Exactly. And here's the thing: as overclockers, we can ignore the power optimizations and are able to push the performance to the limit of the chip.This.
Think about this as well. From 2003-2008 GPUs went from about 75w to almost 250w TDP just by themselves. That is over a 3x increase in power consumption. CPUs went from ~90w to about 90w and increased from 1C to 4C in that same period.
Intel or AMD could certainly build more CPUs with a 250w TDP, but the focus has been on efficiency...
How 'good' would our top-end GPUs be if the hard TDP limit was ~130w like a CPU?
Your understanding of the semiconductor industry is completely off.
The performance of GPUs increases so much because the applications they are used are are very, very parallel. You can quite literally take a GPU, do a full node shrink, double the units, and receive twice the performance assuming perfect scaling. This involves zero architectural changes. The only design changes are accommodations for new libraries.
So the GPU comparison isn't even remotely relevant.
As far as ARM goes, the tricks that they are using for increased performance gains are tricks that Intel used up decades ago. They'l run into the same wall Intel did.
Why do you make baseless argument after baseless argument?
Do you think i even care about those things? Nope.
Intel is milking its market because it doesnt have much competition. Problem is that people are moving on to new more exciting things in other areas of technology. Thats why their profits sunk 27%
Do you think i even care about those things? Nope.
Intel is milking its market because it doesnt have much competition. Problem is that people are moving on to new more exciting things in other areas of technology. Thats why their profits sunk 27%
If you don't care enough about the subjects you are using to back up your argument... maybe you shouldn't be contributing to the discussion.Do you think i even care about those things? Nope.
Intel is milking its market because it doesnt have much competition. Problem is that people are moving on to new more exciting things in other areas of technology. Thats why their profits sunk 27%
You sure are reading a lot into what amounts to the vagaries of buying/selling activity of people over a span of merely 90 days.
Intel will be glad to know that qualcomm, samsung, and nvidia have disappeared from the planet! *Collective sigh of relief at intel headquarters, monopoly has been established*
their depressed share price is enough to show what millions of investors think about the future of intel.
Their depressed share price shows what a few analysts think about selling headlines...
their depressed share price is enough to show what millions of investors think about the future of intel.
If intel came to market with more compelling reasons to upgrade then they would sell more chips.
No need to explain. But honestly, you made the odd comparison, not me. You implied that IB provided more value than SB which I don't agree on. They're so close that the cheaper one is the better pick in almost all cases.Again...
SB > IB = TICK
SB > Haswell = Tock
Nehalem > SB = Tock
Nehalem > Westmere = Tick
At the moment most of us can only make educated guesses. But it's a fairly safe estimate that Intels focus for Haswell was to improve on IB in low and ultra low TDP areas. Knowing this, I really don't think that the desktop chips will get anywhere close to the area where they'd 'make me cry'.Compare SB to Haswell - then start whining.
(Which may happen - but just be fair in your rants).
It seems quite a few have it in their mindset that Intel would like to stagnate but I don't think that's the story at all. This is Intel's interim step to Skylake which just isn't ready yet. I'm sure they would make the jump if they could but solid engineering takes time and an incredible amount of money. Haswell should be a great overclocker anyway, so it should be well received here. It's certainly more than an standard tock in my mind.
Lots of change (in the next year 1/2) and I look forward to it. Interesting times for the industry
Everything is always 2 generations away.
?IMO, I think the biggest trend here is that Intel, in its desire to go after the mobile space, is basically leaving mainstream x86 in the dust. They simply don't care about whether folks like us are happy about whatever actual performance we get on new chips. They've gone from completely ignoring the power issue to talking about nothing else, and if we get a few percentage points more after a "tock", we should just be grateful and shut up.
I do think it would be different if AMD were in better shape. It's really unfortunate.
Sorry, I misread it as "power user", not "power issue".I meant power as in wattage. We went from the P4 era ("50 pipeline stages! 10 Ghz! Who cares about watts!") to now, where all they seem to care about is performance per watt and battery life.
IMO, I think the biggest trend here is that Intel, in its desire to go after the mobile space, is basically leaving mainstream x86 in the dust. They simply don't care about whether folks like us are happy about whatever actual performance we get on new chips. They've gone from completely ignoring the power issue to talking about nothing else, and if we get a few percentage points more after a "tock", we should just be grateful and shut up.
I do think it would be different if AMD were in better shape. It's really unfortunate.
Well, we'll find out if you're right when they announce their pricing. :hmm::\That should say we should just be grateful and pay up.
That is so true, my sides are hurting from laughing. Be it itanium, x86 mainstream, atom, larrabee (xeon Phi), etc.
Intel's moto really is "our stuff is adequate today, but you should really see the shizzle we got cook'en in the labs for next year and the year after!! "
The only reason they get away with this is because AMD's moto is marginally worse "true enuff, our stuff is not adequate today, but you should really see the shizzle we got cook'en in the labs for next year and the year after!! "
I do think it would be different if AMD were in better shape. It's really unfortunate.
If it was easy to increase performance, AMD would be right up there with Intel.
When Haswell is released, I will likely opt for a performance-optimized model (S), or maybe even a power-optimized model (T), assuming they have the same scheme as IB. I like a quiet, possibly compact computer that doesn't use up too much power - I can't imagine wanting to get a CPU that has a higher TDP than around 65W.
And that's fine. You'll have plenty of options for that.
So will people who want even lower power CPUs.
But what about the high end?
Okay, Intel obviously didn't want TDPs to skyrocket to the moon. But they've made 130W chips before, so why can't they again? What would a Haswell chip with a proper thermal spreader, no integrated GPU and a 130W TDP look like?
My guess is it would look like something everyone on this forum would want to buy unless it was outrageously expensive.
So why isn't Intel doing that? Simple: because they don't care about the desktop market any more, and there's no AMD out there forcing them to care. The issue isn't whether or not AMD could beat Intel. The issue is that now that they are not even trying, Intel isn't either.