|
hey leon de boer
Quote: I can highly recommend learning the whole process on a simple software renderer
I`m familiar to DirectX, I find raytracing intriguing, I want to give it a try.
|
|
|
|
|
Write the code AND THEN refactor (based on analysis of the actual code.)
|
|
|
|
|
That is the agile spirit! Do something quick and dirty, AND THEN discover that it was rather silly...
I admire fearless_ for wanting to do a prestudy do determine the cost of various alternatives, to rule out the obviously invalid ones. Then he dan make a small testbed for the candidates, and from that choose the one looking most promising, and then start coding.
I know very well that agile principles demand that you start your "int main(int argc, char *argv[]) {return 0;}" an make sure it compiles, as a prerequisite before you start asking what the problem is, that you are on your way to solving that problem that hasn't even been ideitified yet. I am so old styel that I prefer to tanke a look at both problem definition and tools available before I race along in that "implementation before problem understanding" race. I know that I am a silly old fool, thinkoing that way
|
|
|
|
|
This has NOTHING to do with agile. This is a tried and true principle of software engineering; first solve the problem and then refactor and optimize.
|
|
|
|
|
I'm sorry for stepping on your toes.
Obiviously, "Code first, refactor later" is not at all related to agile coding.
|
|
|
|
|
I suppose you might find benchmarks on the web. At least you may perform some tests yourself.
|
|
|
|
|
Are you really sure that it matters?
It might, in some very special applications. But most developers tend to overestimate the relative fraction of "calculation" in their code. The data handling often takes up a far larger fraction of the processing time that they thought.
Also: How often do you have a real choice? If your algorithm calls for a sin(), is there anything you can do to avoid it? Not very often! If you do have a choice, implementations vary so much that you should set up a simple test bed and try it out.
One thing I have learnt through simple timing tests: On any CPU architected the last 30 years, timing individual instructions is more or less meaningless. Prefetching and pipelining and speculative execution and whathaveyou can make addition of a simple instruction (e.g. arithmetic) almost unnoticable on the execution time of a tight loop. So don't worry about those.
What really matters is not the execution time of the trig functions (and other mathematiclly complex operations) in themselves, but the number of times you call them. I never programmed any ray tracing myself, but I would not be surprised if there is a lot to be saved in caching, reusse and clever prioritizing to make the most visible effects appear first, details only if you have got the time to do it, before the next update.
|
|
|
|
|
There are lists of the timing properties (not "the time it takes", it's more complicated and you must at least distinguish latency and throughput) of individual instructions for particular CPUs, for example
sqrt more or less corresponds to the similarly named instructions (fsqrt or sqrtsd whichever is appropriate), except it may have some overhead in order to test for negative input and set errno. sin and cos normally don't correspond to fsin and fcos, along with other trig functions they're normally implemented as some non-trivial function of many instructions. The exact implementation varies based on which math library you use. About the only way to get their timing data is to test them yourself.
A popular approach for fast ray-tracers is to use SIMD processing through ray-packet tracing, though there are other ways to incorporate SIMD. Intel has an introduction to ray-packet tracing, it's showing its age with the 128bit SIMD but it could be extended to 256bit. There is a lot of literature about SIMD in ray tracing, and it's probably useful to read some of it in advance because once you're written a non-SIMD ray tracer it becomes harder to hack SIMD in, than it would have been to base the design around it in the first place.
|
|
|
|
|
thanks for positive feedback, there`s plenty to learn
|
|
|
|
|
VS2008, Windows Embedded Compact 7, mfc dialog application that is 4 years old. It's been working fine. Today I've been making improvements, tracking down some long standing issues. Up until this afternoon, I could pretty much debug where I wanted but then suddenly my UI loop stalled in the OnInitDialog. Stepping through, I came to a couple of LoadBitmap calls that I use to load images from the embedded resources. The first call does not error out, it just never comes back from the LoadBitmap call.
This smells like some weird elephanting debris from VS2008. Some references suggest memory issues on the target device, etc, but it's been rebooted, this is a lightweight app... and doesn't use much memory.
Anyone seen LoadBitmap stall like this?
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
Did you tried to load another bitmap, something small in order to find out where is the problem ?
|
|
|
|
|
I actually did... there are only three bitmaps in the project. I added some code:
CBitmap myBitmap;
myBitmap.LoadBitmap(IDB_STOP);
it never returns from the LoadBitmap... very odd. The bitmaps are not large (200x200?), and the really weird part, again, is that this code has been working for many years.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
I would try a little workaround: use CImage instead of CBitmap ... at least you will know where is the problem, or, you will solve the issue .... However, CBitmap is reliable ... I don't think that code from above is causing the problem ... the cause must be in else where.
|
|
|
|
|
Check you have not got a clash on IDB_STOP and you actually have the resource of that ID in the project. All you need is a corrupt resource file and it can't connect the dots.
In vino veritas
|
|
|
|
|
That's why I told him to try to load another bitmap resource  just for testing ...
|
|
|
|
|
good idea, "use the source luke"... looking now - yeah, this looks good. I would expect the loadbitmap operation would just fail and not hang. The resource is built into the exe. Trying the load image... I love you Microsoft.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
So, tried the CImage -> LoadImage approach - no go. Hangs in that call as well. I'm going to sell my house and buy a sailboat....
elephanting code - okay, I figured out why the call hangs. This application is my application loader code in an embedded target. We have two types of upgrades - one with and one without an operating system. So, one thing I need to do is dismount and remount the device partitions (this is Windows Embedded Compact 7), so I can modify OS files. This typically is not a concern, since by the time we're running, everything is running out of ram. Once complete, the device reboots. No harm.
Now I don't know how LoadBitmap is implemented, but it appears that it reads the resource information from the exe file and not what is in RAM. Egg on my face since I missed the #define change. I still argue that LoadBitmap should return gracefully with an error code. Heh it's mfc, filed under "don't do that."
Appreciate all the suggestions.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
I agree LoadBitmap should not fail like that but it is specified the resources are never lifted into RAM. They are tagged in special sections in the file and are excluded from the normal executable load to avoid using memory when not required. Remember a file may contain many resources for different resolutions, languages and the like and much would then be redundant duplicates hogging memory.
In vino veritas
|
|
|
|
|
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
After I install a CBT global hook, does the application that called SetWindowsHookEx[^] need to stay alive for the hook to work for the lifetime of the user session?
Or may the application that installed the hook exit, thus leaving the hook DLL to operate by itself?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
The remarks section certainly implies quite strongly that the application that calls SetWindowsHookEx must keep running (and pumping messages).
|
|
|
|
|
Thank you Richard. I reread the remarks and I see that you are correct.
Quote: However, because a 32-bit application must run the hook code, the system executes the hook in the hooking app's context; specifically, on the thread that called SetWindowsHookEx. This means that the hooking application must continue to pump messages or it might block the normal functioning of the 64-bit processes.
When I install the hook, I must pass a pointer to the hook callback function. However, that's a pointer that's only valid inside the process space of the application installing the hook. How does it use that pointer to call the correct code inside the address space of the application that is a target of the hook? 
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
Richard Andrew x64 wrote: a pointer to the hook callback function.
The system keeps that address (which is the target of the hook) and uses it to call back into your application at the appropriate time (i.e when the relevant hook event triggers). But if you close your application, its address space is destroyed and the hook is no longer valid, so the system will no longer call it.
|
|
|
|
|
Thank you for your response.
OK, so if the system calls back into my application at the appropriate time, why must the system load the DLL containing the filter function into each process that is hooked?
IOW, if the hook function is run inside the installing application, why must the hook function DLL be injected into every targeted application?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
Sorry I don't know the answer to that one. There is an implication that injecting the hook into another process can be done of the call-back function is in a dll. If that is the case then the dll must be associated withe the address space of that process. I must admit it is a long time since I used this feature so my recollection of it is not 100%.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
