Re: AMD Radeon 5700XT not ready yet?
Posted: Wed Jul 31, 2019 1:52 pm
Strange to have less than 1070ti performance with This 5700XT in luxcore.
Show your work, get help, participate in development
https://forums.luxcorerender.org/
Indeed!Sharlybg wrote: Wed Jul 31, 2019 1:52 pm Strange to have less than 1070ti performance with This 5700XT in luxcore.
I really hope there will be better drivers soon so that you can get the best out of your hardware!FarbigeWelt wrote: Wed Jul 31, 2019 3:40 pm According to some prof. HW tester current AMD openCL driver for 5700 XT is weak what may explain the bad LuxCoreRender values.
Actually shared already in the picture of the table attached to a former post.epilectrolytics wrote: Wed Jul 31, 2019 3:45 pm I really hope there will be better drivers soon so that you can get the best out of your hardware!
Can you share some Luxmark results?
Edit: Oops, I see they are already in the table![]()
Code: Select all
LuxMark 3.1
Company Card GPU Hotel Lobby
AMD R9 290X Hawaii 2453
AMD 5700 XT gfx1010 3060
I thought to ask them some day. I may ask AMD soon because it looks like real issue. The low score makes not any sense.Sharlybg wrote: Tue Aug 06, 2019 5:34 pm Any possibility to ask AMD if RDNA 5700(XT) performance under Luxmark are correct or Driver Limited ?
DISCLOSURE: I AM AN AMD EMPLOYEE
I’m not sure what you mean by this. But Here’s what I’ll say. I work on ProRender for AMD, (check out our blender addon!) . Without Quoting numbers in RPR under OpenCL, render times for 5700XT cards are similar and sometimes slightly better than Radeon VII.
And you can look up Radeon VII benchmarks here independently here: https://www.tomshardware.com/reviews/am ... 977-4.html 41
I wonder how this working for GPU rendering engine. Is RDNA better or Turing ?As to which design is better all comes down to how they get used: a program that generates lots of threads performing FP32 vector calculations and little else would seem to favor Navi, whereas a program with a variety of integer, float, scalar and vector calculations would favor the flexibility of Turing, and so on.
RDNA vs TURING - The END of NVidia’s dominance (@ YT by Coreteks)Sharlybg wrote: Sat Aug 10, 2019 8:16 pm I wonder how this working for GPU rendering engine. Is RDNA better or Turing ?