With the current rate of Intel CPU performance increases, could AMD be catching up?

[Homeles]

sorry, it should have been huge, not pure. just edited.

In that, case, you probably won't ever see it again. ILP's tapered off.

Here's your proof for claiming that AMDs power gating is a lame hack.

Good grief man. Idle values, in particular, are highly reliant on the motherboard in use.

Whoops, Intel's in the lead here. And guess what? If you take a look, you'll find that idle numbers will go back and forth, depending on the setup.

In mobile, though, things aren't nearly as pretty:

And in case you aren't aware, power gating isn't only for idle, buddy.

 

[grimpr]

Remember chip design cost goes up rapidly. So while a company might have been able to compete x years ago doesnt mean it can even afford it today.

Its expected that 14nm kills of half the current semiconductor companies. Simply due to design cost.

I dont think so, cpu design goes automated tools. fab tech,tools and research goes up.

 

[OBLAMA2009]

pc's are actually getting slower. most new ultrabook type machines have chips in the 1.7 ghz range. so my guess is, 5 years from now, cpus will run at 500 mhz and get a month battery life

 

[Torn Mind]

pc's are actually getting slower. most new ultrabook type machines have chips in the 1.7 ghz range. so my guess is, 5 years from now, cpus will run at 500 mhz and get a month battery life

Lowering of clock speeds is solely to reduce power consumption. But with node shrinks, power consumption can decrease without necessarily affecting clock speeds.

 

[Sable]

and get a month battery life

If only.

 

[ElFenix]

When did we have huge IPC increases from uarch changes? The change from netburst to Core isn't a good comparison, as Netburst was really inefficient.

Yet Conroe devastated AMD as well.

netburst to core was conroe (for all intensive porpoises, no one cares about core1solo)

 

[MisterMac]

netburst to core was conroe.

...well duh.

Might wanna re-read my context

 

[colonelciller]

Let's play hypothetical and say that AMD could/does catch up AND surpass Intel's high-end-consumer CPU line in 3 years time.

My bet is that Intel would go into war-mode with their multiple billions in potential R&D spend vs AMD and take back the performance lead in a shockingly short period of time. It would be great to be a consumer in such a climate.

...here's where i get shamelessly creative: AMD could then go out in a blaze of glory selling the CPUs at just a sliver above COGs (cost of goods sold)... which would force an unheard of frame-shift in the price consumers are willing to pay for CPUs as lightening fast performance could be had for insanely low prices.

The money fountain (billions in profits) for the CPU makers would be cut off and AMD might then be able to compete since nobody would be raking in wads of cash hand-over-fist.

:hmm: I like this alternate reality I've painted for myself.
Please, no one get too serious debunking this

Dear AMD... engage!

 

[MisterMac]

Let's play hypothetical and say that AMD could/does catch up AND surpass Intel's high-end-consumer CPU line in 3 years time.

My bet is that Intel would go into war-mode with their multiple billions in potential R&D spend vs AMD and take back the performance lead in a shockingly short period of time. It would be great to be a consumer in such a climate.

...here's where i get shamelessly creative: AMD could then go out in a blaze of glory selling the CPUs at just a sliver above COGs (cost of goods sold)... which would force an unheard of frame-shift in the price consumers are willing to pay for CPUs as lightening fast performance could be had for insanely low prices.

The money fountain (billions in profits) for the CPU makers would be cut off and AMD might then be able to compete since nobody would be raking in wads of cash hand-over-fist.

:hmm: I like this alternate reality I've painted for myself.
Please, no one get too serious debunking this

Dear AMD... engage!

That scenario would indicate people want lightning fast speeds @ 80-90W TDP in 3 years time for net cost.

We want it - but i think we're shrinking in size of market tbh.
We're not exactly volume people

 

[grimpr]

In that, case, you probably won't ever see it again. ILP's tapered off.

Good grief man. Idle values, in particular, are highly reliant on the motherboard in use.

Whoops, Intel's in the lead here. And guess what? If you take a look, you'll find that idle numbers will go back and forth, depending on the setup.

In mobile, though, things aren't nearly as pretty:

And in case you aren't aware, power gating isn't only for idle, buddy.

Thanks for proving further that AMDs power gating isnt a lame hack. :thumbsup:

 

[mrmt]

...here's where i get shamelessly creative: AMD could then go out in a blaze of glory selling the CPUs at just a sliver above COGs (cost of goods sold)... which would force an unheard of frame-shift in the price consumers are willing to pay for CPUs as lightening fast performance could be had for insanely low prices.

COGS is just what you pay to bring and maintain the chip in your inventory, nothing more. COGS does not include R&D, so if you are selling just above COGS you cannot even pay R&D, or administrative/sales expenses, etc etc etc.

And what you are suggesting has been tried before, it is essentially that AMD undercut Intel on prices, and that worked wonders for them in K7/K8 times, then Intel learned, and started selling cheaper chips that were also smaller and faster than AMD's offerings. The rest is on AMD's balance sheet today.

AMD's problem aren't in the node only, but also in engineering. AMD's big core chips are bigger even when compared to Intel's 32nm offers, and they are more inefficient too.

 

[Homeles]

Thanks for proving further that AMDs power gating isnt a lame hack. :thumbsup:

Might want to start looking for optometrists in your area.

 

[Puppies04]

Let's play hypothetical and say that AMD could/does catch up AND surpass Intel's high-end-consumer CPU line in 3 years time. My bet is that Intel would go into war-mode with their multiple billions in potential R&D spend vs AMD and take back the performance lead in a shockingly short period of time. It would be great to be a consumer in such a climate.

I really don't think intel have anything to worry about. They could easily release their chips at much higher stock clocks than they are now if AMD get too close to them. In fact I was kinda hoping the 8350 would give intel a kick up the arse and we would see some faster haswell CPU's. Of course This wouldn't help "enthusiasts" who would still be limited to exactly the same max OC's.

 

[Ajay]

Doesn't AMD have access to all of Intel's patented innovations?

Sure they can't introduce them in real-time as Intel files the patents but they can incorporate them with about a 3yr lag time.

So at worst AMD should never be more than about 3 years behind Intel's microarchitecture IPC, unless they choose to not implement Intel IP into their CPU designs. And process node is another matter too.

Intel is going to be on 14nm while AMD will still be trying to get to 28nm. That kind of delta can't be easily overcome with CPU design alone.

In some technical respects, AMD has done remarkably well with its 32nm chips vs Intel's 22nm. But, as you point out, Intel is about to go from having the advantage of an ~31% smaller node to a 50% smaller node in 1 year's time. GF's lack of progress on its 20nm node could be the final nail in AMD's grave. The only thing keeping AMD alive is Intel's requirement for >60% margins, IMHO.

 

[Ventanni]

Might want to start looking for optometrists in your area.

I think we need more data to make a fair comparison. The list is very Intel heavy.

This is the list re-done in, well, something a little more English for me. Intel's 17w chips rule the roost, but those are also ultrabooks, which lack things like mechanical HD's and disc drives.

17w - Intel - SB - 11.2
17w - Intel - SB - 11.14
17w - Intel - SB - 10.93
35w - AMD - Trinity - 9.27
35w - Intel - SB - 8.06
35w - AMD - Llano - 7.91
35w - Intel - SB - 7.59
45w - Intel - SB - 6.63
45w - Intel - IB - 5.94

Based on the limited information, I'd say AMD is on relative footing with Intel, but without significantly more data to look at, you really can't draw any conclusions other than the laptop platform/design has pretty significant impact on battery life.

Note, there's only one Ivy Bridge chip listed there.

 

[Homeles]

Based on the limited information, I'd say AMD is on relative footing with Intel, but without significantly more data to look at, you really can't draw any conclusions other than the laptop platform/design has pretty significant impact on battery life.

It's a similar point to what I said about desktops. You can't really draw conclusions about the idle power of a chip by looking at platform idle power. My point was that if he's deluded enough to believe the absolutely asinine assertion that Trinity has better idle power than Ivy Bridge on the desktop (I mean come on. Really? Are you really that ignorant?), then by his ridiculous logic, AMD loses out just as badly in the laptop sector.

You and I both agree that this isn't the case. There are far too many variables to draw a foregone conclusion about idle power draw of a CPU when the platforms are completely different.

And like I said, power gating isn't all about idle power. Workloads are dynamic. Individual parts of a core are constantly being powered on and off. Intel has a performance per watt lead for good reason, and it's not simply because of fab tech.

And he can live in his own little world all he likes about AMD having a lead, but apparently he's not heard of Haswell and the changes that are coming with it that pertain specifically to power gating. What was the number that Intel's throwing around? 10x better idle power? Yeah. Way to miss that one, grimpr.

 

[Fjodor2001]

What was the number that Intel's throwing around? 10x better idle power? Yeah. Way to miss that one, grimpr.

Not really relevant since its 10x in "connected standby", which is completely different than idle.

 

[Homeles]

Not really relevant since its 10x in "connected standby", which is completely different that idle.

It's still relevant, since it's power gating that enables this. I just misunderstood the claim.

 

[Fjodor2001]

It's still relevant, since it's power gating that enables this. I just misunderstood the claim.

Perhaps, as long as you don't compare it to the idle power consumption of AMD CPUs.

Also, is there anything that says that AMD cannot achieve the same 10x connected standby if they want to?

 

[Homeles]

Perhaps, as long as you don't compare it to the idle power consumption of AMD CPUs.

Also, is there anything that says that AMD cannot achieve the same 10x connected standby if they want to?

It doesn't matter if they're capable of achieving it. What matters in the context of this discussion is that they're not.

 

[Fjodor2001]

It doesn't matter if they're capable of achieving it. What matters in the context of this discussion is that they're not.

It does matter, since if they can achieve it then they can also introduce it in the next CPU generations (if they want to that is).

 

[Fjodor2001]

Some thoughts on this:

1. It's pretty safe to assume that Intel and AMD learns from each other's designs. They cannot steal patents straight off, but can get "heavily inspired". That cannot be done without some lag however. So let's assume AMD is 2-3 years behind Intel in the CPU design area. Based on this:

a) If Intel increases CPU performance by 25% per year, that means Intel has:
1.25^2 to 1.25^3 => 156% to 195% the performance of AMD (being at 100%).

b) If Intel increases CPU performance by 8% per year, that means Intel has:
1.08^2 to 1.08^3 => 116% to 126% the performance of AMD (being at 100%).

So you see with this assumption AMD will catch up if Intel does not continue a high rate of performance increase per year. a) has been the case during 2011 and a couple of years before that, but the question is whether b) can become a likely scenario for 2012/2013 and the years to come?

2. Regarding the process technology advantage that Intel has, it affects several areas. For example:

a) TDP: But with 22/28 nm and below the TDP is getting very reasonable for most people on desktop CPUs at around 75-95 W, and that's even including an iGPU. With Broadwell Intel may be able to cut that another 15-20 W @ 14 nm. But we're reaching a state where were seeing diminishing returns. If Intel would be moving into 8 core territory that would be another story, since then the TDP on AMD CPUs (with iGPU) would likely become too high. But since Intel is staying on 4 cores, the TDP advantage it has is not that important on desktop CPUs.

b) Idle power consumption: Actually the process technology advantage does not have much impact on this. The reason is that at idle basically everything is power gated anyway.

c) Economy: No arguing that if Intel can fit twice the amount of CPUs on the same die area that AMD can (due to a 1 node shrink lead), it has a definitive advantage. But note that AMD does not own any fabs, so if the fabs AMD uses are too far behind technology wise, then AMD can likely also demand a lower price per wafer. Also, the biggest problem for AMD has been that it hasn't even been able to even produce competitive CPUs in the mainstream high performance segment. However due to 1) and 2a) that might change in the coming years.

 

[frozentundra123456]

Let's play hypothetical and say that AMD could/does catch up AND surpass Intel's high-end-consumer CPU line in 3 years time.

My bet is that Intel would go into war-mode with their multiple billions in potential R&D spend vs AMD and take back the performance lead in a shockingly short period of time. It would be great to be a consumer in such a climate.

...here's where i get shamelessly creative: AMD could then go out in a blaze of glory selling the CPUs at just a sliver above COGs (cost of goods sold)... which would force an unheard of frame-shift in the price consumers are willing to pay for CPUs as lightening fast performance could be had for insanely low prices.

The money fountain (billions in profits) for the CPU makers would be cut off and AMD might then be able to compete since nobody would be raking in wads of cash hand-over-fist.

:hmm: I like this alternate reality I've painted for myself.
Please, no one get too serious debunking this

Dear AMD... engage!

Sort of like rebuilding after a nuclear holocaust, right? Seems like cutting off your nose to spite your face.

 

[Idontcare]

b) Idle power consumption: Actually the process technology advantage does not have much impact on this. The reason is that at idle basically everything is power gated anyway.

I'm curious what percentage of "idle time" my 3770k's cores are truly power-gated. In task manager I see what appears to be a perpetual unending background activity around 1-3%.

The cores are reported as being clocked to 1.6GHz, but activity is going on. So long as an ethernet cable is connected to the internet and the A/V software is regularly rebuffing various IP/port queries, I suspect my cores are rarely in a truly power-gated state.

Active-idle power is my concern, and I hope it does continue to decline.

 

[Homeles]

It does matter, since if they can achieve it then they can also introduce it in the next CPU generations (if they want to that is).

Yes, but that has nothing to do with the discussion I was having...