i can run ssl successfully on my dual g4 at a buffer of 0 with no problems (build 1113). it's possible what you're experiencing is a g5 bug. developers?


ssl runs like any other mac program. click the green maximize button to put it back in window mode where you'll have access to the dock and menubars and can move the ssl window around. games generally do this as well (take over the whole screen) so in essence they're running in their own "environment" because they don't use standard ui components; however, they're still mac programs and ssl is no exception.

Thanks for the info SpinThis! It seems to be an issue that some Mac's can't run at the smallest buffer size. So SSL should support multithreading?

I opened up Activity Monitor (10.3.5) last night while using SSL and noticed some interesting things:

1-SSL only uses like 88MB of RAM on my laptop (15" 1.25GHz/512MB)

2-Someone had previously commented that SSL used less CPU in the set-up screen vs. the scratching screen. I found this to be untrue: the computer overall uses less but SSL uses more; WindowServer activity drops, but SSL activity jumps

3-As mentioned in another post, there is no 0 USB buffer setting, far left is actually buffer setting of 1 (count the lines). I get no audio with buffer setting 1, but works fine with setting of 2. I know 1113 updated Mac USB drivers, but I never tested this with any other version

My question: Can I configure SSL to use more RAM? I seem to have plenty to spare.

i don't think ram is the issue. ssl relys more on processor than ram. you can download a couple apps (check out process wizard and nicer) that will increase the priority the app gets.

Attention SSL Devoplers/Moderators:

Please include these problems to the known issues. Thanks!

Thanks for including the buffer problem to the known issues Sam! Does anyone know anything about the processor issue?

this is not an issue.. u are comparing apples and oranges

just like saying "my KIA revs at 5000rpm @ 60mph, while my porsche is @ 2000rpm at the same speed, what gives?"

LOL

I got excited that there were some new posts for a second. It's funny how some people enjoys wasting their and others time. Could someone that actually knows about both platforms fill me in? SSL developers? I don't think other cross platform apps use more cpu on the mac than on pc. Thanks!

since am a software engineer by day DJ by nite... i know a bit about poth platforms more about PC than macs...its totally dependent on how the programmers write the app... but u are still comparing apple to oranges...some "cross platform" (crossplatform would mean exactly the same code would work on two+ platforms without change in the code...the only apps that I know off that work on both MAC and PC are "HTA"s) apps may use more cpu in Macs than PC and vice-versa

Thanks for giving solid info tashafa! :)

My guess it that the code was written for PC and ported over to Mac.

Thus explaining it's better performance on PC.

Didn't Steve already inform you that it's co-developed on both here - www.scratchlive.net ?

He actually maintains the mac codebase himself.

I went back and read Steve's posts again and they make more sense to me the second time around. I think some of the more technical things he mentioned that I didn't understand threw me off the first time and I really didn't pick up that he was just saying it was written from the beginning to be cross platform. So thanks for bringing it back up Josh! My bad! So SSL was written to be cross platform but the code happens to perform better on PC as far as CPU usage? This is what I'm really trying to find the answer to. Could you please please explain Josh? Steve? Thanks for you time! Much appreciated!

Hi depakote, let me attempt to answer your burning question once and for all.

Scratch LIVE is a performance app. It tries to give you the best performance your computer can offer. It is not designed to use as little CPU as possible, it is designed to use your CPU to the fullest capacity until everything is creamy. When people compare different graphics cards, they don't try to work out which one is using less power, they see which one spits out the most frames per second. This is also how people like to compare computers sometimes.

The way you should see how your G5 stacks up against a PC is to run them side-by-side both decoding two mp3s simultaneously, and see which GUI has the better frame rate. We don't display frame rate as a number, but you can compare the smoothness of the turning animation and make a subjective judgement.

Scratch LIVE would never sacrifice audio for the on-screen graphics, but if you have CPU left over, we use it and improve the graphical smoothness. But your actual question was
"Why does Scratch LIVE use more CPU on a Mac?"
and the answer is
"I don't think it does."

Thanks for you time AJ! Much appreciated! :) I understand what you are saying (SSL uses any left over cpu for graphics) but it still doesn't explain why people using PC's (ones that are less powerful than my machine) at the same buffer settings have more cpu available. The only explanation would be that their machines weren't using left over cpu for graphics but that isn't possible. So this questions continues to burn...

I think there is way too much worrying about this one, they are entirely different hardware architectures(big endian vs. little endian). If it works without a problem then why the quest for this answer? Do you feel cheated cause it uses more processor on the mac? Did you have plans for using left over processor cycles for something else? :)

Renting left over processor cycles might be a marketing niche ;-)

Besides that, depakote, you are really comparing pears and... apples ;)

SETI are always keen for cycles...

You also can't take on face value the results that people post on these boards. The only way you should come to any conclusion about this is if you can sit two computers side-by-side and compare the results yourself. Also, the amount of CPU being used it not all about scratch live, because the operating systems are so different. People have been arguing for years about how to benchmark different systems against each other. Running scratch live on two machines and looking at CPU usage is pretty close to the least scientific benchmark I have ever heard of.

Did you make sure that no other processes were taking up CPU time? Did you run Process monitor to se what percentage scratch live was using? Did you think about the fact that all graphics on the Mac are double buffered by the operating system, and that graphics are handled by the OS in completely different ways on each platform? Did you take into account that CPU meters are pretty unreliable? There are so many variables involved in your system of comparison that your results are not useful for saying anything about how the code runs on each system.

It's difficult to say for sure what's going on inside a computer given how complicated they are, and three different benchmark tests will give three completely different answers as to which machine is faster and by how much. This is what you have to deal with when you are comparing apples with Brussels sprouts.

We could create a version of the software that measures the framerate of the graphics while making sure that everything is set to a standard set of preferences, but I don't think the people on this board would approve of that use of our development time, and if it works for you, and the graphics are smooth, then why worry what your CPU meter says?

Yes, I would like to see SSL run as well on a Mac as a PC. What Mac user wouldn't? I would like to try and use SSL with Reason and Ableton Live.

all three simultaneously?

If I understand you correctly AJ, you are saying that there is not a way to even remotely compare both systems? I find that very hard to believe! Yes, I understand that both systems are different but if what others say is true then there is a huge gap between the two with cpu usage. An amount that I would think would be an issue and not just the differences of both systems. Why don't you test the two? You are the developer!

Yes, but not using each full blown at the same time of course. I'm getting ahead of myself and need to lay off of these boards and get to work!

AJ didn't say that there is no way, just that there has been alot of serious discussion about it, and no definitive method decided upon. I don't see how you can state that there is a huge gap given this fact, and given my own personal experience with numerous systems, both mac and PC I am inclined to disagree. You say we should test it, but what exactly are we testing? Mac vs PC architecture, OSes AND our own application builds, all at the same time? I don't see how we could do this without inventing a whole new testing regime, which given that we invented it, would thus be completely worthless as far as providing independently verifiable results.

How do you come to that conclusion? Are you comparing the megahertz speed? As said by other before, you cant simply compare those two os'es independently. Try using a same OS, which has been compiled by the same compiler with the same settings etc. Before you compare those you have to have the same starting conditions. How else could you explain that different manufactureres have different results for benchmarks? One manufacturer uses this set of drivers, the other a different set, different compilers etc etc.

Along these lines, I was wondering what language is SL written in? I am assuming C? And what compiler are you guys using? A Different one for each platform or something like Code Warrior?

GCC within XCode on the mac, Visual C/C++ 7 (.net) on Windows.

I think SSL is awesome but think you guys are really copping out on this one! You may not be able to make an exact comparison of the two systems speed wise because they use different processors but you can come close. My Dual 2Ghz Powermac is definitely as fast or faster then the PC's people were speaking of. Yes, they also run on different systems but that's no excuse for SSL not running as well on one as the other. Even if what people said about their PC's is correct I still think your justification on this one is way off. You are basically saying it doesn't run the same because they are on different systems. Don't you want it too? If they don't then maybe you should write separate code for each! Peace.

Depakote, you say you want to run reason, ableton live and SSL at the same time. Do you know of any PC users that are able to do that?

Depakote, Why not just run take the music in reason and mix it down so that its an aiff and then play it in SSL. Keep the studio/live performance seperate from the djing/live performance, unless you got seperate computers. Most electronic musicians I know that play live usually have a mini disc/ DAT/ CD of the music as backup, as some have had sequencer crashes in the middle of performance. I think running all three would really be looking for problems, but if you can do it, then great.

i think depakote is complianing that his Dual 2ghz Powermac cpu usage is more than a PC... due to the fact that he has a misconception that the PC is slower than his MAC

but the bottom line is that u cant compare the two.

anyways on another note... i remember when the new Dual G5s came out and apple where touting it was the "fastest computer"... I remember they had to stop saying that cause they were porven wrong by AMD and Intel.. how did they do this? they must have benched it somehow? off to google...

www.geek.com

But, sir, AMD, Intel, and DEC already beat us to 64 bits, and both AMD and Intel have CPUs that can outperform our G5 under many conditions.

meant to quote that

members.cox.net

Like I said before I don't expect to be able to run all three apps at once, I just hope to be able to run different elements of each together. I never said that I have the fastest computer either. I just know that it's a fast machine that is at least as fast or faster then the PC computers people were mentioning. You may not be able to do what I want on a PC either but definitely have a better chance having more cpu available. Like I said before even if what the PC users said isn't accurate I still think it's a poor excuse to say that SSL functions differently on a Mac because it's a different system. I would think you would want then to run the same on both wether that meant to write each differently or whatever. It's like blaming any problems for the fact it's on a different system.

So which elements of Reason, Live and SSL would you like to run concurrently? Can you explain in more detail what you'd like to do?

And BTW, if we disagree with you (which we do), it's not making excuses, it's called explaining the situation as we see it.

The issue isn't what I'd like to run with SSL, it is that SSL seems to be using more cpu on the Mac than PC. We will just have to agree to disagree on this because I real feel like you guys are beating around the bush. You have done great work and hope to see you continue to try and work out possible issues without losing ground between the two platforms. I have nothing left to say and retire from this topic. Peace.

Thank goodness *locked*