Forget Intel Haswell, Broadwell on the Way

My cousin's desktop (has 3770K I think) had issues with opening multiple tabs on chrome at the same time (either the program crashed or froze) and was probably because of an update. It's funny because my A10 4600M never had an issue where I constantly open multiple recent tabs.

 

What are you talking about Intel SoC being as strong as Xbox One. It is nowhere near. Are you talking about the Jaguar cores, because that's a bad comparison. As far as graphics go Iris Pro is only as fast as Desktop Richland (8670D in A10 6800K is slightly faster when not memory bandwidth limited), while Xbox One is slightly faster than a Radeon HD7770 GHz, making it at least twice as fast as Richland if you count the stream processors alone. Once you count improvements from better architecture and the fact that console games won't be memory limited, we are looking at more than twice the performance of the Iris Pro, and this is not taking into account all the optimizations console games get due to API allowing developers to optimize for the hardware.

 

again i stated in between and i was referencing broadwell version on 14nm...good god read my post. Also did i mention i was particularly referencing the GPU? Also i was making a general point. -_-

GPU wise the haswell version appears to be ~1/3 of the 7770 GHz card so i wouldn't be surprised if the broadwell version was ~ half of the 7770 GHz card, which is a joke considering an iGPU card is even half of that. That is really sad when an integrated GPU is half of the next gen console in graphics performance.

 

Jobine, 5 tabs? lol...

I don't open a web page - I open a folder of shortcuts (IE warning: Warning, Opening 38 tabs could take a long time... and your computer may run slowly... ha ha). Then, I open another folder... and so on. A couple of hundred tabs is about right for a 3 or 4 monitor browsing session...

Having a system that hesitates even a little is like a slap to the face (I've paid enough for the performance, didn't I?) - and having a system crash is the ultimate embarrassment (in front of clients).

When we have a THz cpu with a matching 1024bit O/S, with DDR512 and NO Storage device (RAM is King baby), then I may be happy doing the work I'm doing now.

Hope I'm not still working when we have these (and one day, we will) but I'll be sure to be playing with them even so... (probably it will be the processing power of an entry level phone by then).

But the point is: I've never met a cpu that was 'too powerful'. I can always bring the system down to it's knees (and sometimes, without even doing anything 'work' related.


CPU+RAM=WORK DONE

Give me >64, 30GHz cores with a terabyte of Hex Channel DDR5 RAM today and in less than an hour, I'll need to upgrade to the next best thing.

No, not greedy - just that life is too short to spend waiting for a machine...

 

yes that elitist stuff and more

and no we don't measure cpu performance with that ineffective igpu

 

When an ultra low voltage cpu/gpu combo like an i3 3317u can power three displays in a 4"x4"x2" form factor (See: NUC), 'ineffective' is not what I call it.

Power that capability up with a desktop, i7, quad core/eight thread processor and I don't need to 'measure' what the video does anymore.

Elitist? Nah, one boy's 'elitist' is another girl's 'modest' around these parts.

 

I think by the time broadwell (new rockwell name, right? ) rolls around we'll be seeing huge battery life improvements.
That'll be 2014... and I'll be buying in... 2015-16.
Honestly, battery improvements are going to be pretty crazy.

 

Well I know you didn't mention the GPU but I assumed that because that is the only important point of consoles. Nobody cares what CPU it uses, as long as it isn't as complicated as the one in PS3. And the fact that the Broadwell chip is coming out a year later and costs 50% more than the whole console (- kinnect) already tells how irrelevant it is. Sure it only uses half the power (47W vs 95W) but it also has only half the performance, and considering efficiency only goes down with higher clocks and the fact that Intel's TDP rating is BS and under GPU load it goes way above the TDP even the Xbone looks good.

 

I'm getting close to the point with Haswell on notebooks, at least, that I don't really care that much about battery gains anymore.

 

Big GPU improvements they say

 

Tiller said a few posts up that he can never have enough CPU performance. I can never have enough battery life. I'd love it if I could use my next computer for a solid week before having to plug it in again. Battery tech evolves slower than semiconductor tech, though.

 

djembe,

our wants are not that far apart as they may seem: if we had super powerful processors but, only used them for 'normal' workloads - we could have the battery life and the instantaneous performance we both expect.

The trick is; making the complete platform to use negligible power when idle (and have it stay at or near idle 99.9999% of the time).

 

I suppose that would work, but you'd need even more aggressive power gating on all components (not just CPU or GPU) and some iteration of transflective memory displays.

 

which begs the question of why intel showed at their developers conference their works on wireless and new display tech with .... drum roll... memory!

 

Oh, I know the tech exists. It's just all prototypes now, much like the countless methods that have been announced for improving or replacing traditional lithium-ion batteries with other vastly superior solutions (aluminum-air, carbon nanotubes, graphene, etc). The question is not "can it be done?" but rather "can it be effectively mass-produced?"

 

more like "can it sell and make a profit?"

 

the question with the wireless tech is when, given broadwell soc approach... I'm betting skylake

the question with the monitors is, will it actually leave alpha phase?

the question with batteries is when are they going to make something decent? from the looks we forgot how to make em for several years ago now


but for me the question is are those effective techs? doesn't matter if on paper they can fly unicorns, it matters if the produceable and reproducible product is actually more interesting than what we have available.

and with batteries, its a mass failure all around.

 

precisely. A great many companies and developers have announced and even demonstrated seemingly incredible tech advances that never made it into mass production. And while I fully support research and testing, it doesn't have an impact on end users unless it can be put into production.

 

which means can it be effectively massed produced....it isn't effective if no one wants to buy it. Like green cars....take out the thousands of dollars that the federal and state governments redistribute to get one of them sold and you have no industry. Fact

@Kara...there is only a limited amount of material on this planet and universe....if the material doesn't exist with the proprieties we want...it can't be made. Can't complain about battery tech when it is limited by physics. It is like complaining that cars can't get 200 miles per gallon....only so much energy is in the fuel.

 

not really on both accounts.

for one most of the new battery research has gone away from super duper high tech materials, to more manageable ones.

and no one is actually dreaming and researching on materials that don't exist. nor we are actually thinking of using the rarest materials on earth (unicorn horns for example) on battery, given that almost everything now uses one.

the research speed is abysmal, the proof of function aint something that is really interesting

you should also have paid more attention on your economy classes regarding barriers of entry

 

not sure what your getting at exactly but your complaining on the research speed of battery tech is completely baseless.

There are only a couple battery techs that i can recall that are being researched that are viable in any form in the near future and a couple that are pure dreaming. The only 2 i can recall off my head that would actually allow some game changing are the nano tube/sphere and li-air. I believe only the nano tube/sphere is even close to being commercially available. That was supposed to be out a couple years ago according to some statements made by the developer (Dr Yui IIRC).

The economy has a very large vested interest in getting better battery tech so it would surprise me to see it magically be done any faster. Any company with a new battery technology would have a huge competitive advantage.

 

its simple, you made some points:

1- You point out if no one wants to buy it, it isn't effective
A: I didn't answered it, given that if that were true a lot of things that we have now, wouldn't be available. I did answer the next point that basically lack of knowledge on barriers of entry

2- You raise the point of limited quantity of materials
A: didn't bothered its true

3- You raise the point of physical theory limitations
A: didn't bothered its a chemical theory and application limitation

4- You raise the point of impossible materials
A: unicorn horns

The I go on and make the statement that research speed is lacking, which was what prompted all that.

here is a very basic framework of how research works
1) propose idea
2) experiment with materials
3) build general theory
4) build proof of function, concept
5) research what went wrong
6) correct general theory
7) build a prototype
lets say that the prototype is successful, it usually isn't
8) research on production
9) feasibility tests for production
10) see if anyone wants to actually use it, despite already having invested money on that

speed of research is still lacking by the simple and mere account that nothing significant has been done for quite a long time, we have very minor refinements on a tech that is from the 70s. quite pathetic.

for battery coverage on whats being developed you should read ars technica more, they give with good detail whats going. and djembe listed a few.

 

Any updated eta for mobile Broadwell? And what is the latest information on Broadwells features?

 

Best guess is mid to late 2014. The most important thing about Broadwell is supposed to be the integrated graphics being competitive with Nvidia & AMD midrange. More info here.

 

And better power use while in use (rather than at idle).

 

any see a review of the new independent voltage vs non independent voltage? I am curious i anyone has tried to compare IB/HW in single thread action to see if it has helped drop the TDP draw for single thread. my IB uses 25w for single thread so i am curious if that has dropped due to the IVR.

 

Intel’s 14nm milkshake: It’s better than yours

 

Broadwell dual and quad coming in 2H 2014

 

Ugh. Annoying that another upgrade focused on power improvements and GPU improvements is getting delayed. Has there been any rumors at least of Broadwell potentially being when we see mainstream quad core CPUs in Notebooks, or any improvements to the amount of L4 cache that is present for the GPU?

If not, even if there is another big leap in GPU performance, with Skylake's DDR4 right around the corner, I feel like you would be a fool to upgrade for GPU improvements with Broadwell.

 

I think Broadwell is when we finally see quad cores replace dual cores everywhere except maybe ULT. I think the M-series will be quads only. I remember reading rumors from a long time ago about this, but who knows if it will be true. It should happen since we already have 35W quads.

 

5770K is confirmed to be an i7 quad. Not sure about the i3's.

Also Broadwell is rumored to support DDR4, but only on the server side. Now i'm debating if i should make a Z97 mITX Xeon build for my next desktop instead of waiting for Skylake.

 

We have to wait til June ish?

This
sucks....BAD.

 

June (ish) would be good - I think it's more like October 2014 though...

 

So the Haswell laptop i purchased in June will be considered "recent" for a year and a half?

Damn.

 

'Recent'? Maybe, but 'current'... no.

 

I have had a couple of haswell notebooks and the battery life and speed were not really noticiable different from ivy bridge, so no real need to upgrade, now ivy bridge to broadwell might be worth the upgrade we will have to wait and see.

I am back to AMD at the moment A10-5750 and A6-1450 , i keep switching back and forth between Intel and Amd , for power Intel is the way to go though

I have had at least 50+ notebooks

John.

 

Yeah, but when i upgraded from 2.0ghz Single-Core Conroe Celeron to 3.4 ghz Quad-Core Haswell i7 i felt like:

 

If there's no need to upgrade, would buying a new laptop now be advisable? I'm planning to get the Y510p SLI, but I'd rather it not be outdated so quickly (with Maxwell and Broadwell coming out). I already have a Macbook for school duties (minus some programming things), and want to get into PC games, but yeah. Seems like the main upgrade here will be for iGPU, which I don't even care for very much.

Sent from my SAMSUNG-SGH-I747 using Tapatalk

 

If you current notebook is running fine, and it`s powefull enought for what you need, spend the money on somthing else, wait till broadwell and then think again, it`s a bit silly asking me really as i always buy the new shinny notebooks

John.

 

Same for me, going from 2.26 Ghz Pentium M (Dothan) to 3.4 Ghz i7 (Ivy Bridge)

 

Good. More time to save pennies for my next machine.

 

Dothan was pretty good for overclocking (the magic pin-modding), I had a Dell Latitude D610 with ~1.6GHz CPU modded to 2.4GHz... Anyway waiting for 14nm Brodwell, hopefully by than AMD also done with 20nm Radeon GPUs and I can have a new super efficient laptop .

 

pretty sure the new GPUs come out this summer i thought.

 

Maxwell is coming out in Q1 it seems. As for AMD they are still milking their rX 2XX series

 

pretty sure that hopeless is right

 

Never said he wasn't.

 

neither did i, but someone let the cat out

now go and fetch him

 

Very true. I remember "upgrading" my 1.86 Ghz Pentium M 750 with a modded 1.7 Ghz Pentium M 735. It ran solidly at 2.26 Ghz ever since.

Regarding 20nm graphics, TSMC is claiming up to 30% energy consumption decrease from 28nm to 20nm at the same performance level and then up to 35% energy consumption decrease at the same performance level from 20nm to 16nm. The performance increases for each die shrink are smaller (up to 15% from 28-20nm, up to 20% from 20-16nm).

 

Core i3 Broadwell 14nm in Q4 2014

 

That is do not even yet 20nm, as I heard Maxwell is still the good old 28nm. It would be nice to see a mobile GPU roadmap from AMD and Nvidia too.