|
slow isn't bad. as long as it's consistent. and, slow testing increases the incentive to make the process even faster!
|
|
|
|
|
so insert a Thread.Sleep(1000000) in the startup code and all speed variations become moot.
Luc Pattyn [Forum Guidelines] [My Articles] Nil Volentibus Arduum
Please use <PRE> tags for code snippets, they preserve indentation, improve readability, and make me actually look at the code.
|
|
|
|
|
|
I was afraid you were thinking that. In my experience, and I have been concentrating on performance for as long as I'm active in electronics and software, optimizing something based on observations under abnormal conditions is most often a waste of time. For instance, the way multi-threaded operations evolve will be different when you switch from a faster to a slower disk, or vice versa.
Side note: think of the highways; when traffic increases and things tend to slow down, they decrease the speed limits to get the cars moving faster. Sounds weird at first, it actually works.
Yes you can try and optimize simple things (say a numeric algorithm, one thread, no blocking calls) under synthetic conditions. However, as soon as your app has some level of complexity (i.e. many threads doing uncorrelated stuff), overall app performance optimization requires realistic conditions to be relevant. And that makes optimization an iterative process: as soon as you make a second decision, you have to question your first decision again. If not, you may end up in a local minimum rather than the overall minimum.
For real observations, a reboot is the only thing I know of that provides the cold start numbers (an alternative I mentioned earlier would be using N identical files, using a different one for each of N runs). If that is too cumbersome, then, as a quick and dirty approach, you could just ignore the first (cold) run and assume all other runs are comparable (which will depend on what other apps your system is also running), but there are no guarantees as I explained before.
Luc Pattyn [Forum Guidelines] [My Articles] Nil Volentibus Arduum
Please use <PRE> tags for code snippets, they preserve indentation, improve readability, and make me actually look at the code.
|
|
|
|
|
yes. optimization is tricky, and iterative, and is sensitive to the input data. i assure you, i've done this before.
Luc Pattyn wrote: optimizing something based on observations under abnormal conditions is most often a waste of time
the only thing abnormal about what i'm doing is that i'm running the same data through the system over and over. our customers will typically not run the same data twice - they'll process a data set once, send the results off, and move to the next data set. the Windows cache data from one job will not help them with the next job. so, in order to better match their experience, i need to eliminate all the bits of lookup data that were cached by Windows during the previous run before i start the next.
|
|
|
|
|
In windows xp, when i just select the folder, it display the total umber of the file and the total size in this folder. But I didn't see this in windows 7. Unless I select all file in this folder then it will show. Are there any option that I can do this in windows 7?
|
|
|
|
|
It is automatically visible in my case...
You should only click on the third icon on the right side of any window (just under the search bar) and select "details".
Apart of that at the bottom there is information of the number of files if you have not selected any of them. Also, if you select one file or more than one file the bottom information will change accordingly.
Hope this helps.
|
|
|
|
|
Joan Murt wrote: You should only click on the third icon on the right side of any window (just under the search bar) and select "details".
thank you for your reply. how about the total size of the file?
|
|
|
|
|
Once you are on the details view, you can right-click any of the column headers and select "size". If there is not an element that suits your needs then you can go to the "More..." element and select it from an extended list.
HTH!
On a side note, of course if you select more than one file you will see this information (size of all of the selected files) in the bottom of the window.
|
|
|
|
|
I don't think this is offered in Windows 7 Explorer; another example of improvements to the product.
I must get a clever new signature for 2011.
|
|
|
|
|
hi,
sir i am a student of final year(3 rd)polytechnic in information technology from the vidt\ya bhawan pollytechnic college udaipur rajasthan .
i want make a software as a "webcam security" so i want to know i shuld choose which of the platreform.
webcam security- the program will be start at the login time and when user enter the password then pess the enter then the wbecam shuld take a image and save the administrator directory.
this all work will be on the hide form.
so the user dont identify .
plz sir tell me information
mob-7737622335
mail ID_krishna.taniya@gmail.com
|
|
|
|
|
krishna.taniya wrote: plz sir tell me information
This is a rather difficult question to answer, since you have given no indication of what language you will be using to develop in, or what level of experience you have. I would suggest you go to the articles page and search for webcam programs as a starting point and proceed from there. But please remember not to submit any of the CodeProject articles as your own work.
I must get a clever new signature for 2011.
|
|
|
|
|
i want to ask that how to place an exe in windows 7 startup so that the user (user with USER RIGHTS- LIMITED ACCOUNT ) can't remove that exe from windows startup only the administrator can remove that exe.
If i place in "startup" folder than user can easily remove it
|
|
|
|
|
It's ok to have it in the startup folder, you just have to set the access rights on the file accordingly.
Rightclick on the file and press properties, then select the security tab and remove all rights except read and execute for the user/user group.
|
|
|
|
|
Depending on the tasks to be performed by your application, a Windows Service could actually be the right way to go.
|
|
|
|
|
Hello all,
I'd like to know how do you cope with this situation:
* Two network groups.
* Some computers in the first group and some computers in the second group.
* No DHCP enabled in any of the groups.
* One ADSL (no wifi) that must work in both groups (192.168.0.1).
* 192.168.0.xxx is the address range in both groups.
The first group has to be able to remote connect (remote access) to computers of the first group.
The second group has to be able to remote connect (remote access) to computers of the second group.
A) Does this mean that I can have only 255 computers in a network?
B) Without a hardware solution, can I have a set of software settings that would allow me to achieve the previous?
Could you give me some pointers/recomendations? and please avoid this: http://xkcd.com/138/[^]
Thank you in advance...
|
|
|
|
|
A) Actually you have a maximum of 252 computers in the network if you're connected to the internet.
0 is the network address and 255 is the broadcast address, and then your router needs one address.
But I think we need a bit more information on what's defining these two groups. As the're supposed to share the ADSL connection it's not a geographical grouping.
So my guess is that you're having two companies that needs to be kept apart while sharing the infrastructure.
Have a look at VLAN[^]. While normally being a hardware solution, there are also software to do the job. I don't know what security implications you can expect though.
|
|
|
|
|
Joan Murt wrote: A) Does this mean that I can have only 255 computers in a network?
No. This range, when coupled with the default subnet mask 255.255.255.0, gives you 254 addresses. But there's no reason you have to use that subnet mask.
Given an address for your network of 192.168.0.0 and a subnet mask of 255.255.254.0 you have 510 addresses available - that is 192.168.0.1 through 192.168.1.254. Visually (if I can render this):
11000000.10101000.00000000.00000000 IP Address = 192.168.0.0
11111111.11111111.11111110.00000000 Subnet Mask = 255.255.254.0
ANDing gives the network address:
11000000.10101000.00000000.00000000 (192.168.0.0)
Hosts can be addressed from
11000000.10101000.00000000.00000001 (192.168.0.1) to
11000000.10101000.00000001.11111110 (192.168.1.254).
The bits to the right of the least significant 1 in the subnet mask are available to be assigned to hosts on the subnet, while the space marked by 1s identifies the network. If you need more, you can always change the mask to 255.255.252.0 or 255.255.248.0, etc. Each step will double the available host addresses.
Members of a workgroup can access other hosts within the same network, and the subnet mask defines what digits in the IP address should be interpreted as the network identifier, and which are host identifiers. If your users need access to the Internet, include a router in your system with a LAN address within the network you've selected, and set the Gateway for each host that requires access to match the router's LAN address.
By the way, not using DHCP will cause you a lot of grief. If you don't have a server in the group to perform that task for you, configure one router to serve as the DHCP Server. Manual configuration of more than 5 computers on a network is a major hassle, not worth anyone's time.
Will Rogers never met me.
|
|
|
|
|
Really good answer, clear and concise. Thank you.
I understand the DHCP bit in which you recommend me to use DHCP (which I see as a positive thing). The only question it comes to my mind is: "how do you get access to a computer if the computer's name doesn't give you access to it?" I mean that sometimes the only effective way to access a computer is trough the IP and not the computer name...
|
|
|
|
|
Oh dear, do you not have a DNS Server in the network?
You can manually add hosts to the Hosts file, but you have to do that for every computer in the network. You can enable NetBIOS over TCP/IP, too, to allow the old Windows network discovery tools to work. Unfortunately that exposes you to a bunch of attack vectors, and it's just a whole lot smarter to avoid all that. Some routers have a DNS Server funcion built-in, along with the DHCP Server functions. If you can't have a real server to control your domain, I'd really recommend installing a router that can do all these things for you.
Will Rogers never met me.
|
|
|
|
|
We are going to be sharing a conference room with another company and would like both parties to be able to schedule the room with only visiblity into the room scheduling.
|
|
|
|
|
You've got a huge problem. Your two companies would have to share the same Exchange server infrastructure, meaning the Exchange servers would have to trust each other users. If you don't want anything other than the conference room being shared in Exchange, you have a massive security effort to keep the seperation in place.
If this is all too much, for a single conference room, put up a scheduling sheet in the room or have a secretary maintain it.
|
|
|
|
|
Maybe you can use an online calendar app just for this conference room like Google calendar
Wout Louwers
|
|
|
|
|
I'm using AnkhSVN and just recently RocketSVN but have the same problem with both.
I want to make some major changes to an application so I want to create a Branch.
The way I normally do this is that I created the Branch then switch and work from the Branch
but whenever I try to merge my changes back into the Trunk am not able for one reason or another.
Should I create the Branch then continue using the Trunk then Merge back if I want to keep changes?
This doesn't seem right to me so I haven't tried it.
I've used SVN quite a bit but I just can't seem to get the Branching through my head for some reason.
How should I/How would you do this?
|
|
|
|
|
that's a very good question!! i always have the same issues with branching and end up making a mess of the merge! ...hopefully someone with real svn expertise can advice here...
|
|
|
|