A bit disappointed at Flash's rubbish 60fps playback of AVC sources... but well, if it doesn't run so good, the download should sort you out.
Hmmmm... very limited as to what I can say about the game before 5pm tomorrow, but I have not actually unlocked *any* cars yet. I'm not sure prize cars are even in it in the conventional sense. All the cars I've got I've had to buy...
It seemed to run pretty well in the PS3s browser actually ... probably because it doesn't take that much memory and it has a custom written AVC decoder running on SPE. Could download the video from the browser too though so I'll be watching that next. It's actually uglier than I remember from seeing on off-screen videos but that's to be expected. It probably looks nicer on a bit laggier LCD.
Tomorrow's feature is a technical discussion of Uncharted 2 based on the demo code. Tons of discussion points and performance tests that I am sure you will all appreciate
I was going to cover MAG but as I've spent the best part of 18 hours now trying to download the thing and failing miserably, I'm on the point of giving up on that.
I was going to cover MAG but as I've spent the best part of 18 hours now trying to download the thing and failing miserably, I'm on the point of giving up on that.
Back to trial HD, I made the calculation for the Edram occupation, that's what I call tight memory management.
10MB=1024*1024=10485760Bytes
1280x680x4x3=10444800Bytes
Sebbbi I've a question I've the feeling that you faced more CPU than GPU limitations.
I don'tunderstand properly the threading model you used for Trial HD:
sebbbi said:
We have split our workload to all the six of the Xbox 360's hardware threads. We simultaneously process physics in one thread and graphics setup, graphics rendering, game logic, sounds, networking and particles in the others. One of our threads takes care of the timing, scheduling and data synchronisation between the other threads.
One of the three CPU cores is completely reserved for the physics engine. In the most physics-heavy levels the physics engine utilises the majority of its CPU core processing time. In the final content optimisation we were most often bottlenecked by our graphics setup thread (it does viewport culling, shadow map culling and scene setup), but in some really physics-heavy levels our physics thread was exceeding 100 per cent usage as well.
That's really something I would like to have more details about.
Does it mean that one core only run one thread with does physics and graphic set up, or than one is reserved for physic and graphic set up (each one has it own thread)?
How look would look the typical CPU occupation so?
For example
Core 0: Game logic + rendering (static)
Core 1: physics + graphic set up (static)
Core 2: Synchronisation thread, particle, networking, sound (+OS) share the two available threads?
Or is it less static?
In the future do you consider the implementation of a task based (finer grained than thread) system or you think the gain would not be worse it for nowadays (and near future) CPU dealing with between 4 and 8 hardware thread?
Did you have some room left in regard to the GPU performance for Trial HD?
I'll also have another question but I'm not sure it make nor that I can state it properly.
When I read your comment about the bandwidth and memory usage for texture, the overall size of your level, RAM occupation (tiny if we consider the active content you have room for HUGE levels actually) and the amount of computation your engine does
sebbi said:
A quickly made graphics engine checklist:
- Deferred renderer highly optimized for Xbox 360 architecture
- Locked always 60 fps (frames per second) tearless (vsynched) graphics output
- Floating point high dynamic range (HDR) rendering
- Real time per pixel phong lighting from all light sources
- All objects and scenery cast dynamic (ESM) soft shadows
- Sophisticated 12 channel per pixel material system
- Parallax mapping on all surfaces
- Anisotropic filtering on all surfaces
- Tone mapping emulating eye iris adjustment to lighting conditions
- Post process filters: depth of field and overexposure light bloom
- Advanced GPU powered particle engine (up to 10k+ particles per frame)
- Particle per pixel lighting, dynamic shadows and proper depth ordering
- Custom made material and geometry compression algorithms
a question grows in my mind, "in general don't you think that ones are trying to do to few with to much?" It's a feeling I've no clue about the average ratio between memory and computation in other games. I don't expect you to go hard on the other industry members but do you have this feel. Like being on the diet is the best thing you can do for your tech. Could it be that in bigger team and to make content creation more convenient for the artists it's easier and more cost effective to simply stuff your RAM use carelessly textures, etc. In the end relying on too much textures to achieve the look you want? (more convenient on the artistic creation front?).
It might a completely wrong feeling.
On a different topic (completely ), the tech (to me) seems really suitable for a hack&slash/RPG game do Redlynx ever consider this possibilty for their tech?
Sounds like dangerous ground to me as each game will have its own unique feature set, you're not comparing apples with apples. Maybe a like-for-like comparison with the same track and the same car could be interesting, but it's not the sort of thing you can really hang a whole comparison feature around.
Car companies buy up their competitors' cars, rip them to shreds, see how they work and attempt to do better. I'm sure it's the same with Turn Ten and (maybe) Polyphony. Now that's a process that I'd like to write about. I wonder just how much TT "reverse-engineered" GT...
Sounds like dangerous ground to me as each game will have its own unique feature set, you're not comparing apples with apples. Maybe a like-for-like comparison with the same track and the same car could be interesting, but it's not the sort of thing you can really hang a whole comparison feature around.
Car companies buy up their competitors' cars, rip them to shreds, see how they work and attempt to do better. I'm sure it's the same with Turn Ten and (maybe) Polyphony. Now that's a process that I'd like to write about. I wonder just how much TT "reverse-engineered" GT...
I didn't mean an overall comparision of the games per se (ie. not like this vs review). But a technical showdown that discusses different art/technology approaches between the two.
An apples to apples comparision with same cars, same tracks (and compared to real life) sounds great. I'm sure a lot of gamers out there will want to see an objective comparision of the two, especially when it's written by someone who knows their stuff.
I'm sure a GT vs Forza comparision piece would be a front page item on Eurogamer.
You could even compare the handling models & physics of the two, perhaps with the help a guest expert (eg. race driver).
I didn't mean an overall comparision of the games per se (ie. not like this vs review). But a technical showdown that discusses different art/technology approaches between the two.
An apples to apples comparision with same cars, same tracks (and compared to real life) sounds great. I'm sure a lot of gamers out there will want to see an objective comparision of the two, especially when it's written by someone who knows their stuff.
I'm sure a GT vs Forza comparision piece would be a front page item on Eurogamer.
You could even compare the handling models & physics of the two, perhaps with the help a guest expert (eg. race driver).
I think any comparison between the two, unfortunately will devolve down to a versus article in the minds of anyone that is a fan of one franchise or the other. If you say game A does this and game B doesn't. Certain people will immediately think you are putting down game B and by association they'll think you're also putting down consol X that game B runs on.
Personally, I wouldn't mind seeing something like that, but I'm already prepared for the internet shite storm if it gets made.
I think it's a good idea to have an tech interview with Bungie, Naughty Dog, and Guerrilla Games. Ask them what tech they used and how to overcome hardware limitation. It would be nice if we know their opinions about what else we are gonna see in the rest of the life of the current gen consoles.
Grandmaster will you be doing a feature on Operation Flashpoint 2? Codemasters seems to have done very well at repurposing their EGO engine to make an open world miltary sim.
35 km draw distance, with full day/night cycle, weather effects, impressive particle effects (persistent smoke plumes and fires), unscripted AI that decides when, where and how to engage you.
And has over 50 vehicles and 70 weapons (with a realistic ballistics system)
Sounds like dangerous ground to me as each game will have its own unique feature set, you're not comparing apples with apples. Maybe a like-for-like comparison with the same track and the same car could be interesting, but it's not the sort of thing you can really hang a whole comparison feature around.
Car companies buy up their competitors' cars, rip them to shreds, see how they work and attempt to do better. I'm sure it's the same with Turn Ten and (maybe) Polyphony. Now that's a process that I'd like to write about. I wonder just how much TT "reverse-engineered" GT...
Can they think of a non-fps game that uses this kind of large online system? (RTS doesn't count) Bonus points if it doesn't involve any form of weapon at all.
And if they come up with something good, tell them to make that next - too many fps games!
but that's to be expected. It probably looks nicer on a bit laggier LCD. 
