@VirtualMark Np my friend. You are a lucky individual then and i am happy it runs stable. Not all cpus overclock the same but thank you for your input. Therefore i stand corrected on this one, to my defense too many times auto overclock from various UEFI bios has fkd up trying to run cpus at 5 ghz lol. It happened with Asus Deluxe, Gigabyte UD-3/5 and Asrock Extreme 4 on various chipset itterations. At least i am not alone on this, the web is full of reports about this as well.

@dondada Sure mate, points understood, we agree for the most part anyway.

@martel80 I am sorry but where exactly in your posts you say anything significant about cpu architecture except your first post? If you know what science stands for, you then know the most important branches in any science is research and experimenting. No technology is proven until it's applied. That's why we test the cpus under various scenarios.
But sure you can be whoever you want if you like np by me. As for me, i am a musician, dj, audio/master engineer and producer (yes, in this particular order). If you want to see my studio or listen to what i make for me or others, i 've got many photos and some songs/remixes/album releases' posts in FB, i am not hiding. Just ask. As for the spelling, English is not my native language. Still, i bet i can speak and write it more elegantly than you do and at least write sentences that make sense.
Now i was wondering, can you tell us what kind of ice cream does Jim Keller prefer ?

 

Last edited: Feb 25, 2016

Yeah to be fair I did find that when I increased the speed past a certain point, a lot more voltage was needed to keep it stable. It wasn't worth the tradeoff imo, as it would mean running the fan at a higher speed all of the time, and possibly shortening the life of the CPU.

But Sandy Bridge was quite good for overclocking. Some of the newer chips didn't OC as well for one reason or another.

My chip still runs new software ok, but that's because there hasn't been much improvement since then. 30% isn't a great deal really, it's not like the old days when chips would double in speed every few months. I always want more computing power, I'd love to see some more advanced music programs come out. U-He Diva is just one example of what can be done when you throw more power into a synth, and I'd like to imagine that future software could do even better.

 

Last edited: Feb 26, 2016

I'm sorry that my intervention caused a diversive rumpus.

I don't particularly care about any of it, re Martel80, it's just that sometimes I read or hear something and a 'spidey sense' goes off inside me that something isn't right. I was interested in the wider topic and debate, originally. I didn't mean to upset, Martel80, but it's the internet, and it easily happens, I know. Anyway.......

I don't have any emotional inclination towards CPUs, or their manufacturers. I think anybody denigrating AMD generally is ignoring the critical real world fact (and constraint) of very limited budgets. Most people want bangs for buck (value) I suspect, and there's a wealth of seemingly very tough choices between AMD and Intel - at the various price points. Especially with AMD's (sort of) shared socket thing providing a straightforward upgrade path which I suspect must have really helped repeat and up-sales. It certainly was a factor in keeping me with AMD for so long - not decisive, but a factor.

Though undoubtedly, I'd go Intel if I had the dough. Or rather, a full analogue desk in the south of france in some chateau overlooking the fishing village..........

I really wanted i7 - because I was so enthusiastic and determined about making chewns - but I'm merely a bedroom musician, I do it for fun. So when I priced an i7 system I just couldn't justify it, as compared to FX8350. I could get nice SSD, a few huge SATA drives, a nicer monitor. A *real* Les Paul. Horses for courses, yeah.

I used to build some systems for people and always say: "Set a budget! And tell me what you use it for." PC's are systems.....and a gazillion possible configurations. It's *a* particular system vs another - at a certain price. It also has to be maintained. I recommend just set a price, find a decent solution (without too much fannying around) and go for it. Getting *the* absolute best deal is ....impossible? Why bother wasting (too much) time on it? Spend time....but not too much, innit?

Back in the 80s a crappy fourtrack seemed to cost as much or more than a decent PC today. So even fx8350, 24/96 across 100 channels and 100 mixer inserts is just totally insane to me now. I can't grumble at all. I always think, if I'm running out of CPU then you just have to bounce - though I hate doing it on PC. That's what you always had to do before....... and it's going to happen sooner or later anyway? Poor workflow? Over-ambition? A dog's dinner? My projects are never very grand, mind you. But I have noticed a reluctance on my part to render down to audio/bounce using software - I just want the system to do it all, in real time too!! And at 192!!! Hell, why not?

But it isn't mission critical to me - I don't earn a living at it - and I have plenty of other calls on my dough. Everyone does? And whilst I am interested technically (from software engineering, first computer 35 years ago (oric1 wooo)) the fanboyism is always a bit tedious. Xbox vs PS etc. Yawn. 35 years ago it was Speccy vs C64. Meh - whatever. More than anything it suggests 'prejudice' - a largely inconsequential and trivial one, but still....

For 'enthusiasts' you'd surely have "Go Intel" as the most basic rule? But seemingly everyone is an enthusiast on a tight budget.....so it gets more complicated. The Zen stuff sounds great - but they'd be a poor outfit if it didn't at least sound good in the marketing. Everything always sounds good in marketing, doesn't it? And in computer stuff especially, it seems. Promises, promises.....I know, because I bought Dig Dog for the Oric and I still haven't forgiven them.

Oh, and Dinky Kong. (I should have known better?)

There's an old adage that goes "work expands to fill the time available". Well, people have always been wanting more processor power, no matter how much it increases. You can very easily accomplish now what you could only wish to do a decade ago. But now it isn't enough. So when will it ever be? Just like VSTs, synths, whatever - when will you have "enough" of them?

So, what's the point in always building a home but never living in it? You can *always* do more of whatever, but is it worth the time?

Anyway, re Zen and in envy of i7s, I hope AMD play a blinder. A whole new architecture could really be something. All improvements are welcome, no? On the wider note, Moore's Law is under threat from physical law? Reduce scale much more and quantum tunnelling through the insulator apparently prevents electrical signals reliably being carried in the conductor. So, new architecture might give relatively much better returns in future than ones from mostly scale reductions which we've enjoyed until now. I suspect all the %improvements are coming down towards plateaus.

One thing I'm drawn to with raising cores (Zen) is the problem of parallel computing - yes, it's fantastic at some things, but it's complicated. Computing is ultimately a serial process, doing stuff in "parallel" is just breaking up a serial task into various other serial threads, run in "parallel". It isn't really parallel. What I see is software companies writing making big about their multi-threading, until there's a problem, and then it's "Windows manages the threads and CPU allocation - not down to us." Earlier in the thread there's the discussion/argument about threading (various tracks into mixer) and the logical route of audio being processed through DAW/Windows/CPU. I would suggest anybody who has the least handle on how that really is all working (especially in combo) has much better (and important) things to be doing than telling the internet about it. [I don't, obviously!] Some VST's (Omnisphere?) claim(?) to take a core per instance, whatever. Some VSTs claim to be multi-threaded. How is the DAW running across several cores? How are mixer tracks allocated across cores? The final summing must eventually be done in a single core? Each final bit of the final sum must be allocated a single register in the CPU - the final calculation - before it outputs it?

Personally I don't think anybody knows.

It's hardware, it's software.......and a world of sh1t that most people (musicians especially?) really don't want to (and shouldn't need to) care about. God, it all gets in the way as much as it helps, doesn't it? lol

BTW, no, I haven't had a few cans. Well, alright then........

/ramble tamble

@martel80 - I never suggested you hadn't worked for AMD. What I did was point out a series of inconsistencies in your posts. They remain but it's no big deal. I'm sure you're a nice guy and that you have your talents - I don't mean any personal harm. You seemed to be making "a claim from authority" (as it's called in rhetoric, I think). Any such claims are liable to be tested - I tested them and personally found them wanting. That's all. I could, and perhaps should, have just kept it to myself (it messed the thread up). But, you know, it's the internet......I thought I'd contribute. However, you're a real person - not "the internet". One forgets..........

So, perhaps you can accept my apology for apparently upsetting you somewhat? I didn't intend to upset you, the real person, I was just responding to what I saw.....the internet. There really is no malice intended, though I appreciate it perhaps doesn't seem like that.

 

Oh well, lol

 

Latest info;
AMD Zen and Intel Kaby Lake to go toe-to-toe at end of 2016

by Mark Tyson on 3 March 2016, 12:31
Tags: Intel (NASDAQ:INTC), AMD (NYSE:AMD)
Quick Link: HEXUS.net/qacy4f
Add to My Vault:
A new report out of Taiwan suggests that AMD Zen and Intel Kaby Lake will clash rather directly before the year is out. It seems that Intel's mass production timescale for its Kaby Lake processors has slipped, and now these 14nm Skylake successor chips will only emerge in force at the same time that AMD makes its first 14nm Zen architecture processors available.

​

Intel's processor plan slippages seem to be continuing as insider sources speaking to Taiwan's DigiTimes indicate that the Kaby Lake range of processors will not begin volume production until the end of 2016. Previously it was expected that these processors roll out for launch in Q3 2016. However the initial flow of Kaby Lake processors will only start with a trickle (of U-series chips) from "small volume production in mid-June." Intel Kaby Lake mass production begins in November or December, according to DigiTimes upstream supply chain sources.
Back in January we heard from AMD CEO Lisa Su that the Zen processor architecture would debut in high-end desktops such as gaming PCs before 2016 is out. The latest industry sourced report points to AMD releasing AM4 socket-based Zen architecture eight-core high-end Summit Ridge and Raven Ridge-series processors, manufactured on Samsung Electronics' and Globalfoundries' 14nm processes, in early 2017. Thus the overall picture suggests that PC systems makers will get their hands on the processors to include in pre-built systems before enthusiasts can buy the chips to make their own. A staggered market release like this isn't a new phenomena at all.

​

On another positive note for AMD, the new report from Taiwan indicates that motherboard makers are optimistic about the Zen platform's "improved performance and strong price competitiveness". Such confidence could well result in more platform supporting components, such as motherboards, being built for Zen and AMD processors becoming more popular in PC builds from 2017 and beyond.

 

That means I'll be building my new PC, probably Zen based, at the end of 2017, because it's best to leave the new platform to ripen a little until they iron out most of the BIOS bugs, and improve the new CPU making process/yields, so it will be more overclockable and work even better.

Nevermind. I'm too occupied with the studio upgrades anyway, and I'm still satisfied with my Phenom II PC.

 

AMD Zen 8 Core 95W Summit Ridge CPUs Launching In October – Already Going Through Validation
Read more: http://wccftech.com/amd-zen-cpu-8-core-summit-ridge-launching-october/#ixzz4B7MZ06fO

This is the second major milestone that we’ve heard about concerning Zen’s development The first being the tape out of the Zen core / microarchitecture back in 2015 and the second being the tape out of the eight core Zen based SOC in January. This means that not only has the core design been finalized, but development of the first product chip with Zen cores has been completed. This is the eight core 95W Zen SOC which will be coming to the desktop later in the year. AMD has several other Zen based SOCs – systems on a chip – in development. A 32 Core Zen server CPU, a sixteen core Zen HPC APU and a quadcore Zen consumer APU. All of which should be taped out sometime later this year.


AMD Zen Desktop CPUs Are Shaping Up To Be Everything That Enthusiasts Have Been Asking For.
Zen has been one of the most hotly anticipated AMD products in a decade. It’s the company’s first attempt to compete at the high-end CPU space in five years. In many ways Zen also represents the company’s first truly innovative next generation architecture that’s also on process node parity with Intel since the Athlon days. What has lend the tech community even more optimism is that Zen harkens back to AMD’s glorified Athlon days in another even more crucial aspect. Zen is a brand new clean-slate design that’s been led from the get-go by accomplished CPU architect Jim Keller. The very same person that brought us the original Athlon XP and Athlon64 processors. AMD’s most successful products ever.

 

AMD demos Summit Ridge Zen processor



Image: AnandTech

As previoulsy remoured, AMD today showed off an engineering sample of what will be the firm’s Summit Ridge CPU, an 8 core Zen-based CPU with 16 threads.

At the Computex event in Taiwan, AMD CEO Dr. Lisa Su showed off the new processor that will use AMD’s new AM4 socket, making it a drop-in platform replacement for its Bristol Ridge processors.

“Zen is very, very special project and product for AMD,” said Su. “We are in the early stages of bring-up but the product looks really good.”

“This product is eight cores, 16 threads, it is FinFET technology, and it is integrated as part of our new AM4 desktop platform. Zen is a new high-performance CPU that scales across multiple market segments, and we’re working very hard on our server version of Zen, which is also going very well,” said Su in a speech at Computex.

Su confirmed that the firm is still focused on a 40 percent IPC (Instructions Per Cycle) increase, but there is still work to be done on the CPU. According to AnandTech, AMD is going to start sending engineering samples to top-tier customers in the next few weeks, with wider sampling to the the larger OEM base by Q3 this year.

AMD hasn’t specified a date for when it is available for retail.

Su said that Zen will form the basis of a range of products AMD has up its sleeve. Zen will be the basis of its eighth generation APU as well as appearing in server and embedded products in the not too distant future.

“You will see much more from us in the coming months as we move towards launch,” said Su

 

I thought Jim Keller is working for the electric choo choo car guy at Tesla Motors, no? Very weird...assuming we will not have to only use terminals to the cloud like a computing utility, how does this CPU add to us at the consumer level, say, using PT, or Cubase...or Fallout 4?

Is it just the 40% jump in IPS? What is that in FLOPS? I haven't tracked that kinda stuff in a long time...

When I see stuff like this, with no y axis labeling and non-informational x axis labeling...I go uh oh, time to vet stuff. The implication from the bulldozer core to Zen is 5x jump in performance by the way the graph is presented...when the 40% jump looks pretty extreme against nothing to measure it by...red flag to me...but I'm a statistician that doesn't use stats to lie but to find emergent truth.

Robert Palmer, the ex-CEO of Digital. “Designing microprocessors is like playing Russian roulette. You put a gun to your head, pull the trigger, and find out four years later if you blew your brains out.”

 

Last edited: Jun 16, 2016