AMD Energy Monitoring Driver Slated To Be Removed From The Linux Kernel

[willmore]

3DNow! - Market rejected
Clustered Muti-threaded - Market Rejected
Fusion and HSA - Market Rejected
OpenCL- Market Rejected
ROCm - Market Rejected
Mantle - Market Rejected
Infinity Fabric - Market Rejected
Properly feature supported and marketed Consumer products - Market Rejected

x86_64 - Market dominated

You're wrong about OpenCL and Mantle became Vulkan which is pretty darn popular as well. Not sure what your arguement is for Infinity Fabric as AMD uses it on every Zen processor. Are you expecting it to be adopted by Intel or something to say that it's successful?

[Teggs]

AMD's position on this seems incredibly weird. Why not let all of your users, mobile to server, access this information? Does that thought bother someone in AMD management for some reason? Also, everyone and their dog can do this on Windows and has for some time, so I won't buy this 'security' crap unless the functionality goes away there as well. Not expecting that, honestly.

Further, Guenter Roeck is responsible to the Linux community, not to AMD. He could choose to mainline the driver with his changes, make it functional on desktop Ryzen, and tell AMD to go fly. I wish he would.

[Jumbotron]

x86_64 - Market dominated

You're wrong about OpenCL and Mantle became Vulkan which is pretty darn popular as well. Not sure what your arguement is for Infinity Fabric as AMD uses it on every Zen processor. Are you expecting it to be adopted by Intel or something to say that it's successful?

I'm not wrong. OpenCL has been rejected by the market for CUDA and increasingly OneAPI from Intel. Hell, even OneAPI is OpenCL 3.0 compliant. Not so AMD's ROCm.

Mantle was NOT accepted by the market which is why AMD allowed Microsoft to use portions of it for DirectX in order to be used in the AMD powered Xbox. Also after AMD open source it Khronos took it and used it as the basis for Vulkan. It was just a "Why reinvent the wheel" at that point. Vulkan didn't need Mantle to become Vulkan. It was just a convenient jumping off point.

IF won't be of any use outside AMD. The market has gone with CXL, GenZ and NVLink. IF is FANTASTIC tech but it's still just a superset of HyperTransport.

[MadeUpName]

I'm not wrong. OpenCL has been rejected by the market for CUDA and increasingly OneAPI from Intel. Hell, even OneAPI is OpenCL 3.0 compliant. Not so AMD's ROCm.

Mantle was NOT accepted by the market which is why AMD allowed Microsoft to use portions of it for DirectX in order to be used in the AMD powered Xbox. Also after AMD open source it Khronos took it and used it as the basis for Vulkan. It was just a "Why reinvent the wheel" at that point. Vulkan didn't need Mantle to become Vulkan. It was just a convenient jumping off point.

IF won't be of any use outside AMD. The market has gone with CXL, GenZ and NVLink. IF is FANTASTIC tech but it's still just a superset of HyperTransport.

Pretty sure Intel went openCL. Thats 2 vendors to 1. But I do agree AMD needs to get it's act together on the software side.

[coder]

probably because of ease of use. can't make a panel applet, ...

You're answering the wrong question. I think we all get why it's nice to have it accessible to non-priveleged users, but the key question is how it's better to completely remove the module than at least having it accessible to root?

[ThoreauHD]

And what happens when they use AI to interpolate a repeating pattern based on dissipating thermals... ?

The ride never ends.

This sounds like a hardware design problem. If the Heisenberg uncertainty principal needs to apply to your power data, you are doing it wrong.

[coder]

where AMD will ultimately run out of runway AGAIN like they did after the Opteron/Athlon era is in the fact they have NO coherent software and/or compiler stack and NO coherent or even obvious or evident strategic vision for A.I. or ML apart from their GPUs and their purchase of Xilinx and their FPGAs.

Huh? And huh?

OpenCL- Market Rejected

BS. Plenty of people are using OpenCL, in spite of many Vendors and Platform companies rejecting it!

ROCm - Market Rejected

It's still a going concern. If/when it becomes the easy/default setup for AMD GPUs out-of-the-box, it could still win.

Mantle - Market Rejected

Mantle was only ever done as a tech demo. It absolutely achieved its aim of moving the industry off OpenGL and DX11.

Infinity Fabric - Market Rejected

Again, huh?

Man, you really have a way of seizing on this narrative you want to push, and somehow spinning the facts to try and support it. Don't get ahead of the facts. If you can't find enough solid facts to support your case, then maybe you don't have one!

[coder]

You're wrong about OpenCL and Mantle became Vulkan which is pretty darn popular as well. Not sure what your arguement is for Infinity Fabric as AMD uses it on every Zen processor. Are you expecting it to be adopted by Intel or something to say that it's successful?

If you haven't noticed, he's a total Nvidia/ARM shill. Probably has a big pile of NVDA shares.

[coder]

Mantle was NOT accepted by the market ...

If AMD really intended Mantle to be a viable alternative API, please explain why they only ever offered it for GCN cards, even while they still had some TerraScale GPUs in their product lineup!

IF won't be of any use outside AMD.

There was never any suggestion that it would! If they only intended it to be used within their own products while there was no viable, standard alternative, that doesn't make it a failure!

Just cut the BS. Please.

[Jumbotron]

If you haven't noticed, he's a total Nvidia/ARM shill. Probably has a big pile of NVDA shares.

Horseshit. Nvidia is as evil as Intel. But at least Nvidia can execute unlike Intel. And the Big Mo' (momentum) is explicitely and clearly with ARM and its ecosystem and partners.

AMD CAN innovate and exexute. MIGHTILY! But only once per decade and not even for the entire decade. After Zen 4 and 3nm AMD will be tapped out. The Compute Paradigm going forward is Heterogenious and Composable with A.I. baked into the die or as a Discreet Compute part as part of the composibilty of the entire wafer and or SiP (System in Package). AMD does not have any hardware or software roadmaps at all for this paradigm other than cribbing off Intel AVX extensions and CDNA cards. Old school. As of now the best bang for the buck old school x86 compared to Intel but still old school. Intel is going to recover as it did when it dropped the disasterous Pentium 4 and adopted the Core paradigm. Their composable Tiles architecture with BIG.Little scheme ala ARM is going to be a hit. Along with OneAPI and their GPUs Intel will catch up to AMD by the time Zen 4 is rolling out at scale for both HPC and consumers. Around that time Nvidia will have completed their purchase of ARM and will be on their way in designing Grace 2, Ampere Next and their BEASTLY DPU.

And lastly, I have NEVER owned a single stock and never will. I don't even have an IRA or 401K. Never have, never will.