Yes, leave my beloved Logic alone!!mjones4th wrote:Oopssorry!
Logic 7 released
-
- KVRian
- 980 posts since 25 Feb, 2003
?????????????
-
- Beware the Quoth
- 35449 posts since 4 Sep, 2001 from R'lyeh Oceanic Amusement Park and Funfair
mjones4th quoth
I may not be the most knowledgible person when it comes to the history of the microprocessor, but I do own a bachelor's degree in this stuff,
So do I. Concurrent with the original RISC processors, so I've read the original papers.
I may not be the most knowledgible person when it comes to the history of the microprocessor, but I do own a bachelor's degree in this stuff,
So do I. Concurrent with the original RISC processors, so I've read the original papers.
Last edited by whyterabbyt on Thu Sep 30, 2004 4:10 pm, edited 1 time in total.
An idiot on Set Theory:
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
-
- KVRAF
- 2172 posts since 14 Feb, 2003
Sure!
(Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck,Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck,) Just kidding.
(Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck,Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck, Macs are better, PCs suck,) Just kidding.
-
- KVRAF
- 2172 posts since 14 Feb, 2003
Cool! so we're peers! (well not exactly, I just graduated in 2001, lol!whyterabbyt wrote:mjones4th quoth
I may not be the most knowledgible person when it comes to the history of the microprocessor, but I do own a bachelor's degree in this stuff,
So do I. Concurrent with the original RISC processors, so I've read the original papers.
I dunno, in what I've seen, the RISc architecture is superior, and maybe that's why Intel and AMD are cozying up.....? So to me it makes more sense that a pure RISC approach is superior?
-
- KVRAF
- 4265 posts since 21 Oct, 2001 from my bolthole in the south pacific
Rubbish - Opteron cleans the G5's clock at a slower CPU speed. The core's speed is not hobbled by being fed strategically ordered micro-ops - quite the reverse.Which reaffirms my stance that RISC is better for CPU applications. Also, the cost (in terms of performance) of breaking down CISC instructions into a RISC-like instruction set is still very real, and still leads to a significant disadvantage for the X86 compatibles.
I find this talk about RISC in this era of Altivec and SSE3 amusing. All these processors are having new SIMD units added to do the real work relevant to us DAW users. These all have a shitload of NEW instructions. There does not seem to be a lot of REDUCING intsruction sets going on of late :^)Not to mention that RISC architectures greatly simplify coding....
Another reason why the AMD Opteron is kicking butt is that it has the memory architecture to keep the core fed with data and instructions. CPU cores spend a large portion of those cycles doing nothing useful because the rest of the system cannot keep them busy. That's why this RISC/CISC debate is so NOT what is being discussed by people pushing the envelope. It is marketing guff from the early 90's.
And I used the plural "people" because you are not unique in this regard and I had no desire to personalise this discussion.
Eg
-
- KVRAF
- 2172 posts since 14 Feb, 2003
Okay, my tail is tucked, geez, LOL! 
I studied all this stuff in the mid-to-late 90s when, as you claim, RISC vs. CISC was marketing fodder. Admittedly, I am not a practicing engineer, and I have only marginally kept abreast of the industry (but all that is soon to change, Johns Hopkins Grad Engineering, here I come! Yeah, suck on that lemon baby
)
So I am arguing based on a couple of assumptions that, according to you, egbert and whyterabbyt, have been invalidated by technological advances. If so, then thanks for enlightening me.
So on that note, thanks for the knowledge, and sorry for my continued misconceptions...
RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, Just kidding!
I studied all this stuff in the mid-to-late 90s when, as you claim, RISC vs. CISC was marketing fodder. Admittedly, I am not a practicing engineer, and I have only marginally kept abreast of the industry (but all that is soon to change, Johns Hopkins Grad Engineering, here I come! Yeah, suck on that lemon baby
)So I am arguing based on a couple of assumptions that, according to you, egbert and whyterabbyt, have been invalidated by technological advances. If so, then thanks for enlightening me.
So on that note, thanks for the knowledge, and sorry for my continued misconceptions...
RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, RISC is still better and CISC still sucks, Just kidding!
-
- KVRer
- 11 posts since 8 Jan, 2003
why DAP is better/has more promise than FXTeleport:
(1) DAP is included for free with the DAW, and you can use as many nodes as you want. FXTeleport costs money and they charge an additional amount for each node. Someone above was saying that SIR was better than Logic's convolution reverb because it was free. I guess that only applies when it's free PC software.
(2) With DAP, there's no need to remotely control the node. People over at the FXTeleport site are always talking about the need to add some kind of remote control over the node. Some people use VNC. This is less elegant than DAP, which doesn't need this.
(3) DAP comes from the maker of the DAW, the OS and the computer. Not the case with FXTeleport. Multiple sources for the pieces of software increases the chances of glitches and adds to support woes (companies point the finger at someone else).
Of course, DAP needs to get AU support to live up to its promise.
Don't get me wrong, FX Teleport is doing a great thing and I hope they come out with a mac version, but a company in Apple's position can leverage an idea in ways that a small, single ISV cannot do.
This is exactly what Apple is going to do - leverage good ideas in ways that nobody in the PC world can do (or do as readily) because of the multiple parties involved in making everything.
Oh, and by the way, the idea that apple is "free from competition" somehow is, well, retarded. Apple has so much competition from wintel they are lucky to be alive. If they had no significant competitive pressure, what we'd see is something like what we with Microsoft (i.e., a company that doesn't need to improve its product because 90 percent of people buy it anyways and they don't feel they can switch). When your customers "can't switch", that's when you stop improving things (e.g., long Longhorn delay, etc.). That's Microsoft's gig. Not apple's
(1) DAP is included for free with the DAW, and you can use as many nodes as you want. FXTeleport costs money and they charge an additional amount for each node. Someone above was saying that SIR was better than Logic's convolution reverb because it was free. I guess that only applies when it's free PC software.
(2) With DAP, there's no need to remotely control the node. People over at the FXTeleport site are always talking about the need to add some kind of remote control over the node. Some people use VNC. This is less elegant than DAP, which doesn't need this.
(3) DAP comes from the maker of the DAW, the OS and the computer. Not the case with FXTeleport. Multiple sources for the pieces of software increases the chances of glitches and adds to support woes (companies point the finger at someone else).
Of course, DAP needs to get AU support to live up to its promise.
Don't get me wrong, FX Teleport is doing a great thing and I hope they come out with a mac version, but a company in Apple's position can leverage an idea in ways that a small, single ISV cannot do.
This is exactly what Apple is going to do - leverage good ideas in ways that nobody in the PC world can do (or do as readily) because of the multiple parties involved in making everything.
Oh, and by the way, the idea that apple is "free from competition" somehow is, well, retarded. Apple has so much competition from wintel they are lucky to be alive. If they had no significant competitive pressure, what we'd see is something like what we with Microsoft (i.e., a company that doesn't need to improve its product because 90 percent of people buy it anyways and they don't feel they can switch). When your customers "can't switch", that's when you stop improving things (e.g., long Longhorn delay, etc.). That's Microsoft's gig. Not apple's
-
- KVRAF
- 4265 posts since 21 Oct, 2001 from my bolthole in the south pacific
The idea that DAP is "free" is so transparently false I can't believe you are putting it forward here. The cost of this feature is factored into the cost of Logic and the hardware.cesjr wrote:why DAP is better/has more promise than FXTeleport:
(1) DAP is included for free with the DAW, and you can use as many nodes as you want. FXTeleport costs money and they charge an additional amount for each node. Someone above was saying that SIR was better than Logic's convolution reverb because it was free. I guess that only applies when it's free PC software.
It is a bit like saying the address book on your mobile phone or the feature where you swap info over bluetooth with another mobile phone is "free" with the phone. The cost is bundled in with the hardware.
If you buy additional Macs to exploit this DAP capability Apple will make their usual margin on them.
In the case of FX Teleport - the price of the product and keeping them in business is out in the open for all to see. It is absurd to compare this scenario with SIR which is plain and simple freeware from a very generous guy. Apple isn't in the freeware business.
Eg
-
- KVRer
- 11 posts since 8 Jan, 2003
Hey, if you don't like the word "free", how about "no extra charge." Bottom line, if you buy Cubase you don't get DAP. You have to pay extra to FXTeleport. Them's the facts.egbert wrote:
The idea that DAP is "free" is so transparently false I can't believe you are putting it forward here. The cost of this feature is factored into the cost of Logic and the hardware.
It is a bit like saying the address book on your mobile phone or the feature where you swap info over bluetooth with another mobile phone is "free" with the phone. The cost is bundled in with the hardware.
If you buy additional Macs to exploit this DAP capability Apple will make their usual margin on them.
In the case of FX Teleport - the price of the product and keeping them in business is out in the open for all to see. It is absurd to compare this scenario with SIR which is plain and simple freeware from a very generous guy. Apple isn't in the freeware business.
Eg
Edit - there is sytem link, but my understanding is that it's not really comparable because it doesn't run over TCP. You need audio interfaces on each PC.
-
- KVRAF
- 4265 posts since 21 Oct, 2001 from my bolthole in the south pacific
Hey Cesjr - I'm with you most of the way - this DAP thing is interesting - that's why I am reading and posting here.cesjr wrote: Bottom line, if you buy Cubase you don't get DAP. You have to pay extra to FXTeleport. Them's the facts.
Edit - there is sytem link, but my understanding is that it's not really comparable because it doesn't run over TCP. You need audio interfaces on each PC.
It has been suggested in this thread that it will support 3rd party AUs in time and that would make it more interesting.
I think parallelism at the machine level is going to be necessary in the near term because machines aren't getting faster at the rate they were a few years back.
Processors will all be Multicore within 2 years but memory and I/O will need to expand massively to deal with the sample libraries and so on that are already here and the monsters yet to come. It is perhaps more practical to have several inexpensive machines with several lots of RAM and disk resources than one monster machine - look at the price of a Quad Opteron server for example. Sun has just released a monster for about 20K US fully spec'ed.
FX's products/System Link/DAP - even Midi over LAN from MusicLab and other comparable products - these are all products trying to assist people to harness the power of several machines in one studio production. I think these systems will continue to be refined and - who knows - maybe FX's ideas will be absorbed/cloned by Steinberg in System Link II or something
Regards,
Eg
-
- KVRer
- 11 posts since 8 Jan, 2003
egbert, I'm definitely with you there.
It's not like DAP is the second coming. At the same time, it's not nothing. It's something pretty cool, and not just a FXteleport clone. (Not suggesting you were arguing otherwise).
Actually I think one of the cool things going on is that Apple may think it can waltz into audio and kick Digidesign/Steinberg/etc.'s butt the way FCP kicked Digidesign/Adobe's butt. I don't think it will be so easy, which should make for a good battle! Good for everyone, all around.
It's not like DAP is the second coming. At the same time, it's not nothing. It's something pretty cool, and not just a FXteleport clone. (Not suggesting you were arguing otherwise).
Actually I think one of the cool things going on is that Apple may think it can waltz into audio and kick Digidesign/Steinberg/etc.'s butt the way FCP kicked Digidesign/Adobe's butt. I don't think it will be so easy, which should make for a good battle! Good for everyone, all around.
-
- u-he
- 30213 posts since 8 Aug, 2002 from Berlin
Havn't followed the whole thread, just bumping in 
This is for instance always not the case where the AU is wrapped around other formats like VST, because the code for drawing to the window port is commingled with the dsp portion.
I personally hope that DAP will give the dev community a kick towards more clean designs. It would help the users (DAP) and it would help the devs (encapsulation).
In the case of high order filters, the SIMD fun starts when you calculate 4, 8 or 16 filters in parallel... but if you also have to take modulation parameters into account, this might again becoime more expensive because you have to store more intermediate data.
I know from many examples where SIMD code was simply slower. But then, there are other examples where code just flies
Cheers,
Urs
Currently this is screwed up by the majority AUs that don't fully follow the specs. The communication between AU and its AUView must be safe to run in seperate memory spaces. This means that any data passing must be done by value, not by reference.egbert wrote:It has been suggested in this thread that it will support 3rd party AUs in time and that would make it more interesting.
This is for instance always not the case where the AU is wrapped around other formats like VST, because the code for drawing to the window port is commingled with the dsp portion.
I personally hope that DAP will give the dev community a kick towards more clean designs. It would help the users (DAP) and it would help the devs (encapsulation).
Yes and no. It really depends on the application. A recursive digital filter for instance with its inherent dependencies from instruction to instruction can oftenly benefit more from optimized scalar code, due to less pipeline stalls.egbert wrote:All these processors are having new SIMD units added to do the real work relevant to us DAW users.
In the case of high order filters, the SIMD fun starts when you calculate 4, 8 or 16 filters in parallel... but if you also have to take modulation parameters into account, this might again becoime more expensive because you have to store more intermediate data.
I know from many examples where SIMD code was simply slower. But then, there are other examples where code just flies
Cheers,
Urs-
- KVRist
- Topic Starter
- 36 posts since 22 Sep, 2001 from Massachusetts, USA
No, I don't think so, I know so, I know that this was done.whyterabbyt wrote:Also, Logic can make accurate CPU demand calculations for all of its plugins since they are built-in and made by Emagic.
You think so? Im surprised, I would have thought that would be hard to predict in real-life utilisation.
I guess you have no real information here. It works great. If you "wouldn't really want to think about" it, though, then uhhh... I guess don't.This can't be done with 3rd party plugins of any format, but it is a nice thing, and makes for the highest possible efficiency with lowest possibility of overloads.
Perhaps, but that sounds like your implying dynamic load balancing, which I wouldnt really want to think about when you're talking about real-time audio streams
Yup, gotcha: you can't read.Logic's implementation is much nicer than this, but then on the downside simply
does not allow for this option.
If restricted=nicer then thats true. Otherwise its not.
Uhhh, can you read?
Uhhhh, yes. You're saying its 'nicer' even although its restricted. Im saying its not 'nicer' at all because it is restricted.
Uhhhh, can you follow that?
Again, if only you would respond to what I actually write (this isn't worth my energy)...But still present something with that (major IMO) limitation as 'leaps and bounds better' than, say FXTeleport, which has very many of the features you're vaunting, and without this restriction.
In VST, the plugin itself is the owner of the GUI and responsible for spawning it and releasing it. In contrast, in AU, they are separate components and the host manages the spawning and release of both.No, it's absolutely not. There are other hacks to make things work, but it is definitely definitely not a separation of DSP and GUI in separate processes, this is absolutely not technically feasible with VST's architecture.
What prevents it?
-
- KVRer
- 10 posts since 3 Oct, 2004
-
- KVRer
- 10 posts since 3 Oct, 2004