Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
HT on the P4 showed some decent improvements once they tuned the OS schedulers to grok the difference between physical and virtual processors. It wasn't so much a throughput thing as a responsiveness thing -- if you had a CPU bound task permanently pegging the CPU then the system didn't bog down as much. I recall some benchmarks showing mp3 encoding at the same time as gaming showing a decent speedup, too...

I wasn't sure this would be as effective in a dual-core world, purely because instead of a virtual core taking up the slack on unused resources for responsiveness, you've got a real core doing it... so, if the actual throughput change is negligable and the responsiveness isn't really improved...? I guess if you now have two processes/threads that peg CPUs, but it seems unless it improves throughput, it's a diminishing returns thing...

Of course, HT was there to mask the huge latencies introduced by that massive 30+ stage pipeline, which Merom (and presumably Penryn?) isn't lumbered with...
I was out of the loop for that time. The short pipelines should so much more of a benefit.

Northwood was surprisingly capable in retrospect. Prescott was a dead end with the heat generation issues.
 
Presumably because Intel's engineers can see practical applications for it, or because Intel's marketing guys are pushing for it.
Well rumor is that it is only disabled on Core 2. It's a dirt easy way to add some more smoothness if it doesn't cost the ability to put in a logical core.
 
I was out of the loop for that time. The short pipelines should so much more of a benefit.

Northwood was surprisingly capable in retrospect. Prescott was a dead end with the heat generation issues.

Well, it's during pipeline stalls that HTT really shows the benefit, so the shorter pipeline should reduce the effectiveness -- assuming all else is equal, which we don't know yet.

As for Prescott, yeah... even Intel finally admitted that when they killed Tejas off.

I found that round quite amusing because I'd been predicting Intel would do just what they did for a couple of years or so :)
 
Well, it's during pipeline stalls that HTT really shows the benefit, so the shorter pipeline should reduce the effectiveness -- assuming all else is equal, which we don't know yet.

As for Prescott, yeah... even Intel finally admitted that when they killed Tejas off.

I found that round quite amusing because I'd been predicting Intel would do just what they did for a couple of years or so :)
I'm beginning to understand more the benefits of hyperthreading and its relation to the pipeline now. Thanks for that.

Now since OS X has rather good scheduling, it should benefit from it. Weren't the original Intel developer kits just Power Mac G5's with Pentium 4's in them? It'd be nice to get the word on OS X's effectiveness at using hyperthreading.
 
Well rumor is that it is only disabled on Core 2. It's a dirt easy way to add some more smoothness if it doesn't cost the ability to put in a logical core.

From memory, so this may be inaccurate (and I'm too lazy to look it up), adding HTT to the P4 required an extra 5% of die-size.

I do wonder about how noticeable it'll be on a shorter-pipeline dual-core CPU though.
 
From memory, so this may be inaccurate (and I'm too lazy to look it up), adding HTT to the P4 required an extra 5% of die-size.

I do wonder about how noticeable it'll be on a shorter-pipeline dual-core CPU though.
"According to Intel, the first implementation only used an additional 5% of the die area over the "normal" processor, yet yielded performance improvements of 15–30%."

http://en.wikipedia.org/wiki/Hyperthreading
 
The biggest issue is putting a quad-core chip that's cool enough into the form factor. If left unchecked a Core Duo can cook the insides at 190° F. It's designed to handle that but it's not fun to see.

We wanted to see Conroe in the iMac a several months ago. We know that the current form factor can at least handle a 45-55 watt TDP processor from the specifications of the PowerPC 970FX.

My guess is that the best quad core we can get for the next two years would fall into that upper area.

I remember, I wanted Conroe too, but it didn't happen. Obviously 65W is a bit too high. If they can't put a Conroe in it, maybe they'll put the 45nm shrink in, but they won't put in a quad core unless we are extremely lucky. At least until we either have a massive redesign of the iMac, or we have sub 60W quad cores. Of course, if we end up with a four core "Mini Mac Pro" (as I keep flogging, but it falls on deaf Apple shaped ears) then that could mean the iMac stays as a Mac Mini with a screen (not quite that bad, but you get the idea).

Apple seems to have decided that the iMac will be a silent, skinny, sexy computer, but doesn't care how fast it is. We can only hope that we find a quad core mini tower in the near future to fill the ever increasing gap between 2 core iMac and 8 and eventually 16 core Mac Pro.
 
(A whole lot of stuff about e-waste)

Personally, I like the idea of fuel cells (hydrogen fuel cells... For some reason Americans sometimes call fuel tanks fuel cells, but whatever), but on an unrelated track, hydrogen comes from either fossil fuels, or electricity, which comes from fossil fuels. No win. The reason I like hydrogen fuel cells is I like Hydrogen, because it can be made from water, and it can be used for nuclear fusion. Electricity turns water into Oxygen and Hydrogen, Hydrogen fuses into Helium or other Hydrogen isotopes, makes electricity. Water goes in, electricity, Helium and Oxygen comes out, with enough spare hydrogen to start a hydrogen fuel business on the side. No risk of Chernobyl style "Nuclear Meltdown", not that modern fission reactors are dangerous anyway. Environment wins, we win, everyone wins, except petroleum companies, but who gives a stuff about them? But everyone today hears the word Nuclear and frieks out. Totally unfair.

The main reason why fuel cells aren't used today is they have a terrible power to weight ratio. Hydrogen being a gas makes it almost impossible to store in anything but an extremely heavy pressure vessel. This is why Hydrogen will NOT be used in commercial passenger jets for a very very long time. Similarly it is expensive to originally purchase. In theory it is good for the environment, but no-one will buy it because it's just not practical. Like Nuclear fusion, the technology is not yet here to make it work. When it is, it will be awesome.

It is similar with computers. Materials are used because they do the job the best, with credit to affordability. Until there are materials that do the job better than the materials we use today that are also good for the environment, we are stuck with heavy metals and plastics. I might add also that Apple doesn't sell CRT screens, but I'm sure you alredy know that (well at least to my knowledge).

Now I've forgotten almost everything you have said, and am struggling to link random rants together into a proper argument, because I think I got a bit carried away.

Basically, computer manufcturers use what is available to them to make the best product they can. I find it unlikely that an "environmentally friendly" computer would even work. The best thing we can do as a community is to keep up efforts in improving recycling techniques, and eliminating any dangerous chemicals from products that don't have to be there, or can be phased out for something else. CFC's are a good example of this. It seems likely in the future that carbon compounds such as nanotubes will not only replace many more hazardous substances used today, but will perform far better at their intended role than the old material.

So yeah. I congratulate anyone who actually reads all of this way too long and tedious post.
 
I remember, I wanted Conroe too, but it didn't happen. Obviously 65W is a bit too high. If they can't put a Conroe in it, maybe they'll put the 45nm shrink in, but they won't put in a quad core unless we are extremely lucky. At least until we either have a massive redesign of the iMac, or we have sub 60W quad cores. Of course, if we end up with a four core "Mini Mac Pro" (as I keep flogging, but it falls on deaf Apple shaped ears) then that could mean the iMac stays as a Mac Mini with a screen (not quite that bad, but you get the idea).

Apple seems to have decided that the iMac will be a silent, skinny, sexy computer, but doesn't care how fast it is. We can only hope that we find a quad core mini tower in the near future to fill the ever increasing gap between 2 core iMac and 8 and eventually 16 core Mac Pro.
Well then again, Merom was a pin-compatible drop-in to Socket M. The GMA X3000 should resolve some of the graphical issues of the Mac mini. It's just going to take a stock RAM of 1 GB to really use it.

The issue is that the acoustic properties of the hardware are taken more into account with the post iMac G5 iSight internals. The Rev. A/B G5's sacrificed quietness for fitting a G5 into a 2" thick case. The redesigned internals are MUCH quieter even when using the PowerPC 970FX.

If Apple ever decides to put a desktop chip back into the iMac, it should work. It's just that they need to give up on trying to keep their machines so quiet.

iMac Internals
 
Well then again, Merom was a pin-compatible drop-in to Socket M. The GMA X3000 should resolve some of the graphical issues of the Mac mini. It's just going to take a stock RAM of 1 GB to really use it.

The issue is that the acoustic properties of the hardware are taken more into account with the post iMac G5 iSight internals. The Rev. A/B G5's sacrificed quietness for fitting a G5 into a 2" thick case. The redesigned internals are MUCH quieter even when using the PowerPC 970FX.

If Apple ever decides to put a desktop chip back into the iMac, it should work. It's just that they need to give up on trying to keep their machines so quiet.

iMac Internals

With a bit of luck, Apple will shove a quad core chip in the iMac, and give it a better graphics card. With a bit more luck, Apple will give us a box made for a quad core chip, which supports DX10 graphics cards, such as the soon to be awesome X2800 series, or whatever they will be called.

I should think that only one will happen, because one will in Apple's mind nullify the need for the other.
 
Average user reads emails once in a while,surfs few low demand websites,prints a recipe now and then,listens to a funny .mp3 song and plays a game of solitaire on sunday morning.Thats about it.

Average user rips a dozen DVDs to H.264 with two pass encoding and doesn't understand why each DVD takes hours and hours...
 
Penryn has no Hyperthreading.

According to The Inquirer: "Our previous article was wrong. Penryn does not have and will never have hyperthreading. "
 
Average user rips a dozen DVDs to H.264 with two pass encoding and doesn't understand why each DVD takes hours and hours...


Wrong.

Average users dont rip dvd´s. And they think that H.264 is birdflue.
What you describe is a poweruser.


There was a article that tried to study the character of the "Average" user.
Their conclusion in short was that "uses computer 3-6 times a week,10 minutes at time,uses browser for banking and checking out the local news and checks his/hers emails almost every time."

The average user is looking very much like my parents...


Erasmus,nowhere did i say "a 10 year old computer". I talked more about 5 year old ones...

Many programs will benfit minimally form the advent of the multicore-ism.
Some programs just dont translete well to that enviroment,just because their inherit nature. (mail,browsers..)
I would dare to say that cleaning up and optimizing code in those areas would translate to greater speed achievements than the implementation of the quad/octo cores... :)

BUT. It is not cost effective for the companies to do it,because there is abundant aviable power in the computing! And for companies like apple,who provides software AND hardware,it would be double cost prohibitive,because it would slow down the cycle of people updating your hardware!
Why bother,because everything just works?

And it´s NOW that we get back to the enviromental issues here...
:D


But let´s not get started on that..

Dont get me wrong, i am not a anti-developement kind of guy, (more of a sustained developement) but the prevailing atmosphere in the commerce is not healthy at the moment,and with the introduction of qartal economics it is straightforwar going to hell,imho.

[/end of rant]
 
Erasmus,nowhere did i say "a 10 year old computer". I talked more about 5 year old ones...[/end of rant]

OK, my point was this, as succinctly as possible. Theoretically, a 10 year old computer has enough power to do what your "average consumer" does. Today, however, software is so bloated with translucent effects and pretty animations that a 10 year old computer cannot run Safari, and even 5 year old computers (ie. pre- G4 iMacs) must be starting to feel the heat right about now.

The point is some time down the line even the average user is going to want more than 4 cores, as firstly even basic software will require more and more power to run (I mean just look at Vista!), and the "average user" will want to do more demanding things. My computer in theory should have no trouble running those timy little apps called Widgets, yet it is incapable of doing that little water effect when you drop one on the desktop. Of course this won't happen for a number of years.

This brings me to my second point, that the "average" user's computer will not have more than four cores for a number of years. Unless I've missed something, there are no quad core laptop chips in the works that we know of yet, at least until 2010ish. Today, the "average user" machine, the Mac Mini, uses laptop components, and out of date laptop components at that. The "prosumer" machine, the iMac, uses modern laptop components.

The point in conclusion is that low end computers will become 8-core just about when the average consumer starts to have a need for it, or it is forseen that the user will have a need for it within the computer's lifespan.

I do however agree that the "average user" does not rip DVDs in H.264 ;) but they would if they knew they could play it on their iPod Video :D .
 
Unless I've missed something, there are no quad core laptop chips in the works that we know of yet, at least until 2010ish. Today, the "average user" machine, the Mac Mini, uses laptop components, and out of date laptop components at that. The "prosumer" machine, the iMac, uses modern laptop components.

There are Penryn samples that are quad-core. Whether they're destined to be desktop or mobile parts, who knows.
 
There are Penryn samples that are quad-core. Whether they're destined to be desktop or mobile parts, who knows.

I believe Penryn is a code name similar to the name "Merom". Merom was actually the mobile part, but Intel ended up heaping the mobile, desktop and server "Core 2 Duo" architecture parts together, and just called them Merom, almost like saying they were the Merom generation.

Or take the name Pentium. How many unique chips can be heaped under that name? Focussing even just on Pentium IV, there's still Pentium M's and dual core Pentiums and 64 bit Pentiums etc etc etc.

I feel confident that we will find that the mobile part of Penryn will be dual core exclusive, with maybe a low power single core down the line. Only the desktop and server parts will have quad core options. When Nehalem comes out, it will also split into different parts, as will each successive generation, which tend to have one overall name, over the individual names of each part.
 
I believe Penryn is a code name similar to the name "Merom". Merom was actually the mobile part, but Intel ended up heaping the mobile, desktop and server "Core 2 Duo" architecture parts together, and just called them Merom, almost like saying they were the Merom generation.

Indeed - hence the second half of my statement. :)

Having said that, Merom was the codename for the microarch, but eventually they split them into Conroe, Woodcrest, etc.

Similarly, we already know of Wolfdale and a few others, so whether Penryn is intended as an umbrella name or not is rather unclear.

The fact these are Intel code names and not really intended for public consumption means the waters will likely remain muddy for some time.... Having said that, if they can do a low-voltage part with good yields, I wouldn't be astounded to see a high-end mobile quad core part within that generation.

Or take the name Pentium. How many unique chips can be heaped under that name? Focussing even just on Pentium IV, there's still Pentium M's and dual core Pentiums and 64 bit Pentiums etc etc etc.

Pentium, of course, was never a code-name, but a trademark and a brand name, so it's apples to oranges.

I feel confident that we will find that the mobile part of Penryn will be dual core exclusive, with maybe a low power single core down the line. Only the desktop and server parts will have quad core options. When Nehalem comes out, it will also split into different parts, as will each successive generation, which tend to have one overall name, over the individual names of each part.

I agree this is easily the most likely outcome, but I don't think it's a huge stretch to go quad-core. It's all down to the TPM really. Battery life wouldn't impress, but that's never stopped Intel before -- Pentium 4M and Celeron M were both less than impressive. There's a significant and growing desktop replacement market that intends to run off AC power for the vast majority of time. My former boss did just that with 17" Toshiba "laptop" that is enormous -- I suspect it can sleep a vacationing family of three if necessary.

Executive summary: I agree with you, I was mostly just being a smart-ass :)
 

I agree, I suppose. I am also often just being a smartass. It's fun.

OK, the Pentium bit was a bit of a stretch, but you seem to have understood where I was coming from.

There will undoubtedly be quad core laptops made using Penryn, just like there are laptops running X1900s and 19" screens. The thing is that Apple is unlikely to do this, at least IMO.

There is hope however for a quad core chip in the iMac. Hopefully we don't have to wait too long to have quad core lappies too.
 
I agree, I suppose. I am also often just being a smartass. It's fun.

OK, the Pentium bit was a bit of a stretch, but you seem to have understood where I was coming from.

There will undoubtedly be quad core laptops made using Penryn, just like there are laptops running X1900s and 19" screens. The thing is that Apple is unlikely to do this, at least IMO.

There is hope however for a quad core chip in the iMac. Hopefully we don't have to wait too long to have quad core lappies too.

Indeed -- I was disappointed, albeit unsurprised when the newer iMacs with C2D were Merom rather than Conroe. I'll assume it's a heat thang, since Conroe's both cheaper and faster... of course, it could be something related to discounts and sticking with Merom/whole Napa platform garnered huge discounts...
 
http://www.computerworld.com/action...ewArticleBasic&articleId=9010258&pageNumber=3

Finally, a mobile CPU known by the code-name "Gilo" (pronounced GHEE-lo) has raised a considerable amount of intrigue. Nothing is known about Gilo except that it is a 65nm CPU, leading to widespread speculation that this will be Intel's quad-core mobile processor. Intel has refused to confirm this speculation.
We we get a mention of Gilo. It might still be on the roadmap though as a cover for something good.
 
http://www.computerworld.com/action...ewArticleBasic&articleId=9010258&pageNumber=3

We we get a mention of Gilo. It might still be on the roadmap though as a cover for something good.

If Gilo is on 65nm process, then the die size is too big to go into a MBP, without a complete MB redesign. I doubt we'll even see a quad at 45nm, we'll have to wait until 32nm or smaller. Perhaps the economies are right to put a quad in a MBP 17in, but unless there is a huge performance advantage over dual core, I don't think it would make it into even the larger laptops from Apple.

As anandtech mentioned, Intel has been far more forward with letting out this kind of info as in recent past. Suggesting more of a marketing campaign is in operation here, considering falling profit margins and stiff competition. All good for us, but I can't see that 45nm is moving along much faster than indicated in the recent past. Late '07 or early '08 we'll see these 45nm chips in the MBP.

I'm more interested in Santa Rosa anyway, 45nm process is too far out in the future to be certain of anything (will there be delays in production up ramping, yields of the faster Ghz range, etc). Santa Rosa is supposed to ship this spring, and I want to see Apple ready to upgrade the MBP with that immediately.

http://www.computerworld.com/action...ewArticleBasic&articleId=9010258&pageNumber=2
 
To the best of my knowledge Kentsfield processors are only a little bit thicker then their Conroe counterparts. It's the same 65 nm process and die area.

I doubt we'll see Gilo for the Napa platform.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.