Matias Goldberg

Indeed, what Tom Sloper said.

 

You may want to study psychology and social methods for studying, for example your test subjects can't receive help from you AT ALL. If they fail to install the game, you can't help them. If they fail to open the game, you can't help them. If they get stuck on a level, don't interrupt them to give tips. Even if they ask you and insist, you have to say "it's up to you" (but record the question, it's very likely they're confused by something that needs to be addressed). Even if they failed to do something as simple as installing the program or get to the first level from the main menu, there's a 95% chance it's your fault and has to be fixed.

 

Additionally, you may want to include statistics gathering in your game. For example FPS games record areas where players get killed more often; you can record average life meter, length to pass each level, areas where player spend most/less time, deaths per match/level/etc. Whatever applies to your game.

For example if they spend too much in an area, it may be because it's poorly balanced (an area having too much to explore), players get stuck on something, it's a hard level, its layout looks like there's a hidden/hard to reach area (even though there's none; aka it looks more interesting than it should be), etc, etc... you get the idea.

Also, if I use an unsigned int to count frames, will it reset to 0 once it overflows? Need to increment a frame counter without using if.

If you have to ask that question, you're definitely not ready to talk about multithreading. Sorry to be this harsh, but multithreading is an extremely advanced topic, even more so if determinism and lockstep is involved; and you are not familiar with the basics of two's complement integer arithmetic. That is, a beginner question.

Start your first game single threaded, and possibly of smaller scope.

Your analysis is correct. However you're exaggerating how bad it is.

game is stepped forward, displayed, stepped forward twice, displayed, etc. --> the display will appear shacking!

If you read the preferred way of updating the simulation and rendering in Fix your timestep, even in single threaded scenarios, it is possible that if the rendering took too long, the physics will start updating more often than rendering.
In other words this is a problem that appears in single threaded programs as well. It's not shacking, it's frame skipping.
 
However, I agree that without proper care, the update order can be pretty chaotic, and indeed it will look like it's shacking; which is exclusively a multithreading problem. However, let's see it in more detail:
 
 
First, Rendering only needs 4 elements from Logic, if you need more, you should rethink the design:

• Transformation state of every object: Position, Quaternion and scale. That's 40 bytes (64 bytes if you choose a matrix 4x4 representation)
• The playback state of the animation (if animation needs to be sync'ed from Logic). That's anywhere from 0 to 32 bytes
• A list of Entities created in a frame
• A list of Entities destroyed in a frame

Second, forget the idea that you need to render exactly what the simulation has. If your game can avoid that restriction (99% chance it can), you can relax the synchronization.
 
Third, locks aren't expensive, lock contention is.
 
Now, creation can be handled without invasive locks: Logic creates a list of entities and at the end of the frame, it locks a lightweight mutex, updates Graphic's list and releases the lock. Chances are, Graphic thread wasn't accessing that list because it has a lot to do. At the end of Graphic's update... it locks, clones the list, and releases the lock.
In both cases, it takes almost no time to work inside the locked resource and it consists of a little fraction of all the work they both have to do, so lock contention is extremely low. (Furthermore you can avoid mutexes entirely using a preallocated space and interlocked instructions, and only lock if the preallocated space got full, but I won't go there)
 
There's a catch here, and remember my second advise. You don't care that you're rendering exactly what is in the simulation. Suppose Frame A is simulated, and created 3 objects, but Graphics was too fast and looked into the list. Then loops again, uses renders frame A but without those 3 new objects. Do you really care? those 3 will get added in the next frame. It's a 16ms difference. And not a big difference because the user doesn't even know those 3 objects should've been there.
 
Same happens when destroying objects. Note that a pointer shouldn't be deleted until Graphics has marked that object as "I know you killed the foe, I'm done rendering it"; so that you're sure both threads aren't using the pointer. Only then you can delete the pointer. In other words, you've retired the object from scene, but delayed deleting the ptr.
Otherwise, as you say, a crash will happen.
So in this case, an object may be rendered one more frame that it should. Big deal (sarcasm).
 
Now we're left into updating position & animation data. You have two choices:

• You really don't care about consistency. Read transformations without any locking at all. Don't care about race conditions. The chance that logic is updating the transform at the same time graphics is reading it is minimal (you should be copying the position from your physics engine to a copy, all inside Logic thread; then reading that copy from graphics thread). If memory is aligned, you won't get NaNs or awkward stuff. But you may get very rare states (it's race conditions after all) for example a position in a very far place from where it actually should be..... but it only lasts for a frame! Chances of this happening often is extremely rare because cloning that transform is very fast even for thousands of objects. So just a flickered frame. Mass Effect 3 is a very bad example of this flickering getting really noticed. They must be updating position from the physics engine data directly, instead of cloning it into a list or they use a memory representation other than a std::vector or plain old array (thus increasing cache misses and time spent iterating), which increase the chances of reading data in an invalid state (I'm telling you an example of an acclaimed AAA game doing this and royally screwing it up).
• You do care about consistency. Use a lightweight mutex when copying the physics transform to another place inside Logic thread, and from Graphics Thread do the same. In other words, is the same as above but with locks. Lock contention is again very low.

I've tried both and #1 really works ok if done properly (don't take my word, try for yourself! it's easy to switch between both, just disable/reenabled the mutexes!).
Note that #1 isn't a holy grail of scalability, because it can still slowdown your loop a lot due to cache sharing and forcing them to flush too often (which only happens if both threads are accessing the same data at the same time with and one of them writes to it).
 
Same happens with animation, but it's a bit more complex because you really don't want time going backwards in some cases (i.e. when spawning a particle effect at a given keyframe it could spawn twice), I won't go into detail. Getting that one right and scalable is actually hard (but again, solutions rely on the assumption that lock contention will be minimum).

Remember, you don't care that you're rendering the exact thing, but 99% of the time you will, and when it screws up it often gets unnoticed and fixes itself in the next frame.
 
And remember, synchronizing points 1 to 4 should only be a tiny fraction of what your threads do. Logic thread spends most of it's time integrating physics, a lesser part updating the logic side, and only then syncing.
Graphics spends most of it's time issuing culling, updating derived transform of complex node setups, sorting the render queue, and sending commands to the GPU; and only then syncing.
 
Note if you read transform state directly from the the Physics engine data you'll have terrible cache miss rates or will have to put a mutex to protect the physics integration step, and that does have a lot of lock contention.
 
All of this works if there are at least two cores. If the threads struggle for CPU time, then the "quirckiness" when rendering becomes awfully noticeable. Personally, I just switch to a single threaded loop when I detect single core machines. If you've designed your system well, creating both loops shouldn't give you any effort, at all. Just a couple of lines.
 
And last but not least there's the case where you really care about consistency, and you absolutely care that you're rendering exactly what is in the simulation.
In that case you can only resort to using a barrier at the end of the frame, clone the state from logic to graphics, and continue. If both threads take a similar amount of time to finish the frame, multi-threading will really improve your game's performance, otherwise one thread will stall and must wait for the other one to reach the barrier, and hence the difference between single-threaded version of your game and a multithreaded one will be minimal.
 
 
You asked how this is dealt with, and there is no simple answer, because multithreading can be quite complex and there's no simple answer. There are numerous way to deal with it.
A game may put locks for adding & removing objects & updating the transform, another game may not use locks for transforms. Another engine may use interlocked functions to add & remove objects without mutexes.
Another game may just use barriers. Another game may not use this render-split model at all and rely on tasks instead*. There's a lot of ways to address the problem; and it boils down to trading "visual correctness" vs scalability.

 

*Edit: And it can be like Hodgman described (all cores contribute to task #1, then to task #2, etc) or they may issue commands using older data and process independently (i.e. process physics in one task, process AI independently in another task using results from a previous frame, etc)

Win32 threads: Sleep(0)

No God, NO!

Use SwitchToThread to yield (supported since Win XP)

Sleep( 0 ) is a terrible way of yielding. If you're looking to avoid consuming CPU cycles (i.e. lower power usage) prefer Sleep( 1 ) over Sleep( 0 )

Hi everybody.
 
I'm writting a simple reliability layer on top of UDP. Not that I want to reinvent the wheel, just most of the data I'll be sending can be sent unreliable but there are bits of data every now and then that need to be sent reliably; and mixing tcp & udp gets even more troublesome.
 
For establishing a connection, I'm well aware of the 3-way handshake that TCP uses "SYN; SYN-ACK; ACK"
However, what I don't get, is why the third ACK is needed. Yes, I know it's the ultimate proof that the connection has been acknoledged by both parties; but I don't believe it's absolutely necessary.
 
This question has already been raised in StackOverflow, however the answers there don't satisfy me:
 
Let's see a communication without the 3rd ack:

Case: Client didn't get SYN-ACK
In this case, Server thinks the connection is established, Client think it's not. Just send another SYN from client until the SYN-ACK is received
 
Server may start sending data to client because it doesn't know the SYN-ACK got through. Client won't ACK that data because his connection isn't yet established, so server will continue to keep sending data over & over again until timeout. When the Client successfully gets the syn-ack; it will ack that data, hopefully before the timeout. If timed out, server-side it will just look like one connection timed out, and another came in. It's important that server doesn't use the SYNs from client as proof of heartbeat.
 
Typical view on this: The server needs to allocate resources (for tcp). More SYNs received -> more resources. However I send a random ID generated client-side with the SYN. That way the server identifies the SYN & IP with associated resources using that ID (and deals with client reconnecting and starting a new session as they'll change their ID; or with another client getting the same IP)
 
ACKs from normal messages always send the ID along the sequence number. So that if client reconnects (or new client got the same IP & port old client had) the server won't think an ack received from an old session is acknowledging packets sent from the current session.

 

A connection could be hijacked only if a machine gets the same IP address (and port), happens to use the same ID the other client has been using, and all happens before timeout (or there is a man in the middle that can see all data, spoof the IP, and still read the answers from server because he can see all data going to that spoofed IP).

But it's not like TCP is foolproof to hijacking either. Granted, this method is a lot easier to hijack because the ID, IP & port is repeated in every ack. Furthermore I'm interested in preventing "accidental" hijacking, not directed attacks.
 
The only disadvantage I can see: Potentially much higher bandwidth consumption (because the server may start sending data while client won't acknoledge it), while TCP needs to account for congestion control (which I don't care).
In TCP, everything is silent until SYN-SYN-ACK-ACK has been performed.

 

Bigger ACKs as a disadvantage could also be mentioned, but TCP overhead is already much bigger than UDP, and again, most of the data I'll be sending is unreliable (no need for ack), while every now and then I send some reliable data (needs guaranteed delivery, guaranteed to arrive in order)

 

 

Am I missing something? Why is the third ACK needed?

Thanks!

60fps in milliseconds stored as 'long' truncates to 16 from 16.666 which is highly inaccurate

Also, long(16/16.666) = 0

Anything above 60fps wouldn't be able to move (bug), anything between 58.82fps and 29.41fps would be multiplied against '1', which means the fast machine running at 58.82fps would move almost twice as fast as the machine running at 29.41fps (bug)

 

That time scaling code is completely broken.

Also, fix your timestep, and represent your time delta as a 32-bit long in microseconds, not milliseconds, and once you're multiplying/dividing, keep it as a float, not as integer.

I'd recommend using floor() rather than casting to int back and forth. The casting can be several times slower (if performance is an issue)

 

Note that floor( -3.3 ) = -4; while (int)-3.3 = -3; but floor's behavior is usually how you want snapping to work if negative numbers are possible.

Yeah, no way PS 2.0 can run 15 light equations. There's not enough instruction slots.
 
I'm not sure if it's available in XNA, but try [unroll] and [branch] in your loops and see which one is faster. What GPU do you have?

We both had assumed that F₀ was the Fresnel term and F_spec was the specular color.

Somehow I'm not surprised. I'm sick of going through GDC/SIGGRAPH slides and even independent academic papers where they present a formula/equation without (or with incomplete) footnotes that clarifies the meaning of each coefficient and variable. Half of them you have to guess, or dig through all the referenced bibliography to find out.

The only term however I heard about it being used practically was for HDR and tonemapping, where you would calculate the average logarythmic luminance of your entire scene from the cameras viewport, for the HDR equation.

Don't forget about YUV format. Y is for luminance, and contains the luminance at the given pixel. YUV (and it's variants YUV422, YV42, etc) is the preferred choice when dealing with JPEG, MPEG-2, MPEG-4 & H264 compression, also when broadcasting TV.

Also lately the YUV422 format is comming up in games because it halves the bandwidth and memory requirements for slight losses in quality.

The 2nd option already knows the input is a parameter and there's no doubt it can't be part of the command in the statement, so SQL injection is not possible.

 

In the 1st option, you would have to properly escape the string, and there's the risk that you're not escaping it properly.

To check if your formula is energy conserving, you have to do the integral of your formula just like Fabian Giesen did in this blog (check comments, quick link: http://www.farbrausch.de/~fg/stuff/phong.pdf).

 

That's a looooot of math and has the risk of making your head explode, but it will make you be 100% sure.

d3ddev is a null pointer

Well, it takes time to analyze resources unless you stored them in plain readable formats (i.e. mesh formas that there are already known viewers, png files, you get the idea).

So if someone finds your easter egg through hacking, it would've took him some time by then. If he's the first one, everyone's going to be "holy cow, I did not notice it!?!?" reinstall the game, and go see to confirm it. The easter egg still did it's job.

No password hashing? No SQL sanitizing? No sql prepared statements? Regardless of complexity, safely storing a password is a serious issue, and I strongly encourage that this should be taught from start.

It's not funny when a newcommer follows a tutorial, happens to have moderate success with his first attempts; and then all his user passwords are stolen and all the sql database was destroyed.

It isn't that hard either, specially considering nowadays there is a plug 'n play solution in phppass.
Prepared statements are as easy as normal queries, and they should be preferred when teaching.