Page Location: Home > Dreamcast Lounge > Topic
Previous  1, 2
ultrapro
unknown
Posted:
Tue May 31, 2005 2:39 pm
quote : #16
profile : pm
Posts: 1674
no. im not looking for anything in particular, be it improved graphics or framerate, they look no different to me. the difference between the 2 is not noticable imo.
  _________________
Catch me on XBOX LIVE as Ultrapro
icepick
rank 13
Posted:
Tue May 31, 2005 4:20 pm
quote : #17
profile : pm
Posts: 218
Gareth.de; Isn't the behavior of the game dependent on how the game was developed? I've heard (read) people talking (typing) about "frame priority" and "game time priority" or something like that... (The following verbose explanation for the benefit of others.) For example, scroll shooting games prioritize the frames, so when there's a lot going on and the frame rate drops, each frame is still displayed and as such, the game speed slows down. This is good, because if the movement suddenly got jerky, people would be getting dead pretty fast, and it wouldn't be their fault.

In other games, the game time itself has priority, so if a lot is happening and the frame rate drops, the game speed remains normal which means that frames are dropped and movement appears jerky until the situation has cleared. This would be especially important in online multiplayer games, where synchronized timing is important. Now, in either of these instances, I doubt that time speeds up past the designated speed of the game, no matter what the clock speed is. It'd be too much of a drag to have to fill space where the graphics system and CPU are free, in cases where cinematics/scripted camera movement is involved (such as the opening of the first level of Trizeal). If game time is tied in with frame time, I'm pretty sure that there's at least a frame rate limiter (say, 60 FPS) if not a more complex system in place for allowing frame rates higher than a certain rate with no game speed increase.

Another example: Starting a racing game in time trial, and driving up against a wall so that there's hardly anything onscreen would greatly free up the CPU and graphics system. If frame rate was tied in with game speed and no limiter was in place, the stopwatch display would start to count faster. A situation where this does happen is in Wipeout XL for PC -- I can't even play it on my computer, because the game apparently has no limiter in place and even on a non-overclocked PC, the game is way too fast to play. This shows that it really depends on the way that the game was developed; Regardless of whether it's on console or PC, stock speed or overclocked.

I've played very little of Shenmue, so I don't know of whether or not the people walk more slowly in graphically-intensive areas... but, I doubt that someone of Yu Suzuki's experience and skill level would produce a title using methods such as those that you've mentioned. All modern consoles have a realtime clock; Wouldn't action sequences and speed variables be based on this, as opposed to the speed of the CPU or the frame rate? Granted, I got sort of confused when thinking about this, as I tried to pick these theories apart myself. In any case, I haven't watched the video (I'm not interested in overclocking consoles), so if they overclock the CPU by meddling with the real time clock (which seems even more absurd), then all that you've said could be true, and it would be mostly useless with everything besides custom software. Ergh.
  _________________
•• –•–• •
Dreamcast 2007-02-22: Triggerheart Exelica!
Dreamcast 2007-03-08: Karous!
Gareth.de
unknown
Posted:
Wed Jun 01, 2005 3:04 am
quote : #18
profile : pm
Posts: 45
Hi. You make some interesting points - dropping frames to retain speed is indeed a tactic used by some developers to maintain the speed of the game. Other developers (like Epic) will dynamically reduce special effects in order to maintain framerate. In either case though, be it the framerate or the game's speed that the developer is striving to maintain, the point is that they have to do it because the CPU (yes I know the GPU, memory etc. are relevant too but we must ignore them in this case as it's the CPU that was overclocked in the demo) simply cannot cope with the number of calculations being asked of it - which is where overclocking comes in.

You mention PC games in your post. Like I said, consoles are not like PCs in that the CPU speed *does* affect the way the software is being run, though to be honest on PC it does depend on the game. For example, run Quake 3 on a Pentium 3 500mhz and again on an Athlon64 3700mhz, with no graphics card in place on either machine, and the visuals will look exactly the same and will play at the *same speed*, but at a much higher framerate on the Athlon64. Try an older game on an older OS though - software that was developed when the difference between the highest CPU speed and the lowest was like 50mhz, the difference in game speed would have been negligible enough to ignore so no limiters were used. Today though the difference is much greater and steps must be taken to prevent games running too fast. If you want to test this then please feel free - run a DOS game on a fast PC. I remember playing Hi-Octane on a P233 when I was younger and I loved the game. Once I upgraded to an 850mhz Athlon though, the game was unplayable as it was just too fast. It would be interesting to see how it would run on my 3.4ghz HT P4 today. Wink

I also remember Eidos had to use a framerate limiter on the first Tomb Raider on PC when run on a Voodoo graphics card, but that's going into GPU territory rather than CPU speed.

Going back to CPU speed, if you take a look at Unreal Tournament and use the TimeDemo to obtain a framerate counter and then walk right up to a wall, you will see the framerate go up and over 600fps on a decent machine. My own machine will run it at around 1300fps, and at this speed it does also effect the speed of the game. Epic obviously don't use a limiter in this game. To prevent all of my games doing this, I simply use V-sync on my graphics card which synchronises the game's framerate with the refresh-rate of my monitor - 75mhz. This gives a silky smooth gaming experience at all times, no matter what's happening on-screen.

Anyway, back to the Dreamcast. If you had taken a look at the demo posted here then you would have answered most of your own questions - in that yes, all they did was increase the clock-speed of the Dreamcast's CPU from 200mhz to 240mhz, and the difference is a faster demo. There is no framerate counter visible so we'll have to ignore framerate issues for now, but since the point I was making about consoles was that the game speed itself would increase, this is irrelevant. Just take a look at the demo. If you're not very good with detail then use a stopwatch to time the two samples and I think you'll agree that the 240mhz demo is running quite a bit faster than the 200mhz one time-wise. This is because of a higher framerate of course, but unlike PCs, a console does not retain the original pace of a game when the framerate is increased.

Using framerate limiters is besides the point in this case. They are relevant when the CPU has reletively little to calculate. This post is about what happens to the speed of a game if you overclock the CPU, and if overclocking a CPU would result in less slowdown in games. The simple answer is that no, it would not eliminate slowdown, but the framerate would be higher both during slowdown and at times of little CPU stress. The sad thing is though, so would the speed of the game itself.
 
icepick
rank 13
Posted:
Wed Jun 01, 2005 3:16 pm
quote : #19
profile : pm
Posts: 218
Someone to rival my own verboseness. Cool

Gareth.de wrote:
Like I said, consoles are not like PCs in that the CPU speed *does* affect the way the software is being run, though to be honest on PC it does depend on the game.

Couldn't it depend on the game for consoles as well? It'd be a different story if consoles didn't have any sort of timekeeping ability, but there are likely a great number of potential sources for time-based calculations in any given modern system (I'm not very familiar with console hardware design). A multitude of components are set to a certain speed with the use of quartz crystals, and any of these could be used (or perhaps are even meant to be used) to limit the speed of the game to the developers' desired rate--Not even counting the real time clock. Right?

Plus; If Unreal Tournament was not limited and the game could run as fast as the hardware would let it, wouldn't this cause problems during online versus play, since nobody would be playing the game at the same speed? I know that when running a "timedemo" in a game like Quake (in the sense that a recorded demo of the game is played back at maximum possible speed, to see how capable the hardware is), the game will run by faster than normal -- But, I thought that this is because the demo was recorded at a set speed (say, 60 FPS), and when playing it back, the FPS could go as high (or as low) as possible with the hardware, thus speeding up the demo and providing a useful number (the time taken to complete the demo) to measure the overall performance of the hardware with the game.

As for the DC demo posted here; I could imagine that these people programmed their custom benchmark without any sort of limiter, thus resulting in the demo running by faster than normal (such as with the Quake timedemo function). They might even have done this on purpose, since it would illustrate the speed increase resulting from the overclocked CPU moreso than showing a scene from, say, Shenmue...
  _________________
•• –•–• •
Dreamcast 2007-02-22: Triggerheart Exelica!
Dreamcast 2007-03-08: Karous!
Gareth.de
unknown
Posted:
Wed Jun 01, 2005 5:07 pm
quote : #20
profile : pm
Posts: 45
Quote:

Couldn't it depend on the game for consoles as well?


Possibly... but the whole point of a console is that they are all the same. Once you start getting consoles with different CPU speeds and other hardware differences, game development becomes much more complicated as you have to produce software that will push the fastest machines but that will also run ok on the slower ones. Then you get stuff like different resolutions, texture quality settings, etc. etc. and before you know it, you have a PC!

The beauty of consoles is that the developer knows the game will play exactly the same way on every machine on earth as it does on their development kit at their studio.

For these reasons, it's easier to assume everyone has the same hardware and not bother with limits and timers, 'if' and 'or' statements.

Although I'm once again about to drift off the topic here, the rumours of different Xbox 360 versions is worrying because how can a developer take full advantage of the HDD if there will be some machines that do not have one? Hopefully even the cheapest version will have a decent HDD or the developers will be very miffed indeed.

Quote:

It'd be a different story if consoles didn't have any sort of timekeeping ability, but there are likely a great number of potential sources for time-based calculations in any given modern system (I'm not very familiar with console hardware design). A multitude of components are set to a certain speed with the use of quartz crystals, and any of these could be used (or perhaps are even meant to be used) to limit the speed of the game to the developers' desired rate--Not even counting the real time clock. Right?


Right. But just one question - why? Why go to the trouble of using timers etc. when every console should be the same? I really do think that the ability to change console hardware would be opening Pandora's box and give developers and gamers alike a real headache. We already have to play catch-up with PC processors and graphics cards every few years, plus the new generation of consoles that come out every 4/5 years. Having different versions of each platform running at different speeds would be an unecessary evil.

Quote:

Plus; If Unreal Tournament was not limited and the game could run as fast as the hardware would let it, wouldn't this cause problems during online versus play, since nobody would be playing the game at the same speed?


I never actually tried it while in multiplayer... and I am unlikely to be able to try it again too. Since obtaining CAS 2.0 3200mhz RAM the damn game refuses to run for more than a few minutes before hanging my entire system. I can't really be arsed changing memory timing settings just to play UT as all the newer ones work fine. Whether or not they have a timedemo function I don't know.

Quote:

They might even have done this on purpose, since it would illustrate the speed increase resulting from the overclocked CPU moreso than showing a scene from, say, Shenmue...


Well of course. Wink Would you spend the time making a tech demo to illustrate the difference when overclocking the CPU and then implement a frame-rate limiter? Thought not. Razz It would have been kewl if they included a frame-rate display on the screen though.
 
lordnikon
rank 87
Posted:
Wed Jun 01, 2005 5:19 pm
quote : #21
profile : pm
Posts: 5902
Type: NTSC-U/C
Yes different hardware specs do result in worse network performance.

This is a big reason why PC online gaming over dialup can be a real nightmare, and why Dreamcast Online gaming runs very consistant.
  _________________
I'm already numero uno on Dark Helmet's hit list...
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
Display:   
 
View previous topic - View next topic
Page Location: Home > Dreamcast Lounge > Topic
Previous  1, 2