ATi's CCC: Not the greatest but not that bad either.

[beggerking]

Originally posted by: Matthias99

Originally posted by: beggerking
umm.. isn't that what I said? to the system, vm + memory = memory. hence, whenever you need any piece of info from vm, you need to go look into your paged file on your harddrive..

Matthias99, it looks like you are the one who don't know anything.

not ALL memory access out of the paging system.. its just unnecessary in physical memory because physical memory is fast enough to find info instantenously. paging system applys to vm to speed up fetching information out of harddrive.. it has nothing to do with physical ram.

ALL memory access -- at least in any modern OS that supports paging -- is done via the paging system. Outside of the memory subsystem code, the OS sees one big memory space (usually 4GB in size with a 32-bit operating system), parts of which are paged in and out on the fly as needed if there is not enough physical memory. WinNT-based kernels can actually support more than 4GB of address space, but this is generally a big PITA and requires special application support (through PAE).

Your earlier posts were implying that, somehow, part of CCC would be stored in the swapfile (which you keep referring to as "VM") all the time. This is just not correct.

Please stop trying to blather on about things you don't really understand.

if information is not already in a swapfile(which to the system is physical memory) or in physical memory, your system can't use it..

your reasoning contradict yourself.. all modern system page because they are big and require VM. "page" means to use harddrive as memory as to "map" harddrive spaces as "pages" of physical memory.

for physical memory, its already "paged" and readily accessible.

 

[Matthias99]

Originally posted by: beggerking
if information is not already in a swapfile(which to the system is physical memory) or in physical memory, your system can't use it..

Your CPU can only directly access data that is in RAM (well, actually, it can only directly access data that is in a register, but it has built-in hardware to move data between registers and RAM). Data in the swapfile is, by definition, data that was in RAM and was paged down to the disk.

your reasoning contradict yourself.. all modern system page because they are big and require VM. "page" means to use harddrive as memory as to "map" harddrive spaces as "pages" of physical memory.

for physical memory, its already "paged" and readily accessible.

All modern operating systems support paging because trying to write applications without virtual memory spaces sucks. The paging system is used even if the 'swapfile' is nominally disabled; it's just that in this case the amount of paged memory is limited to the amount of RAM installed in the system, so nothing is ever pushed out to disk. You still get the other benefits of a virtual address space, such as allowing applications to map I/O to certain address ranges, etc. without having to worry about where things are physically placed.

ALL memory allocated by any modern OS is paged. Some pages are in physical RAM, and others will be placed on disk if there is not room. Paging does NOT just mean using your hard drive as virtual memory, although many people who think they know something about VM and OSes assume that.

In any case, your implication that part of CCC will permanently reside in the swapfile (rather than in RAM) is simply absurd.

 

[Josh7289]

Originally posted by: sindows
I'm not sure what you mean by 1:1 pixel scaling but if you're talking about haveing a LCD monitor and you want to display a lower than native res but don't want it to fill up the screen, ATI's cards are certainly able to do it. I'm able to "show" say 800x600 pixels while the rest of the monitor is blacked out. I've done it on ATI's onboard video, X800XL, and X300 using various drivers, the most recent being 6.1. I'd be glad to help you out but I'm not too entirely sure of what problems you have.

It must only work with specific cards, then, as myself and others never got it to work. I had a Sapphire X800GT, and it only worked on a few resolutions with that.

Anyway, I'm using the faster 6600GT now, and this feature works at all resolutions on this card, so I'm happy with it and with Nvidia. Why can Nvidia make it work on all cards, but not ATI? It's not that hard to do...

 

[beggerking]

Originally posted by: Matthias99

Originally posted by: beggerking
if information is not already in a swapfile(which to the system is physical memory) or in physical memory, your system can't use it..

Your CPU can only directly access data that is in RAM (well, actually, it can only directly access data that is in a register, but it has built-in hardware to move data between registers and RAM). Data in the swapfile is, by definition, data that was in RAM and was paged down to the disk.

your reasoning contradict yourself.. all modern system page because they are big and require VM. "page" means to use harddrive as memory as to "map" harddrive spaces as "pages" of physical memory.

for physical memory, its already "paged" and readily accessible.

All modern operating systems support paging because trying to write applications without virtual memory spaces sucks. The paging system is used even if the 'swapfile' is nominally disabled; it's just that in this case the amount of paged memory is limited to the amount of RAM installed in the system, so nothing is ever pushed out to disk. You still get the other benefits of a virtual address space, such as allowing applications to map I/O to certain address ranges, etc. without having to worry about where things are physically placed.

ALL memory allocated by any modern OS is paged. Some pages are in physical RAM, and others will be placed on disk if there is not room. Paging does NOT just mean using your hard drive as virtual memory, although many people who think they know something about VM and OSes assume that.

In any case, your implication that part of CCC will permanently reside in the swapfile (rather than in RAM) is simply absurd.

Yes , CPU can only access register via system memory. hence any information is cached to memory from swap file before it was accessed.

so what do you say "page" is used for? its for easy access / fast access to swapfile. of course you need a page of 'links" in your physical memory point to all pages of swapfile in your hd, but memory(physical) doesn't need to be paged..

your definition of paging = "the pages of links"

Yes, Swapfile are permanent on hd until its erased.

 

[Matthias99]

Originally posted by: beggerking
so what do you say "page" is used for? its for easy access / fast access to swapfile. of course you need a page of 'links" in your physical memory point to all pages of swapfile in your hd, but memory(physical) doesn't need to be paged..

your definition of paging = "the pages of links"

You just really don't seem to have a good grip on the whole concept of paged virtual memory.

Read the following, slowly:

In any modern OS, all memory allocated by the OS is allocated through the paging system. The virtual memory manager decides which pages are placed in physical RAM and which ones are "paged out" to the swapfile (if one is available). Applications do not normally have any say in these decisions.

Yes, Swapfile are (sic) permanent on hd until its (sic) erased.

While I suppose this is true by definition (hard drives are nonvolatile, so anything written on them is permanent until erased), it has very little bearing on this topic. This:

I see most part of the CCC is allocated in VM, therefore most of the time when you use CCC, it will have to go look for the information in harddrive, hence, slow system performance...

Makes absolutely no sense. The pages for CCC may, in fact, be in the swapfile if it has not been used in a while and you have a lot of programs open. But its pages will be brought back into physical RAM when you access them, and will stay there until they are forced out by something else. It is no different than any other application in this regard.

 

[Powermoloch]

I have no issues with CCC. But one thing that really saddens me that, in Nvidia, their control panel can select AA/AF on the games that I want (within the control panel list ) which I find it pretty sweet. ATi's didn't ....which makes me

 

[BFG10K]

I'm surprised this troll trainwreck thread is still going.

I've got ~2000 MB of system memory, 17 MB is a whole 1% of that, oh my god the bloatness!!

I've got 2 GB too but that makes no difference. The point is that other utilities are faster and more feature rich but take up less RAM, disk space and loading time. ATi states CCC's reason for .NET is for other utilites to plug in but this is BS because ATi Tray Tools has plug-in support too and it's a lean program to boot.

CCC is unnecessarily bloated and slow.

Don't worry BFG, nobody is going to take your "rights" away, you can continue to use your nV control panel and be content in excercising your right to use 3 MB.

Thanks, I will. I'll also continue to slam troll threads like this one.

 

[clickynext]

No problem with CCC for me, I only need to go into it once in a while (a long while), and it doesn't crash anymore so all's good now.

 

[beggerking]

In any modern OS, all memory allocated by the OS is allocated through the paging system. The virtual memory manager decides which pages are placed in physical RAM and which ones are "paged out" to the swapfile (if one is available). Applications do not normally have any say in these decisions.

please re-read ... I never said app has anything to do with how virtual memory are allocated... virtual memory are allocated by OS's memory manager which is itself an application. So any application under OS is actually using memory allocated by the OS. the memory manager looks in physical memory first. If its not there , it looks up the "page of link" which links to the swap file. But if the application itself is too big to fit in physical memory, or is requesting memory for whatever reasons(which it can) that is too big, the OS will have to allocate vm for the app. In other words, CCC is a bloated program that uses too much memory or is itself too big.

--------------------------------------------------------------------------------
Yes, Swapfile are (sic) permanent on hd until its (sic) erased.
--------------------------------------------------------------------------------

While I suppose this is true by definition (hard drives are nonvolatile, so anything written on them is permanent until erased), it has very little bearing on this topic

so that proves to you, swap files are required for virtual memory allocation which in another post you said :

Your earlier posts were implying that, somehow, part of CCC would be stored in the swapfile (which you keep referring to as "VM") all the time. This is just not correct.

.

--------------------------------------------------------------------------------
I see most part of the CCC is allocated in VM, therefore most of the time when you use CCC, it will have to go look for the information in harddrive, hence, slow system performance...
--------------------------------------------------------------------------------

Makes absolutely no sense. The pages for CCC may, in fact, be in the swapfile if it has not been used in a while and you have a lot of programs open. But its pages will be brought back into physical RAM when you access them, and will stay there until they are forced out by something else. It is no different than any other application in this regard.

what I said makes every sense: it is in swap file until you shutdown CCC. and as I have bolded above, which will access harddrive. Since hd is so slow, it bring down system performance.

looks like you just repeated exactly what I just said...

 

[xtknight]

Virtual memory is solely an index system used by the operating system to address physical or page file memory, as needed.

It was explained well in the MSDN article.

In modern operating systems, including Windows, application programs and many system processes always reference memory using virtual memory addresses which are automatically translated to real (RAM) addresses by the hardware. Only core parts of the operating system kernel bypass this address translation and use real memory addresses directly.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

 

[beggerking]

yes. . to address the lack of physical memory by indexing hd spaces to be used as physical memory.

 

[xtknight]

Originally posted by: beggerking
yes. . to address the lack of physical memory by indexing hd spaces to be used as physical memory.

No. To address ANY kind of memory. Physical memory does not need to be lacking for virtual memory to be in use.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

Virtual memory maps a virtual code (which means NOTHING in terms of direct operation) to an address that points to the exact location the memory is to be read or written. When the data is paged out of the physical memory to the page file, the application can use the same virtual memory address, but the virtual memory will now be mapped to the page file instead, where less-accessed memory is stored. When the application is in use again, the memory is paged back out of the page file into physical memory where the application can access it faster. All throughout these processes the same virtual memory 'code' remains the same to redirect the application to whereever the data may be located.

Virtual memory is not the same as a page file or swap file, nor does it only contain addresses for page/swap files.

 

[Kalessian]

Originally posted by: BFG10K
I'm surprised this troll trainwreck thread is still going.

I've got ~2000 MB of system memory, 17 MB is a whole 1% of that, oh my god the bloatness!!

I've got 2 GB too but that makes no difference. The point is that other utilities are faster and more feature rich but take up less RAM, disk space and loading time. ATi states CCC's reason for .NET is for other utilites to plug in but this is BS because ATi Tray Tools has plug-in support too and it's a lean program to boot.

CCC is unnecessarily bloated and slow.

QFT

 

[beggerking]

Originally posted by: xtknight

Originally posted by: beggerking
yes. . to address the lack of physical memory by indexing hd spaces to be used as physical memory.

No. To address ANY kind of memory. Physical memory does not need to be lacking for virtual memory to be in use.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

Virtual memory maps a virtual code (which means NOTHING in terms of direct operation) to an address that points to the exact location the memory is to be read or written. When the data is paged out of the physical memory to the page file, the application can use the same virtual memory address, but the virtual memory will now be mapped to the page file instead, where less-accessed memory is stored. When the application is in use again, the memory is paged back out of the page file into physical memory where the application can access it faster. All throughout these processes the same virtual memory 'code' remains the same to redirect the application to whereever the data may be located.

Virtual memory is not the same as a page file or swap file, nor does it only contain addresses for page/swap files.

Yes. exactly what I said in above post "page of links" which is stored in physical memory that links to memory allocations in pagefiles. My earlier post also mentioned that once the link(address) is found in page file, it pages the information back into physical memory for use. Mathia didn't quote that part but its in the previous page.

though I disagree with the "always in use " part, as I always thought the system looks in physical memory for info before it looks into that table. But since its mentioned in MSDN I say you are right..

 

[xtknight]

Originally posted by: beggerking
Yes. exactly what I said in above post "page of links" which is stored in physical memory that links to memory allocations in pagefiles. My earlier post also mentioned that once the link(address) is found in page file, it pages the information back into physical memory for use. Mathia didn't quote that part but its in the previous page.

though I disagree with the "always in use " part, as I always thought the system looks in physical memory for info before it looks into that table. But since its mentioned in MSDN I say you are right..

Virtual memory is different from what Windows 98 called virtual memory in its settings so it is an inconsistent term between the OSes. The page file definitely isn't always in use.

 

[beggerking]

Originally posted by: xtknight

Originally posted by: beggerking
Yes. exactly what I said in above post "page of links" which is stored in physical memory that links to memory allocations in pagefiles. My earlier post also mentioned that once the link(address) is found in page file, it pages the information back into physical memory for use. Mathia didn't quote that part but its in the previous page.

though I disagree with the "always in use " part, as I always thought the system looks in physical memory for info before it looks into that table. But since its mentioned in MSDN I say you are right..

Virtual memory is different from what Windows 98 called virtual memory in its settings so it is an inconsistent term between the OSes. The page file definitely isn't always in use.

so virtual memory points to null addresses when its not required? but it still exists in memory?

 

[xtknight]

Originally posted by: beggerking

Originally posted by: xtknight

Originally posted by: beggerking
Yes. exactly what I said in above post "page of links" which is stored in physical memory that links to memory allocations in pagefiles. My earlier post also mentioned that once the link(address) is found in page file, it pages the information back into physical memory for use. Mathia didn't quote that part but its in the previous page.

though I disagree with the "always in use " part, as I always thought the system looks in physical memory for info before it looks into that table. But since its mentioned in MSDN I say you are right..

Virtual memory is different from what Windows 98 called virtual memory in its settings so it is an inconsistent term between the OSes. The page file definitely isn't always in use.

so virtual memory points to null addresses when its not required? but it still exists in memory?

The virtual memory points to physical memory when the page file is not required.

 

[beggerking]

Originally posted by: xtknight

Originally posted by: beggerking

Originally posted by: xtknight

Originally posted by: beggerking
Yes. exactly what I said in above post "page of links" which is stored in physical memory that links to memory allocations in pagefiles. My earlier post also mentioned that once the link(address) is found in page file, it pages the information back into physical memory for use. Mathia didn't quote that part but its in the previous page.

though I disagree with the "always in use " part, as I always thought the system looks in physical memory for info before it looks into that table. But since its mentioned in MSDN I say you are right..

Virtual memory is different from what Windows 98 called virtual memory in its settings so it is an inconsistent term between the OSes. The page file definitely isn't always in use.

so virtual memory points to null addresses when its not required? but it still exists in memory?

The virtual memory points to physical memory when the page file is not required.

thanks for clear that up.

 

[Matthias99]

Originally posted by: xtknight

Originally posted by: beggerking
yes. . to address the lack of physical memory by indexing hd spaces to be used as physical memory.

No. To address ANY kind of memory. Physical memory does not need to be lacking for virtual memory to be in use.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

Virtual memory maps a virtual code (which means NOTHING in terms of direct operation) to an address that points to the exact location the memory is to be read or written. When the data is paged out of the physical memory to the page file, the application can use the same virtual memory address, but the virtual memory will now be mapped to the page file instead, where less-accessed memory is stored. When the application is in use again, the memory is paged back out of the page file into physical memory where the application can access it faster. All throughout these processes the same virtual memory 'code' remains the same to redirect the application to whereever the data may be located.

Virtual memory is not the same as a page file or swap file, nor does it only contain addresses for page/swap files.

Thank you. Someone else actually understands what the hell I'm talking about.

Getting slightly back on topic:

Since I'm rebuilding my HTPC (which has a RADEON 9600 installed), and I hadn't installed the video drivers for it yet, I thought I'd take a look and see what happens to memory usage with Catalyst / CCC installed.

Without Catalyst/CCC or .NET installed, sitting at the desktop, my commit charge was about 115M, and the three kernel memory usage counts were about 27M/20M/6.8M.

After installing .NET 1.1/2.0:

Commit charge was around 120M, and the kernel memory usage was about the same.

After installing Catalyst 6.1 with CCC (just sitting at the desktop):

Commit charge jumped up to around 195K(!), and kernel memory usage was not about 34.5M/22.3M/12M. That's a pretty big jump, so I started looking at it more carefully.

With CCC actually open (which only took about three seconds the first time and maybe 1-2 seconds after that):

Commit charge went up to about 210M, with no change in kernel memory usage.

With the 3D preview window open:

Commit charge was about 230M.

Disabling the system tray icon cut the commit charge to ~185M with CCC closed, 190M with it open, and 210M with the preview window up. However, this still seemed high.

Following a tweak guide I found, I disabled the two ATI services and the 'CLI' program it starts up when your system boots (I also kept the system tray icon disabled).

Doing that dropped the commit charge to about 130M, with kernel memory usage of 30M/19M/11M. This seemed more in line with what the driver itself should use; disabling these services keeps it from loading CCC/.NET into memory when you start up the system.

At that point, starting up CCC took about 10-15 seconds, and bumped the commit charge up to the same point as before. When I closed CCC, commit charge dropped to about ~165M and stayed there (I'm guessing that some of the .NET stuff stays resident once it is loaded the first time). This would peg CCC (plus various .NET things it is using) at ~50-80MB of memory usage (with most of that difference being the 3D preview window), with the actual Catalyst drivers at somewhere around 10-20MB.

Now, 'Commit Charge' measures how much virtual memory has been allocated on behalf of various processes running in the system. While CCC by default sets aside a whole bunch of memory, it's not actively using those memory pages unless CCC is actually open. These pages will just get pushed out to the swapfile if there is not enough room in your physical memory, and it should not actually impact performance very much.

That said, an extra 80MB of memory usage just so you can load up CCC at a moment's notice seems a little excessive. It would be nice to have an easier way to disable this than having to manually turn off the ATI startup services.

 

[flexy]

Originally posted by: 5150Joker
With so many people claiming how slow and abysmal CCC is, I thought I'd make a quick video illustrating that it can load under 10 secs on your PC

you must be totally kidding me !!!!

Thats almost 7 seconds or so just for the panel to come up....nice proof how inefficent and what a resource-waster it is. I am with CP or ATT for a long time now and this video just shows me again i will NEVER touch CCC.

 

[flexy]

just as a sidenote, and how many threasds/posting do we have regarding this now ?

As a pretty PC savvy person i do not need an "application" which "demonstrates" me the effect of what AA or AF actually does.
I know that since the beginings of 3D graphics on my computer.

ATI just did a terrible job inventing this and NOT providing an alternative / NOW even abandoning CP....it is STILL only an overlay-application which BASICALLY does nothing else to write some registry-values (for driver options).

This does make as much sense as an "overlay/helper" application for people who use windows which comes up with a demonstration explaining what a double-click is, how to close and move windows around...and other basics stuff which might be interesting for a beginner...but for sure not "forced" on someone who uses windows for years.

 

[beggerking]

Originally posted by: Matthias99

Originally posted by: xtknight

Originally posted by: beggerking
yes. . to address the lack of physical memory by indexing hd spaces to be used as physical memory.

No. To address ANY kind of memory. Physical memory does not need to be lacking for virtual memory to be in use.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

Virtual memory maps a virtual code (which means NOTHING in terms of direct operation) to an address that points to the exact location the memory is to be read or written. When the data is paged out of the physical memory to the page file, the application can use the same virtual memory address, but the virtual memory will now be mapped to the page file instead, where less-accessed memory is stored. When the application is in use again, the memory is paged back out of the page file into physical memory where the application can access it faster. All throughout these processes the same virtual memory 'code' remains the same to redirect the application to whereever the data may be located.

Virtual memory is not the same as a page file or swap file, nor does it only contain addresses for page/swap files.

Thank you. Someone else actually understands what the hell I'm talking about.

Getting slightly back on topic:

Since I'm rebuilding my HTPC (which has a RADEON 9600 installed), and I hadn't installed the video drivers for it yet, I thought I'd take a look and see what happens to memory usage with Catalyst / CCC installed.

Without Catalyst/CCC or .NET installed, sitting at the desktop, my commit charge was about 115M, and the three kernel memory usage counts were about 27M/20M/6.8M.

After installing .NET 1.1/2.0:

Commit charge was around 120M, and the kernel memory usage was about the same.

After installing Catalyst 6.1 with CCC (just sitting at the desktop):

Commit charge jumped up to around 195K(!), and kernel memory usage was not about 34.5M/22.3M/12M. That's a pretty big jump, so I started looking at it more carefully.

With CCC actually open (which only took about three seconds the first time and maybe 1-2 seconds after that):

Commit charge went up to about 210M, with no change in kernel memory usage.

With the 3D preview window open:

Commit charge was about 230M.

Disabling the system tray icon cut the commit charge to ~185M with CCC closed, 190M with it open, and 210M with the preview window up. However, this still seemed high.

Following a tweak guide I found, I disabled the two ATI services and the 'CLI' program it starts up when your system boots (I also kept the system tray icon disabled).

Doing that dropped the commit charge to about 130M, with kernel memory usage of 30M/19M/11M. This seemed more in line with what the driver itself should use; disabling these services keeps it from loading CCC/.NET into memory when you start up the system.

At that point, starting up CCC took about 10-15 seconds, and bumped the commit charge up to the same point as before. When I closed CCC, commit charge dropped to about ~165M and stayed there (I'm guessing that some of the .NET stuff stays resident once it is loaded the first time). This would peg CCC (plus various .NET things it is using) at ~50-80MB of memory usage (with most of that difference being the 3D preview window), with the actual Catalyst drivers at somewhere around 10-20MB.

Now, 'Commit Charge' measures how much virtual memory has been allocated on behalf of various processes running in the system. While CCC by default sets aside a whole bunch of memory, it's not actively using those memory pages unless CCC is actually open. These pages will just get pushed out to the swapfile if there is not enough room in your physical memory, and it should not actually impact performance very much.

That said, an extra 80MB of memory usage just so you can load up CCC at a moment's notice seems a little excessive. It would be nice to have an easier way to disable this than having to manually turn off the ATI startup services.

umm... what you said has nothing to do with what he said... oh well..

 

[beggerking]

and performance is impacted....

 

[Matthias99]

Originally posted by: beggerking
umm... what you said has nothing to do with what he said... oh well..

Other than that he was agreeing with me and you're still trying to say something else (though what, I'm not quite sure).

and performance is impacted....

Not really. If it's in physical RAM, it's there because the space isn't needed by anything else right now. In this case, there would be no performance difference whether or not you had it running.

If the pages for CCC have been forced out to the swapfile, then whatever forced them out is now using that space in physical RAM. In that case, you'll have a performance penalty the next time you go to access CCC, but your other applications won't be affected by it having allocated space in memory. There would be a small performance hit while the data is being paged out, but that's pretty minor, and a one-time occurence unless you keep accessing CCC all the time.

 

[ArchAngel777]

Originally posted by: Matthias99

Originally posted by: beggerking
umm... what you said has nothing to do with what he said... oh well..

Other than that he was agreeing with me and you're still trying to say something else (though what, I'm not quite sure).

and performance is impacted....

Not really. If it's in physical RAM, it's there because the space isn't needed by anything else right now. In this case, there would be no performance difference whether or not you had it running.

If the pages for CCC have been forced out to the swapfile, then whatever forced them out is now using that space in physical RAM. In that case, you'll have a performance penalty the next time you go to access CCC, but your other applications won't be affected by it having allocated space in memory. There would be a small performance hit while the data is being paged out, but that's pretty minor, and a one-time occurence unless you keep accessing CCC all the time.

You know, I really havn't researched virtual memory to the Nth degree... So my money is on Mathias and XTKnight, both of whom should be Elite at this point... After seeing the thread Beggerking created flaming another user, LOL, I don't think I'd trust anything he says at this point.