023-08-24 22:00:24.620> OpenCL : Loading [C:\Users\Wojtek\AppData\Local\CaptureOne\ImageCore\13.1.4.3\ICOCL.bin] Incomplete
2023-08-24 22:00:24.620> (ERROR) Failed to open C:\Users\Wojtek\AppData\Local\CaptureOne\ImageCore\13.1.4.3\ICOCL.xml for writing
2023-08-24 22:00:24.620> OpenCL : Building kernels on dev 0 [NVIDIA GeForce RTX 4070 Ti]
2023-08-24 22:00:24.885> Compiled 0/1298
2023-08-24 22:00:25.683> Compiled 1/1298
2023-08-24 22:00:26.246> Compiled 6/1298
... <cut the Compiled messages>
2023-08-24 22:05:43.594> Compiled 1183/1298
2023-08-24 22:05:44.493> Compiled 1185/1298
2023-08-24 22:05:44.868> (ERROR) Compilation of ResamplePreFilter failed : CL_BUILD_PROGRAM_FAILURE
2023-08-24 22:05:44.868> (ERROR) Compilation of ResamplePreFilter returned : CL_BUILD_PROGRAM_FAILURE
<kernel>:118:29: error: passing '__generic float *' to parameter of type 'float *' changes address space of pointer
1;case 2:return ze94c3fdd77(ze7168fc0b9,zff400e9e97);case 3:return zda883a2d9b(
^~~~~~~~~~~
<kernel>:87:50: note: passing argument to parameter 'ze7168fc0b9' here
}return z7d07bb3247;}int ze94c3fdd77(z558ec19d7e ze7168fc0b9[3],z558ec19d7e
^
<kernel>:118:41: error: passing '__generic float *' to parameter of type 'float *' changes address space of pointer
1;case 2:return ze94c3fdd77(ze7168fc0b9,zff400e9e97);case 3:return zda883a2d9b(
^~~~~~~~~~~
<kernel>:88:1: note: passing argument to parameter 'zff400e9e97' here
zff400e9e97[2]){
^
<kernel>:119:1: error: passing '__generic float *' to parameter of type 'float *' changes address space of pointer
ze7168fc0b9,zff400e9e97);default:return-1;}}void z5417645477(ze9dc86233a*
^~~~~~~~~~~
<kernel>:95:51: note: passing argument to parameter 'ze7168fc0b9' here
z11fc704ac4;return 2;}int zda883a2d9b(z558ec19d7e ze7168fc0b9[4],z558ec19d7e
^
<kernel>:119:13: error: passing '__generic float *' to parameter of type 'float *' changes address space of pointer
ze7168fc0b9,zff400e9e97);default:return-1;}}void z5417645477(ze9dc86233a*
^~~~~~~~~~~
<kernel>:96:1: note: passing argument to parameter 'zff400e9e97' here
zff400e9e97[3]){z558ec19d7e z2dd09f3226=ze7168fc0b9[2]/ze7168fc0b9[3];
^
2023-08-24 22:05:44.868> (ERROR) Failed to create ResamplePreFilter : CL_BUILD_PROGRAM_FAILURESo you can see there's a direct point where the compilation fails. And it is always happening during compilation of kernel no 1190.
Capture One team can you help with that?
1 -
I have exactly the same issue with Nvidia 3060 Ti.
The problem is with the GPU drivers. When I use Nvidia studio drivers 527.56 hardware acceleration in Capture one does not work. When I downgrade to v517.40 everything is working fine.
Quite disappointed with Nvidia GPUs. I switched from AMD a week ago and I am constantly fighting some gpu related issues since then.
0 -
I deleted imagecore and rerun, but it still didn't work. I have sent the log file to support and am waiting for a reply.
Gump, any word back from support? I have the same driver issue with an Nv 1660S--nothing beyond 517.40 works in C1 v 20 (13.1.2.35)*
*Your post alerted me to the maintenance release, 13.1.4. I was not aware of it. Thanks for that anyway. :)
0 -
In the end, I solved it by renewing the license of capture one to 21. In the case of 20, support has ended, and updating to the latest version seems to be the best solution for problems arising from the latest driver or hardware.
0 -
I'm not sure I want to "upgrade" to a newer version of C1. Does anyone know just what features are lost when GPU acceleration is disabled? I can't seem to find that info re version 20.
0 -
I believe drivers old enough for the OpenCL version to be compatible are too old for 4xxx series cards.
0 -
The new dual encoders capture up to 8K resolution at 60 FPS in real time via GeForce Experience and OBS Studio, and cut video export times nearly in half. Popular video-editing apps have released updates to enable this setting, including Adobe Premiere Pro (via the popular Voukoder plug-in) and Jianying Pro — China’s top video-editing app. Blackmagic Design’s DaVinci Resolve and MAGIX Vegas Pro also added dual-encoder support this week. Fortiva Credit Card Apply Online
-1
Please sign in to leave a comment.
Comments
11 comments