ClockBlocker (profiled AMD power-management control)

Any chance we can get python or lua support? Or maybe you can put out a pypi package for CB? This would be a great addition.

 

Rather than asking for a scripting language to implement some type of a profile heuristic yourself, care to elaborate what you're thinking of instead?


For example: Is there something that you want ClockBlocker's profile detection to do that it currently doesn't? I'm assuming that the scripting interface is for altering how detection works or defining new types of rules.

EDIT: More or less this is a purely C++ project right now. While I could add a Python interpreter, I'd have to be convinced of why it's needed first. Execution speed and memory use is important for gamers, and the hit of tools should be "minimal" on a system while running games.

 

What I want is to replace CB's controller so that I can write my own profile support ground up, but I want to use your blocker modules (pypi hint ).

No offence but CB's gui is like something from win 3.1 and it needs to be modernized. C++ is probably why it looks so dated .. low level languages aren't good for gui, you need python or C# mixed with some GL.

 

LOL

--I knew that I remembered you saying something else in the thread earlier here:

Furthermore, use of ANY accelerated 3D-Graphics for the GUI (OpenGL, D2D, D3D) is going to give you problems with the detection of "other" software. Something that I was trying to adamantly avoid through the design choices here.

I also refute that the used language has ANYTHING to do with how the GUI looks, this is dependent on the API or framework used.
(everything is custom drawn, almost nothing stock-MFC is used in here)


--fyi, you still haven't told me what my software's profile system doesn't do that you're intending to basically "throw it all away" and rewrite it to do. To me it more feels like you just have something up your butt about ClockBlocker (or me).

*I'll take your request more seriously if you can provide something valuable that my profile system cannot be scripted to do already.

 

Not a programmer as a profession :3eyes: that doesn't mean I can't write software. You feel like you got the stick up ur butt buddy.

A good gui would have automation and prompts on starting software, not this manual stuff for creating rules. CB is full manual, hard, overcomplicated. I want to streamline it and make it less ugly.

 

Ok, whatever.

Uh? Prompts on starting programs is a good idea?... Sorry, but I'm going to have to disagree with this, and that's also something that I tried hard to avoid doing by design.

-There already IS automation through the fullscreen and 3D detection heuristics. NOTE: This is explicitly prompt-less and needs nothing from the user.
-If there's problems with the heuristics, the user can specify exceptions.
-If a whitelist behavior is desired, then there's the wildcard (mini-regex) feature (inclusion of directories and not just single-programs).


ClockBlocker is built in principle on a "define exceptions" (blacklist), not "define programs" (whitelist) notion. -- this is done to minimize the input required from the user.

---I'd say that ClockBlocker has strong automation in its profile system already, IMO.

 

And I disagree .. a prompt on loading a program is a one time thing, insignificent. Windows does this with windows firewall and it gets praised as the easiest to use firewall, it doesn't annoy anyone.

Nothing is easier than clicking allow or block on starting a program.

 

Then we'll agree to disagree, considering that I dislike Windows Firewall's stock-behavior (filling in one-time prompts for each application).
Doubt I'm the only one that dislikes this too!


-Clicking the firewall-prompt requires that I alt-tab the game "OR" wait till the application ends.
-Often times I may 'accidentally' hit a prompt that I didn't intend to, requiring that I enter the Control Panel to make a correction.
-Rules per application -- unnecessarily litters my firewall rules with applications that may even be temporary or get uninstalled.
* (inflates the number of rules)
-Have to generally check the rules to make sure that I clicked what I "thought" that I did for troubleshooting's sake. (security or when things don't work)
-Having an additional prompt is no easier than having zero-prompts.


This suggestion is getting rejected. Advanced scripting support (Python, LUA, AngelScript, etc) will be considered if and when something useful is proposed that actually needs such.

 

Just because you don't like it doesn't mean it isn't easier for people and not more popular (you should go read reddit about how hard your app is to use). It sounds that because you don't like the most popular way to do software .. you won't add support for it or let someone else create it.. nice

 

I guess I'm about to feed you some more "pointless arguments".
--By the way, editing my posts via quotes (in responses) is not the way to get me on your side, nor the best way to convince me to take my time to do something for you.

Nevertheless, let's take a quick look at your arrogant self-asserted facts (opinions) this time:

-Sure, Windows' built in firewall is the most common, given that Windows is the most common deployed OS on consumer machines. However, if you were to look at "firewall software" as a whole, you'd notice that the majority of competition does not mimic Windows in their behavior. ... So, why is that?

One might argue that the software doesn't behave the same, given noone would buy a replacement product that acted identically to the built-in. (possibly true) Yet, at the same time Windows Firewall's dominant use doesn't imply that MS's decision is what people want. (chicken and egg argument)

EDIT: Keep in mind that I am being VERY generous here and not counting Linux based in-home appliances with firewalls, nor am I considering the Linux & BSD Server marketshare.

I check Guru3D more than any other forum or news site. I've never stated that I would provide support via "Reddit", and infact don't even have a Reddit account. If people are trying to reach me on Reddit or complain about the software's interface on Reddit, that would be a mistake.

Regardless, if you have some key Reddit threads that you want me to know about (or think that I should know about), please post links to them (or PM me them).
** I'm not omnipresent **

You are free to create your own tool similar to ClockBlocker, I promise that I won't stop you. At the same time I feel no obligation to make changes to this tool on your behalf, nor to put in work that I don't feel will improve the software.

-Whenever I personally see a need for a change, I generally do make those changes. Yet you need to remember that I'm working effectively for free, and on my own time on this project.

 

Microsoft has focus groups and spends money researching these decisions, they know better than you. It is a fact whether you like it or not, lol.

Takes like 2 minutes in google man, there's tons of them.

Since you don't get money just opensource your project and we'll be happy. I'm not asking you to add anything. Get off ur high horse.

 

Microsoft can be wrong and I can also disagree with Microsoft. I do not treat MS's decisions as graceful acts of god, as you apparently do.

-I don't care how highly you think of them, it doesn't change my decision.

-I will not "guess" keywords and search for these fabled threads. Links or they don't exist as far as I'm concerned.

No. (given you asked so politely and all)

-Don't bother replying, I won't reverse that.

 

hello, i have a problem with that program. sorry for my bad english, i use this program almost 8 months without problems. but yesterday its stop working. its seems error in ClockBlocker_HelperModule.exe or amdocl.dll, i dont know.
" Application Error
ClockBlocker_HelperModule.exe
1.0.0.1
584c31ad
amdocl.dll
22.19.171.257
593f01dc
c0000005
001fc366
21a8
01d2edbccd2f4f60
C:\ClockBlocker 1.3.3 BETA\ClockBlocker_HelperModule.exe
C:\WINDOWS\System32\DriverStore\FileRepository\c0315159.inf_amd64_b01c1e8cccf04a67\amdocl.dll
eecbb7a7-75fc-46fc-8f44-4b6cdb0a394d "

 

Hey there Perre,

It sounds like there's a problem loading the OpenCL ICD on your machine (probably the drivers changed on you). Most likely if this started abruptly (without you updating anything), I'm going to guess that Windows Update was involved here. Please check and see if WUA (Windows Update Agent) installed a new driver recently within the last day or so.
[from the updates history]


-I'm assuming that this is a Windows 10 PC (given the possibility of WUA replacing video-drivers), but I'll need to know a bit more about what you're running as well.
(*installed video-driver and OS version*)


For a start in what you can try:
=> Disable ClockBlocker (remove autostart, close the software)

-Run DDU (thread link) from safemode, clean / purge your display drivers, restart the machine (non safemode) and then reinstall the latest AMD drivers.

=> Afterwards, with a fresh driver installation, try reenabling ClockBlocker and see if OpenCL loads correctly this time around.


EDIT:
--That said, this looks from the CL component like it's a 17.6.2 install. I'm more curious if you know how it got on there, or if WUA put it there.

 

ok, thanks, but reinstalling driver doesn't help me. no idea, what happening with this files. I solved the problem by reinstalling windows :smoke:, I hope it will not happen again

 

Ah, if that full format / wipe of the machine helped then something definitely got mucked up with the drivers and CL in some way or another. If this happens again, do let me know. -Back in the early days of ClockBlocker (the 1.x releases), some people were seeing similar exceptions (rarely). These were fixed though in changing the way that CL and DCU gets loaded and initialized.


Still may be possible that there's more I can do (improve init further). Another good question would be if other software was able to use CL when this all happened. But yes, we'll wait and see, hopefully it just keeps working for you on this clean install.

 

Any updates on 17.7.2 support? CB has been broken a few days now.

 

If you mean the recent ADL changes, I will be watching this and haven't decided the next action just yet. The ability to explicitly adjust power-states means that there's likely little reason to fix the current methods of preventing downclocking for further driver-releases. (we will see)

Those running older drivers (who cannot run 17.7.2 onward) would of course be able to keep using the current releases of ClockBlocker. If there's a demand to keep using ClockBlocker's profile system, then I'll likely wait and see the direction that this goes first and adjust accordingly thereafter.


Note that I am not abandoning the software, ClockBlocker will be evolving in to something else down the road ... YAP:



This is pre-pre-alpha (non feature complete by a longshot), very early snapshot. So keep in mind that this is not coming soon.

-full task-manager this time
-service based, able to read info on all processes
-client-server
-ground up rewrite (none of the old code)
-similar profile system concept (will keep many concepts of ClockBlocker)
-cross-platform (Windows, Linux, Android) -- eventually...

[starting with Windows first]
Planning to support Windows and Linux as server / client, Android as a client only.

 

omg.. lol your naming dude.. is this like FAP = YAP?
Your program names are epic.


Looks like process lasso, but your not pulling a RP and dropping free software are you? if this is free that sounds sick.

 

The name is totally unoriginal, you can see it in the title of the server process.
... That it sounds like something else is not my fault and totally coincidental. **COUGH**

Actually the concept that I have in my mind is very different from Process Lasso (also a great program), this simply isn't apparent from a single screenshot or through what's implemented right-now. At some point later I promise that this will become more clear that YAP is not just another process manager clone, despite what the title implies.

That said it will however, be able to do anything that a process manager can (adjust process priority, IO priority, kill processes, display process stats, etc, etc). --However, YAP is going to be more about "gaming" process and GPU control.

--Specifically aspects such as gaining control over a machine after a game locks up and takes out the desktop (blocks access to the UI).

That's a bit unfair a way to put it as developers have to make a living, one way or another.


To get to the point YAP will be primarily-free (no startup nags, no ads, no bundleware, no registration nags), just like ClockBlocker (more features than ClockBlocker, this is a larger project).

... At the same time a few people both in the thread and by mail asked to donate, this is a touchy topic I know. My personal-take on this is that I'm not comfortable with just upfront donations in general. If someone gives me something for my work, I'd like to be able to give them something in return for it.... So we'll see what this means ultimately.

Perhaps YAP will get features like the ability to bulk poll machines with a single client, or to do web based monitoring of say, an entire datacenter or a collection of servers. Features like this may be released as a separate branch of YAP or locked down, yet at the same time I don't think the average gamer will consider these beneficial to them either. (if that makes sense)


One thing is for sure, I want the vast majority of functionality to be available free. yet, I also want to be able to offer people something should they want to contribute to the development. So we'll see where that goes.


But to recap, yes, YAP will be absolutely free for a gamer, yet I'll also come up with some cool extras that it can do to show my appreciation if someone wants to throw any money at this project.

Some people call this "freemium". I'm going to call it "froation-ware".. It will NOT be shareware, as it will be full featured as a free software and more powerful than ClockBlocker ever was.

Hopefully that makes sense.