![]() I expect the GPU to be at a clear advantage in his setup as well. Compared to Tordan's setup he has half the CPU power and around two-thirds of the gaming power of the 2060. ![]() But Mazdaman has ordered a Ryzen 7 3400G CPU and nVidia GTX1660 S GPU. If Tordan58 does these tests with his pretty new nVidia RTX 2060, I believe we'll see a large advantage for the GPU. The CPU is marginally faster but by using the GPU I can continue running other things smoothly while DeNoise is rendering using the GPU.īut bear in mind my CPU is a high-end consumer CPU which has only been around for about a year while the GTX 970 GPU has been around for over six years. My Ryis a pretty close match to my nVidia GeForce GTX970 GPU for DeNoise AI. I'd like to build on that hypothesis by saying that it's the relative power of CPU and GPU that matters. My results put in a pretty chart - but the message is clear!ĭespite the fact that OpenVine was produced by Intel for Intel processors, it clearly works extremely well on AMD Ryzen processors as well!Īs we're discovering, the 'AI' aspect of TopazLabs DeNoise clearly gives a decent CPU combined with Intel OpenVine support an advantage over even quite powerful middle-tier graphics card GPUs.ĭave wornish has already rightly reflected that it will take a particularly powerful graphics card to eclipse a CPU running with Intel OpenVINE selected. I've done a more organised test mirroring everyone else and I improved my presentation: It was just a quick test I'm sure it can be refined! This was an ISO 12800 image in a very dark room and the exposure at f/5 was just 1/15th second No Denoise, only tone adjustments in Photoshop and Adobe Camera RAW Using the CPU only (no OpenVine mode) previews took 7 seconds and Applying changes took 134 seconds. Using the discrete GPU mode the results were pretty much identical to using the CPU and OpenVine mode. Using the CPU with OpenVine adjustmenst previewed in around 1 second. I ran DeNoise AI version 2.2.2 in filter plug-in mode from within Adobe Photoshop CC 20.0.9, using DeNoise AI Low Light mode set to on. I run my display at 4K (3840x2160) and 175% text/apps setting. My (desktop/tower\ setup is: AMD R圜PU, 32GB 2666MHz dual-channel RAM, ASUS PRIME B450M-A Motherboard, Crucial MX SATA3 500GB system drive, nVidia GeForce GTX970 GPU, running Windows 10 Professional version 1909. I've done quick test with my 12800 RAW file. You can put your own data into this spreadsheet with a chart to display in this thread:įinally, please consider helping to fund this site by using the following commission link when purchasing any TopazLabs product: Here are the links (use right-click or equivalent 'save link as - filename'):įile 5 JPEG P6180457.JPG ISO12800 Large Normal E-M1 Mark IIįile 6 RAW P6180457.ORF ISO12800 RAW E-M1 Mark II If possible, please let us know what type of system drive you are using (conventional hard disk, SSD, NVME SSD) and whether when testing you are using DeNoise stand-alone or as a plug-in. Whether your computer is a laptop or desktop/tower system.GPU and dedicated video RAM (if present) (GB) and if possible type (DDR5/6, etc.).System RAM (GB) and if you know, type (DDR3/4, etc.) and speed 2400Mhz, for example).When posting results please let us know as much of the following as possible: ![]() No other use is permitted without his permission. Please respect Juan's copyright - he only provides these images for testing by E-Group members. Below are links several RAW and JPEG image files provided by myself and c12402 (Juan Guerrero) for you to use in testing. This thread is dedicated to experimentation with Denoise AI. This leads to discussions on the hardware required to optimise performance, from the choice of CPU, graphics hardware (GPU) and software settings - all of which can have a considerable effect. Folks are asking a lot of questions about its effectiveness and the time it takes to work. There has been considerable interest in the TopazLabs Denoise AI application/plugin of late.
0 Comments
Leave a Reply. |