Getting Hive?

[fluffy_little_something]

I only know what I read, I am no mind-reader

[EvilDragon]

The CPU clock as such does not really mean much. Look at the AMD FX processors, some of them do between 4 and 5 GHz, yet they are at the i3 and i5 level in terms of performance. There are so many other factors to consider.

This is because those particular AMD CPUs are old tech and inferior core design compared to Intel's.

[ghettosynth]

The CPU clock as such does not really mean much. Look at the AMD FX processors, some of them do between 4 and 5 GHz, yet they are at the i3 and i5 level in terms of performance. There are so many other factors to consider.

This is because those particular AMD CPUs are old tech and inferior core design compared to Intel's.

Yes but his point holds. CPU performance has been dominated by factors other than raw clock speed for some time now. Not that raw clock speed doesn't matter, but it's just one factor.

[S0lo]

I only know what I read, I am no mind-reader

Doesn't really require mind reading to read this:

Which would performs faster, an i5 At 3.3ghz or the same i5 over clocked at 4ghz ?

If after that statement you still think that I think that clock speed is ALL AND EVERYTHING to CPU speed, then you have mistaken me for some one else, some one who thinks that a V8 will always win a drag race with a V6 no matter what . And if not, then it's just a misunderstanding

[mcnoone]

Just a reminder.
Hive is the greatest synthesizer ever created in the entire universe, and aliens from the planet Sighlent are at this very moment trying to mentally probe its makers to extract the code for its creation.
Thank you for your time.

[EvilDragon]

But I was referring to the general statement a few posts ago: "I really hope that more 4GHz+ CPUs will be here soon".

Yes, by that I meant Intel ones.

[Urs]

Improvements aren't neccesarily made much in clock speed, but in instruction throughput. An instruction that might have taken 8 cycles on a 5 year old CPU might only take 2 cycles on a current one. Funny enough, simple instructions like addition or bit operations nowadays take 0.5 cycles. Don't ask me how they do that

Moore's law (number of transistors per space doubles every 2-someting years) holds...

[Dasheesh]

What happened is that Intel looked at the situation and said to themselves "everything is miniaturizing, the current laptops are running too hot, and everything is going to go to smaller and more personalized products, so we need a processor for that outcome." They put an "all in" bet on making smaller and cooler and less powerful chips. They are quicker, but they are also crunching less numbers in an instance. IBM and ARM went the other direction and thought they could keep the power and still bring the temp down with smaller transistors. The smaller the transistors the less heat they produce. Intel won initially with the advent of mobile processing. The peak of that being the M series, which dynamically throttles the processing speed to the task at hand. Nobody wants those processors though as you can see.

IBM makes the most powerful processors, but they are impractical for laptops and mobile computing. It's a trade off... power, vs quicker and cooler. Just like everything else digital. The smaller the transistor, the cooler it runs and the quicker the process happens, but you get to a point where you are only doing one process at a time and they don't handle floating point processing very well (that's where intel is at). The pendulum is likely to swing the other way briefly as the materials themselves are improving and allowing for more number crunching at cooler temps. That's not what everyone is working on behind the scenes though. every one is after the holy grail. Quantum computing is very close now and it will change everything you know about the world around you and your existence...just like personal computers did.

[pdxindy]

Quantum computing is very close now and it will change everything you know about the world around you and your existence...just like personal computers did.

Actually, computers did not change things all that much in the actual world around us. Food is still grown the same as before computers... Goods are moved around the world by ships just like before computers. Houses are built by carpenters with hammers and saws just like before computers. I live next to train tracks and the trains look and work the same as before the personal computer.

[Dasheesh]

Pdxindy. All those processes are done with computers now, but I'm not going to argue it out in the here and now. Just thought I would throw out what I witnessed with my own eyes.

[ATN69]

Improvements aren't neccesarily made much in clock speed, but in instruction throughput. An instruction that might have taken 8 cycles on a 5 year old CPU might only take 2 cycles on a current one. Funny enough, simple instructions like addition or bit operations nowadays take 0.5 cycles. Don't ask me how they do that

Moore's law (number of transistors per space doubles every 2-someting years) holds...

Hi Urs, if you don't mind me asking, what are the specs for the computer you use?
CPU, memory, etc. I am just curious since you are developer of plugs.

Cheers!

[bmrzycki]

Funny enough, simple instructions like addition or bit operations nowadays take 0.5 cycles. Don't ask me how they do that

One way this is done is a core can have multiple ALUs and it's becoming more common for the ALUs to be designed for vector registers. Some cores can temporarily fuse the instruction after verifying there is no register aliasing. Couple this with the out of order nature almost all CPUs exhibit and you'll see for certain workloads two simple arithmetic instructions will appear to execute at the same time. At least that's one way it's done. Intel has a lot of smart people there, I have no doubt they've created other clever tricks as they inspect the instruction cache.

[Urs]

Hi Urs, if you don't mind me asking, what are the specs for the computer you use?
CPU, memory, etc. I am just curious since you are developer of plugs.

Cheers!

I use a 2015 Retina MacBook Pro with 2,5GHz i7 and 16GB RAM and a 6 core Dustbin Mac Pro which I don't have in front of me. I think it's got 6 x 2.8Ghz. Whenever I have to work on Windows, we have a couple of "floating MBPs" which have a Windows partition and all tools/DAWs imaginable.

Mind you, I never have to have more than two or three instances of a plug-in open at a time. Last time I made music (in the box) I did that demo track for Repro-1 which had 12 or so instances on the Mac Pro.

The most common computer people at u-he use is the quad core i7 Mac Mini which isn't built anymore. I started a campaign last year to persuade co-workers to opt for newer computers, but they are reluctant (can you imagine?). I think the graphics artist will be forced to upgrade this year, since we need the Mac Minis for our armada of build servers and test machines.

[ATN69]

The most common computer people at u-he use is the quad core i7 Mac Mini which isn't built anymore. I started a campaign last year to persuade co-workers to opt for newer computers, but they are reluctant (can you imagine?)

..you know what they say, it was better in the good old Days. Thanks for sharing the info!

[S0lo]

Moore's law (number of transistors per space doubles every 2-someting years) holds...

Holds but seams not for very long. Just quoting wikipedia here:

Near-term limits
Most semiconductor industry forecasters, including Gordon Moore,[93] expect Moore's law will end by around 2025.[94][95][96]

In April 2005, Gordon Moore stated in an interview that the projection cannot be sustained indefinitely: "It can't continue forever. The nature of exponentials is that you push them out and eventually disaster happens". He also noted that transistors eventually would reach the limits of miniaturization at atomic levels:

In terms of size [of transistors] you can see that we're approaching the size of atoms which is a fundamental barrier, but it'll be two or three generations before we get that far—but that's as far out as we've ever been able to see. We have another 10 to 20 years before we reach a fundamental limit. By then they'll be able to make bigger chips and have transistor budgets in the billions.[97]

Though a few observers put the limits of Moore's law centuries (250–600 years) in the future,[98][99] these suggestions are largely theoretical.