Haha! No, we don't have to depate about that anymore! :D
I just had to say something cause I saw the carillon's test setup. :wink:

Err... So testing with Cubase is not good because it's not optimised to G5 and testing with Logic is good because it has an vintage version of it on PC which is not optimised to A64? Not logic in here. ?hih. There's no point in that comparisation because the older software performs naturally wrose than the updated Logic 7 which supports 64-bit system. As I said only way to do an objective conrete testing is by using 3rd party plugins within an updated sequencer (Logic7, SX3).

Whenn I'll get my system up and running I'll do some benchmarks.

I think you didn't quite get my point here. I'm trying to say that in this case Carillon have included a G5 system in the comparison too, because they know that it doesn't run cubase well (because the code is not very efficient) and they want to give customer an illusion that carillon machines are super fast compared to macs. See? Apple sure has done something similar with some other tests which can be found on their own site, but what really matters is how the system really performs with the software you want to use. Isn't that the most important thing here? Like I said, don't buy a mac if you want to use cubase because cubase runs better on pc. There are also many other programs which mac can run alot faster than any pc. It's only up to what program you want to use.

It's quite obvious that Steinberg has been doing much more work with the cubase/nuendo pc version than the mac version because cubase is the leading sequencer for pc platform and they want to keep it that way. However on mac the logic is the leading sequencer now. Apple is behind both logic (application) and mac osx (operating system) so it has a vantage over every other manufacturer because it knows exactly how the OS is coded and how to achieve the most efficient code for an application. Also what happened when apple bought emagic, they put all the emagic's optional plugins (worth over 1000 euro) together with logic platinum and named it "Logic Pro". Only a very big company can do this radical decision and that was amongst the other things a real sign that apple wanted logic to be n.1 sequencer on mac.

As I see this, I think Steinberg supports mac only because there are still some old cubase users who use mac. I wouldn't be surprised at all if they declared that Cubase will be discontinued on mac and continues only on pc (sure not what I want but it's possible). It was a real shock when apple/emagic told the same about logic & pc. I have read from many places that the new SX3 performs even worse than the sx2 on mac so maybe there's something that Steinberg could improve in their code. Compared to logic, the v.7 runs a little faster than v.6. People have reported that their projects run with less cpu than before.

About the processor technology...

If you didn't already know, a 64bit processor doesn't give any performance boost in calculation in real life. It only increases the memory amount you can use. And like at pc platform there's no 64bit windows available yet for music production, neither is there 64bit osx available for mac! So both Windows XP and Mac OSX 10.3.7 still use 32bit code. And if the operating system cannot work @ 64bit, neither does any program which runs in that OS. Moving to 64bit doesn't affect the performance that much, it has been tested and it's about 1% more power or not even that. The only performance boost will come from the extra memory which the processor can use and if the program even needs that much memory. So logic5 isn't that "vintage" at all compared to logic7, because they both use 32bit code. I know many people who still use logic5 and they wouldn't like to hear it's "vintage".

And like I said, I was about to switch from cubase sx2 to logic 5 pc version. I liked the logic that much and it still performs very well with pc too. Usually an older program is also faster to run. Logic5 runs even a little better with a64 than with intel p4 altough it "hasn't been optimized" for a64. Also the logic5 can run more of the same vst-plugins than the new SX2. Surprised?

So my comparison WAS fair because I definitely wanted to use logic and could have continued using the Logic5 and have a performance of 72 platinum verbs which is already a very good result. However I wanted to get a new version of logic and the new plugins and tools to work with so I had to switch from pc to mac. So if my mac now has twice the power compared to my pc, that's not fair or has to be a misunderstanding? I don't get it. Should I use cubase on mac too to get real results? Come on!

I don't want to sound mean, but you should study a little bit more of what you're talking about, altough you're a moderator. :)
And altough it says in my information bar, that I'm a "complete newbie", in real life I'm not. :wink:

Ok, no problem...

Ah. Yes.

Well, of course the new technology allows improving the structure that the clock speed can be increased. But let's not forget that processing double int's, double floats and other 32-bit data structures is faster with 64-bit system due the improved structure (the accumulator on 64-bit processor is bigger)

I thought OSX already had 64-bit system. The 64-bit Windows XP beta has been around for a while. It has some bugs in running some 32-bit programs, but with it SX3 could run in true 64-bit mode. The official release of Windows XP64 would be late winter/spring. I hope they get it done until then.

Hmm... This is not right. The increase is much more. No-one would start developing 64-bit software (not mention OS's) if the increase would be only 1%. The increase has to be more. Why else Apple and Microsoft would spend hundreds of millions in developing software that actually doesn't improve anything. I'll be gettin my computer around January so I'll make a test by installing basic Windows XP and the XP64 beta and run Cubase SL3 on both systems. I ban bet my arm that the same project on XP64 runs better.

*cough* Sampling *cough* *cough* Animation *cough* *cough* Research *cough*. Few years back someone might have wondered for what the hell a computer needed 126 megas of memory. It was insane amout at that time. The need for memory and processing power increases all the time. When technology allows bigger memory capacity and processing power it's taken in use in no time. I bet that after two years average RAM amount is 2GB or more and HC users have the max.

Sick.

G??! Of course if you have an strong urge to use Logic then compare the old PC Logic with the new Mac Logic. And of course the Mac has twice the power because it's dual. Doh. Hoho.

I haven't said anything about testing only with Cubase. I meant that when doing concrete objective comparisations which are not dependent on the sequencer Mac should run Logic 7 and PC SX3 and do the comparisation with several different 3rd party plugins. That way both systems run the software that's up to date.

Probably, but still there are some holes in the knowledge in both sides. Ie. that 1% increase can't be true.

Edit: I'll add that at the moment the differences between 32-bit and 64-bit systems are not very huge, but the situation will change when the 64-bit software are more common. So we are going to see a good performance boost achieved by software on all 64-bit systems in the next year.

Yes, I noticed. That's why I needed to correct a few things...

As you see, it's your personal opinion. Of course we will benefit by 64bit technology, but your 64bit processor will not boost it's performance by 50% or at least 10% when the 64bit OS arrives. Like I said, in real life the most noticable boost will be in the memory amount which the processor can use. There are many benchmarks already available comparing 32bit vs. 64bit. Search with google for example.

I guess Apple will bring the new 64bit OSX 10.4 Tiger out next spring. I'm waiting for it but I don't expect a huge performance boost since I don't need that much memory yet.

I wrote: "Also the logic5 can run more of the same vst-plugins than the new SX2. Surprised?"

That's right. How can it be possible? It just is...


I don't doubt that we would ever need 8GB ram for example. Of course that day will come and sampling needs lots of ram already. You just use as much ram as you need. It depends on how much you use sampling synths and samplers. Analog- and physical modeling synths and effects plugins don't need very much ram. Some delay and reverb plugins may need more but not that much it would be noticable if you have already at least 1gig of ram. Also the new direct disk streaming option in most soft samplers has helped with ram issues. For example I have sampled huge 24bit JP80x0 Strings (1 minute from every key). I couldn't use them without disk streaming on the moment. I have only 1.25GB of ram, but I'm going to expand that to 2GB. It's well enough for my use.

You don't seem to take this very seriously as you laugh after every sentence. Yes, I'm aware that my mac has 2 processors. I just wanted to point out that PowerMac Dual G5 is not just a crap with it's 2 "lousy" processors against single processor pc computers as the carillon test showed. Nothing more. And it doesn't matter if I compare logic7 performance to logic5 or sx2, because the performance ratio is about the same with 3rd party plugins too.

I just want to correct false statements.

cheers.

Yup. I assumed that it was 64-bit because it was released as a fanfare for the new G5 family.

But if the performance boost is totally trivial then why there are so much effort put in 64-bit OS? It doesn't make any sense. The 64-bit systems can run 32-bit OS's with no problems (including the extented memory) so why to create a new OS that actually doesn't change anything at all? That's my point. If the increase is only one or to percents then why all the hassle? Whats the motive?

Yup. I've seen those. The differences aren't big yet due the software. Only about 10% performance difference.

Many reasons. Less CPU hungry mixing and dithering algorithms for example. It's true that the older sequencers are very light compared to the new ones.

Yup. But as you have noticed the need for RAM increases unnoticeably. The samplings get bigger so the DFD eneds more RAM. There's not long before 2GB RAM is too small. Technology and software develops at frightening rate, one sohuld be always prepared for future. Do not get a system that fits your need, get a system that's over your needs.

Nah. It was just silly to say "twice as fast" when talking about a computer that in theory has two computers in same.

Actually the sequencer affects the CPU time taken by the plugin because the sequencer does mixing & dithering. The differences in the algorithms and the engine in general cause differences in the overall performance. Like I said earlier the new sequencers are heavier due improved algorithms. For example it's quite a big difference between Cubase VST32 and SX3.

I think it looks even more silly when carillon shows some benchmarks where G5 Dual 2.5GHz! is at the same line with single processor PCs and the benchmark has more to do with poor optimization than real calculation power. :wink:

Yes, that's true, but like I also said, Apple has managed to improve the logic's code so much that logic7 needs less cpu power to run than logic6 altough there are several improvements in the audio engine too.

And you can always build a dual processor pc too but that's another story. A computer is still a computer, doesn't matter how many processors it has inside. I think there's nothing silly if I say that my mac is twice as fast as my pc because it is. My point was to prove that a dual G5 is definitely not on the same line with single processor pc. If it was, why would I have bought a mac then...

In contrary, if you want to take a look at some Apple's benchmarks...
http://www.apple.com/powermac/performance/

Edit. It seems that the carillon's test was totally gay... :lol:

1. They used SX3 as a sequencer which performs even worse than SX2 on mac.
2. They used M-Audio Audiophile 2496 soundcard which doesn't have very good mac drivers.
3. They got clicks and pops with the G5 propably because they had used the automatic power saving mode and a poor soundcard. You must always use the "highest performance" setting for the G5 if you run any audio application like cubase, logic etc... (for example with RME soundcards and their superior core audio drivers you can use 0,5-1ms latency and it doesn't even make a big hit on the CPUs. I think the PCs would have been in trouble if the tests were run with those latencies.)

So when they tested the pro tools le, the G5 won the tests with only 60% of it's power. Normally you can run the cpu bars up to 100% with no problems at all. ;)

Yup.

It's silly because we can say Athlon 64 = G5 and of course 2 x G5 = 2 x Athlon 64.

No-one actually haven't even said dual G5 = single PC (except that benchmark page).

Yes we can compare a64 almost directly to G5 what comes to power. However I still don't understand how you're thinking. Nobody says that you have 2 computers if you have a dual processor computer. I have never heard of that. If the computer can run certain amount of plugins for example, why should the owner always think how many plugins each processor can run in he's computer? I don't think he's very interested in that at all. All that matters is that how much the computer can do work. And I was talking about computers, not single processors. I still don't see any silly about that. I think it's quite silly how you think this over. :wink:

My last post to this thread.

the cpu load that a plugin uses is variable and fluctuates alot depending on the max polyphony + number of simultaneous voices played and the effects on the instrument (heh, i.e. plugins on top of plugins).

i can have like, 15 vanguards on my 1.6ghz p4 if i set em all to monophonic bass patches with no effects on them.

or i can add 1 vanguard with 32 poly and strike a massive 13th chord with delay, reverb, EQ + compression on it and it'll slay my CPU ;)

The silly thing was that it was trivial and quite intuitive observation. It's like saying "There's twice as much water when I have two glasses instead of one". ?hih. My comment didn't actually had anything to do with the topic.