"More I read your comments in this thread and more I cannot believe that you for real. This is kind of elaborate role-playing kind of trolling, right ? A person cannot be this delusional and devoted to a corporate entity even if he works there or owns it, right, to make such giant leaps of logic made of fallacies, fact-distortion and straight made-up absurdities ? It's some kind of attempt in sattire by imitation of a caricature of a fan, right ?"
dude, i am neither a fanboy nor do i work (at all - retired since 7 years) nor do i own AMD stock etc (what i could actually). if you think that anything of what i said is made-up or fact-distortion than you are to stupid to understand basic hardware design...
"How crazy must you be to suggest to people to "try" on a whim other expensive pieces of equipment in place of a whole bunch of expensive equipment they already own ?"
stupid? dont people have a friend where they can test a gpu?
"And multiple monitors, gaming mice and "chipsets" (meaning "motherboards") at that ?"
??? try verbs // btw i meant chipsets...
"Even shⅰtty USB mice can't break PCIe bus past its USB controller and it can't go past standard USB's 1000 interrrupts per second (1ms delays for high-spec mice). "MoBos" can't "share pci-lanes for USB and pci-slots", USB controller is a chip which gets PCIe lanes from PCIe controller along with its dedicated interrupt from CPU, only the latter may be shared with a GPU but even that is rare and inconsequential. Every PCIe lane is dedicated to its one endpoint, be it a chip or a slot. There might be a multiplexing PCIe commutation chip between PCIe controller and some devices to make it appear as if there were more lanes BUT it will only limit maximum bandwidth, at worst. On my 2011 PCIe 2.0 system with RX 580 one USB controller (there might be several, especially for 2.0 and 3.x), an additional SATA and a WiFi PCIe-based slotted controllers share same interrupt but that's that, they still have their own lanes. And GPU has everything of its own.
So stop making shⅰt up."
USB-Killer sry its in german but just watch what happens - if the USB Controller isnt in a secured circuit (what most cheap mobos or laptops wont be) a simple USB stick can FRY your fricking PC - stop being stupid
PCI Express lane sharing - again german - but here you only see 16 lanes from cpu to pcie-switch and 32 lanes going from pcie-switch to ports etc; and that is on an expensive board and no it does more than limiting bandwidth: it also increases latency - and if latency is to high and takes over 2-3 cpu cycles your cpu cant handle the pci-e packets anymore and will drop them // cheaper MoBos use different ways and are even worse
btw - do we know if he uses first pcie slot? otherwise he maybe use shared pci-lanes
"All closed-source AMD drivers always had problems because their development process sucks and they have newer fixed that."
in my opinion all proprietary sh*t has problems
btw Radeon and RadeonFeature is based on AMD documentation - also OpenCL should work there (but havent tested it - no Linux Desktop since years)
"We might not be AMD devs but I DO sometimes look at their code in Linux kernel (actual driver) and Mesa & LLVM projects (code compilation, shader interpretation, OpenGL and Vulkan implementaions for GPU; all the "fat", framework stuff and the reason why GPU "drivers" are gigantic), debug stuff and report directly to them. So I know a thing or two about that."
i hope the "report to AMD" function is a way to increase priority of a bug - at least there is hope
would be great if the driver would do the whole job on its own - but what is with bugs that crashes the pc? only possible to collect memdump until crashing - and if it takes longer than 3min the dump will be HUGE - so without repeatable bug the driver is kind of screwed doing it full auto - or amd has to solder 128gb nand only for debug onto any GPU ;D
usually when i report a driver bug on amd.com/report i do upload a dump to my server and post the link into my error description
// about the cp detection - no accusation in any form - i just said that this feature can be used for that - especially because it prevents the criminal from doing anything against it
since 10 years all big SW and HW companies work together on that
many cp is known - Google & MS run pattern-detection on that (not only faces - but everything that has a unique pattern in the picture/video: like distance between window and bed in combined with position of body etc - they save those patterns
MS (especially since Win10) trys to detect those patterns locally when you open a picture
google checks any picture you reverse-search
there was a test where graphics-drivers contain some of those patterns and check the output especially for videos
I knew sooner or later someone would come up with the phrase "they're human, they are working hard to fix it".
Let's put this issue from a business perspective. Imagine a company has to rely on AMD devices to run their business. Imagine they have a problem like this that makes their entire business come to a halt and they have to wait for the next update for it to be fixed (and even then, the solution is not guaranteed). Is the company willing to wait for over 2 months until this issue is fixed? Of course not! They would find the quickest solution to their problem, and here is where the problem lies.
You can tell me they're working hard to fix this and that the temporary solution is there, but this is totally unacceptable, mainly because I'm not getting what I paid for my Rx 580 8GB, which is not a cheap GPU at all.
I'm sorry, but your argument won't hold water here, they have the money, the personnel and the technology to fix this. My patience ran out a while ago, I switched to another brand already and their product have solved this and other problems I had.
Update: Fixed - AMD Radeon Software freeze computer mouse for 1 second, every 5 seconds'
Works with v18.8.1 and previous version v18.7.1 (also after the driver upgrade, I didn't have to reset the option.)
Wickex, you saintly person you.
THANK YOU for posting your fix here!
I clicked the HDCP Status to disabled , as you described, and it cleared the "freezing/stuttering mouse problem immediatly!
Google search bought me here. And I chanced on Wickex post, thank heavens. I noticed a Wickex reply was saying he had fixed it and then it didn't work, or worked for a bit. I hope you fixed it permanently, mate!
I typed how to do it again , to make it clear so more people can find it
NB**The option can only be changed when you have the second or 3rd monitor enabled; it can't be changed if you have only one monitor, I found out**
As he described; Load up display settings window
Goto the Display tab
Click the Specs selection on the right
Click the overide selection
and change the option 'HDCP SUPPORT' to "DISABLED"
My original setup Ryzen 1700, "ati 7800 2gb" graphics card, running 1st (main) Monitor on DVI , 2nd Monitor (display port adaptor--> VGA) , 3rd monitor (HDMI ---->HDMI 720p TV)
This was working fine with all three monitors.
I had intermediatly running the same monitor setup with "AMD fury R9 tri-x 4gb HBM", That had run the monitor setup flawlessly. (until the card failed)
I bought TODAY, this AMD Dual RX580 8gb, plugged it in and , as all you guys have, I get this mouse pause/freeze (it's too long a pause to be called a stutter ( about half a second long pauses) with intervals of about 1 or 2 seconds, when the mouse pointer is simply moved around the screen. It only happens when the monitors are in extended or duplicate mode (set via the AMD software). I noticed when I only set one monitor on the system , the mouse worked as it should , smooth as can be.
Seriously it's seemed so specific that AMD should READ THIS and look into it!
Quote: "HDCP option is 'High Bandwidth Digital Contect Protection'. A signal standard designed to protect HD content from piracy and copying."
juliannb - pn
correct would be pm // personal message
about his complaint - only nonsense: a fix for a critical bug will be released in 1-3 days //
about his other stuff he got a pm
Okay, was wondering what PN stood for as a future reference but turned out to be a typo. Thanks for the clarification.
actually it isnt even a typo - pm is pn in german