|
Hi all,
I am trying to understand how windows allocates processor to programs. Maybe this is not a specific VC++ question, but my application is in VC++.
Coming to details, I am having a dual processor system with single core, two threads each. So, in task manager I can see 4 processors. I usually run one MFC application and one matlab application (Interface in matlab and the actual processing is done in VC++ using a DLL created in VC++). My DLL is having 2 child threads created within itself whenever it is initialized from GUI created in Matlab(so basically its a Interface created in Matlab), and my MFC application is having one child thread apart from its own.
So, my question is, when I run all these at a time; how do you think windows allocates the 4 processors?? From the task manager I can see it allocating 1 processor to MFC application, but not sure about the DLL. Any ideas?? The main reason behind this is, I am thinking of replacing one of our single processor system with a new one which should handle all the above mentioned applications, so thinking of which will be best match, a dual processor or a Dual core single processor (like conroe) system and at the sametime I need high PCI bus transfers too.
thanks,
-Pavan.
|
|
|
|
|
Firstly you need to understand that Windows allocates processor time to threads, not to processes or programs. Each time the scheduler is called upon to reschedule which threads are running, it looks at only the set of threads which are runnable - not blocked waiting for some event to occur. GUI threads are typically waiting for the user to press a key or move the mouse over the window, or for some other window message to arrive. This can also be a timer event created by the SetTimer function. Whenever you call a synchronous I/O function such as ReadFile, if the data isn't already buffered in the system's cache, the OS begins a read from the disk, then blocks the thread until the data arrives, and it's not runnable. If you use the WaitForSingleObject or {Msg}WaitForMultipleObjects{Ex} functions, your thread blocks until one of the conditions is satisfied. Blocking for I/O can happen without any specific command from the programmer if the program references data that isn't currently in memory (whether to be read in from the original program files or from the page file).
The scheduler considers the set of runnable threads in priority order, from highest priority (31) to lowest (0). It only considers threads from the highest priority at which there are runnable threads, unless there are fewer threads at the highest priority than there are processors, in which case it will consider threads at the next lowest priority level which has runnable threads. This means that if a high priority thread never blocks (and therefore stops being runnable), it can prevent lower priority threads from getting CPU time. Best practice is to simply use the Normal priority, the default. The scheduler maintains a queue of threads at each priority level, and simply (if all default values are used) assigns the first N threads in the queue to each of the N logical processors in the system.
The scheduler is invoked when one of three things happen: a running thread's quantum (time permitted to run) expires; a running thread blocks; some other event in the system causes a higher-priority thread to unblock. Windows raises priorities for some unblocking events which cause a thread to temporarily have a higher priority, which increases responsiveness.
This is complicated by affinity (where a thread can be prevented from running on certain logical processors), and by the fact that Windows allocates an ideal processor for each thread, and remembers the last processor that a thread ran on. Windows tries to run a thread on the last processor it ran on, then its ideal processor, to take advantage of the processor's cache - with luck, the code and data that the thread referenced last time it ran (and hopefully will reference again this time) will still be in the processor's cache and will therefore not have to be retrieved from main memory.
Hyperthreaded processors are a bit odd. The execution resources of a single core are contended for by two logical streams of execution. If both threads scheduled on a hyperthreaded core require lots of resources, they will get in each others' way. Windows XP and Server 2003 understand HT and will ensure that one thread is running on the first logical processor on each core before starting to use the second. They also understand multicore processors and will try to schedule threads so that cache thrashing - where the data referenced by one thread causes the data needed by another thread to be removed from the cache - doesn't occur. Finally, both understand NUMA (Non-Uniform Memory Access) systems and use knowledge of which node particular memory is connected to in order to choose a processor which is near the memory that the thread needs to use.
As for bus transfers, PCI is much slower than the processors' connection to the chipset or to memory. On (standard) Intel systems, all processors share a single front-side bus connecting the processors to the main chip of the chipset which connects to memory (North Bridge or Memory Controller Hub). AMD Athlon 64 and Opteron processors include the memory controller in the processor chip and connect directly to RAM, so multiple socket AMD systems are NUMA.
Best practice as a programmer is to make threads block whenever there isn't anything for them to do. If a thread is left runnable it will end up wasting processor time that could have been used by a thread that would do something useful.
So for your example the way that the processors are used depends very much on the exact way in which the threads block and unblock. As a guide, consider adding more CPU resources if the % Processor Time counter (CPU Usage in Task Manager) on your single processor system is consistently above 80%.
Reference: "Windows Internals, Fourth Edition" by Mark Russinovich and David Solomon.
|
|
|
|
|
Hi all,
I have an obscure problem happening in my main data store folder. The folder contains 1136 subdirectories with a total of 2,397,000 or so files spread out over 121GB on a 1TB Raid 5 array. What’s happening appears to be a cache problem. The symptoms are as follows:
A file (lets say ..Folder1\Inv1.tif , 65KB) is loaded then closed.
The Directory containing the file then has its name changed, ..Folder2\Inv1.tif
I can still access ..Folder1\Inv1.tif from any computer that has opened it from that path; however any new computer trying to open that path fails.
The full path is a network share. The location of the accessing computer doesn’t seem to make a difference, Localhost, lan computer or workgroup computer all have the same behavior. I not sure if it is true for Mapped drives and local drives.
My hunch is that the algorithm for sensing cache invalidations is having troubles iterating through such a large folder.
I need some ideas on what the problem might be and a way to test it.
Ronald Hahn, CNT - Computer Engineering Technologist
New Technologies Analyst
HahnTech Affiliated With Code Constructors
Edmonton, Alberta, Canada
Email: rhahn82@telus.net
|
|
|
|
|
havn't tested but this could be it. still looking for ideas
http://support.microsoft.com/kb/839272/en-us[^]
Ronald Hahn, CNT - Computer Engineering Technologist
New Technologies Analyst
HahnTech Affiliated With Code Constructors
Edmonton, Alberta, Canada
Email: rhahn82@telus.net
|
|
|
|
|
Hi, is there any function that could convert Intel celeron speed to Intel pentium speed? For example if I have 2Ghz speed of intel celeron, what is the real speed in Intel pentium?
|
|
|
|
|
This is really a question with no meaning. The differences between the Celeron and Pentium families are in two areas: clock speed of the Front Side Bus, and amount of on-processor cache.
The clock speed of the front-side bus governs how quickly data can be retrieved from the rest of the system, for example, from main memory. A faster FSB allows faster retrieval from RAM, but only up to the speed the RAM is actually capable of.
The CPU does not communicate with the RAM directly but instead to a component of the chipset called the North Bridge in systems where the PCI bus was used to connect components of the chipset, and which Intel now calls the Memory Controller Hub - the bus between chipset chips is now something called Inter-Hub Transport on Intel 8xx and 9xx series chipsets. To fetch data from or store data to memory a request is placed on the front-side bus targetted at the memory controller, which then performs the operation and returns the results in another bus transaction. This allows the CPU to make other requests while it's waiting for the memory to return the results - while the quoted bus speed of the RAM connection may approach or even exceed that of the FSB, there is always a delay (latency) between sending the command and getting the response, and the quoted speed only relates to a single burst of data from the RAM, which is usually only enough to fill a single cache line - if I recall correctly this is 64 bytes on current processors.
To avoid the delays of going to main memory, since even the front-side bus runs at only a fraction of the core clock speed, the processor has a lot of cache memory on it. This memory is of a very fast type which can run at a significant fraction of core clock speed or even at core clock speed. For an operation that fits completely in the processor's cache, a Celeron and Pentium of the same generation with the same core clock speed should perform equivalently. Having more cache will allow some operations to fit into the cache that wouldn't otherwise. This effect is most noticeable on applications that are written with careful tuning to take advantage of the cache.
You will normally notice that a Celeron system is perceptibly slower than an equivalent core-clock-speed Pentium, but it's not something that can be quantified in GHz. The only thing you can do is run a benchmark that's close to what you want to do with the machine and see what the relative performance is. You could try one of the synthetic application benchmarks like SiSoft Sandra[^].
|
|
|
|
|
Thank you very much for your comment. I ask this question because I have one computer with Intel Celeron 2 Ghz and I notes that when I run the computer in a long time, the speed is become slower and slower. The RAM is 128Mb and I'm running windows xp pro. Do you think if I upgrade the RAM to 2Gb (which is supported by the mainboard), my computer will run faster?
Currently I want to purchase a laptop TOSHIBA which have 256Mb RAM and CPU Intel Celeron 1.6Ghz and running windows xp pro. I plan to upgrade its RAM to 1GB. Do you think this laptop would run fasther if I upgrade the RAM to 1Gb?
|
|
|
|
|
128 meg is barely enough for the OS, 256 isn't much better if your multitasking goes beyond keeping solitare and minesweeper open at the same time. In both cases, upgrading to a gig should show a major gain from not needing the pagefile any longer. The second will only help if you're using your system hard enough to go past the 1 gig level. Check the commit charge numbers on taskmanager.
|
|
|
|
|
Hi, If i record a video for one minute, which file format that is require less disk space with good quality picture (*.wmv, *.mpg, *.rmvb, *.mov, *.mpeg, *.dat or ... other)?
And if I record a sound for one minute, which sound format that require less disk space (*.wav, *.mp3, *.wma or ... ohter)?
|
|
|
|
|
Some of the formats you list (e.g. WAV) are containers - they can contain many different types of encoded data, with an indicator to show what format the contents are in. You can have a .WAV containing MPEG I Layer-3 audio data, for example. Even for the other formats, the codec used to perform the compression normally has configurable options to allow you to select how much compression you want.
When most people think of WAV files, they're generally thinking of uncompressed PCM data, which is pretty much the worst choice in terms of space, but requires virtually no processing overhead.
That said, WMA and AAC are generally considered better quality for a given file size/bandwidth than MP3, while MPEG-4, WMV and H.264 are generally considered better quality than MPEG-1 or MPEG-2. HD DVD discs can be encoded with VC-1 (basically a slightly modified version of Microsoft's WMV9), H.264, MPEG-4 or MPEG-2. The greater compression you want, though, the longer it takes to compress, so it may not be able to compress in real time. This may mean it's better to record in a less compressed format to begin with then convert later to save disk space or reduce download times.
|
|
|
|
|
Mike Dimmick wrote: That said, WMA and AAC are generally considered better quality for a given file size/bandwidth than MP3
Thank you very much for your comment. I also found that WMA could record a sound for a long time and consume less diskspace. I found that in my tablet pc, but the sound recording seem not so clear because some people are sitting far away location. I used it to record the sound in the meeting.
|
|
|
|
|
The file format has nearly no bearing on the size of the file. What affects the size the most are the codecs used to compress (or encode) the data and the bit rates selected for the encoding. A higher bitrate will (usually) result in a higher quality playback, but also increases the size of the file.
Dave Kreskowiak
Microsoft MVP - Visual Basic
|
|
|
|
|
Thank you very much for your comment. If the bitrate is importance, why there are many difference sound format available today? Why they do not create only one file extentsion (Example WAV) and just increase or decrease its bitrate to change the quality of the sound and disk space usage?
|
|
|
|
|
I said the bitrate AND the codec being used to encode it.
The file format, and it's extension, doesn't have any bearing on the size of the file, or very little. Why isn't there just one extension?? Because everybody comes up with their own "container format" to differentiate their product from everyone elses.
Dave Kreskowiak
Microsoft MVP - Visual Basic
|
|
|
|
|
Thank you very much for your explanation. Have a nice weekend 
|
|
|
|
|
how to Enable/ disable FTP access to a system in the network
Thanks in Advance
|
|
|
|
|
Not really a lot to go on there.
All I had to do was make sure that the FTP service was installed and running.
Make sure that you don't expose it to the outside world though. FTP sites that allow annon connections will be sniffed out and taken over quicker than you can say 'free FTP site'.
I did it once just to see what would happen and before I could Finnish one piece of pizza someone had started uploading ripped DVDs to the site and had taken it over.
|
|
|
|
|
Ray Cassick wrote: I did it once just to see what would happen and before I could Finnish one piece of pizza someone had started uploading ripped DVDs to the site and had taken it over.
And were the DVD's any good ? 
Steve S
Developer for hire
|
|
|
|
|
I have to admit that I did wait for them to finish and checked them out. Really bad rips and not in English so I deleted them.
They did manage to build a folder in the FTP directory that I could not delete though. They named it COM1 and the OS would not let me kill it.
|
|
|
|
|
Ya Previously i had the FTP
But now some setting are changed ie.. IP/ permissions/
what can it be
Thanks in advance
|
|
|
|
|
Hi all:
I'd like to avoid the reboot while installing the MSI package, I have some clues but they are not worked as expected.
MsiInstallProduct(szPath, _T("REBOOT=ReallySuppress"))
After executing the above line the system rebooted without any warning. How can I really suppress the reboot?
|
|
|
|
|
I noticed that when I install programs such as Skype or MSN it adds some registry entries to HKCR (HKEY_CLASSES_ROOT), and this was done even if I'm in regular user (not power user or administrator). How is this possible? I though only administrator or power user have access to add this key?
|
|
|
|
|
Since Windows 2000, the HKEY_CLASSES_ROOT key is a merged view of HKEY_LOCAL_MACHINE\Software\Classes and HKEY_CURRENT_USER\Software\Classes. The user's settings override the machine settings. You can therefore make a setting in HKCU which the user can write to. It's disappointing that more installers don't do this. Windows Installer knows about the difference and will write to the user's classes keys as appropriate if doing a per-user rather than per-machine install.
The user's settings are still machine-specific - they are stored in the UsrClass.dat file in the user's Local Settings\Application Data\Microsoft\Windows folder. Therefore, even for a roaming profile, they don't roam.
|
|
|
|
|
dude_4453169 wrote: I'm in regular user (not power user or administrator)
Hi, as far as I know, a simple user do not have a user right to install any application. What is the group membership that your user account belong to?
|
|
|
|
|
Hello all, sorry if this is the wrong forum but I've no idea where else to ask.
I've created an ActiveX control using a 3rd party's tools (QT). To distribute the control I create an .inf file and wrap the .dll(the control) and the .inf within a .cab file. In this way, the control is downloaded from the web much the same as the Flash control. This all works great, and distribution does not seem to be an issue. The problem is, when I look within the 'c:\windows\downloaded program files' directory, my control is listed there, but its status is always 'Unknown'. It's driving me crazy trying to figure it out. I've torn apart the registry and looked into everything I could think of, but nothing seems to help.
So my question: does anyone know how the 'Downloaded Program Files' folder determines if a control is 'installed'?
I'd love to ignore the problem, but I want to be able to update the application using '#version' in my OBJECT tag's CODEBASE property, and it doesn't seem to work if the 'status' is 'unknown'.
Anybody have an ideas?
Thanks in advance,
Matt
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
