• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

2.2GHz Chip

link fixed There's an extra 'p' in yours, worth.

I get a kick out of "Austin chip consultant Tom Whiteside said dynamic circuits are used only sparingly in most processor designs because they are so hard to do correctly." This cracks me up. It makes it sound like they invented dynamic circuitry. As if most CPUs aren't totally loaded with it. And, of course, we'll just forget the power aspects of using dynamic exclusively.

They are doing quad phase dynamic which is neat and honestly is hard. But this article doesn't mention that. Instead they say, "``It's a difficult thing to do right, and it takes brain surgeons to do it,'' said Whiteside, a former senior design manager at IBM who recently joined Intrinsity's advisory board."

I like this too:
" Intel Corp. is expected to announce a 2-gigahertz Pentium 4 processor this year, but Intel will use very expensive and specialized manufacturing methods to achieve that speed. " They make it sound like Intel's using GaAs or something. 0.18um 6-layer aluminum is hardly exotic or 'specialized' manufacturing methods. I'm curious what they consider 'non-specialized.

edit: by the way, a better link is this story at the EETimes here
 
erub, they said it's 32-bit CPU core, but they didn't say it's an IA32 CPU core. In fact, I highly doubt that it is x86 compatible (if it was, they would have mentioned it). I doubt it will run any of the standard benchmarks without recompiling.

One comment on my post. It sounds like I'm not impressed with what they supposedly have done. It's not that... I'm simply not impressed with the article. 🙂

They appear to be doing something new and different. It basically sounds like they found a way to do quad-phase dynamic in a standard cell library format. This is pretty impressive and I would love to know how they deal with the timing computation and power dissipation issues of this approach. But it is quite an interesting thing.

The 32-bit 2.2GHz CPU sounds like a bit of a publicity stunt. There are no details about it to be found as near as I can tell. But if they say that Intel's Pentium 4 is on a 'specialized' process (it's 0.18um), then that could indicate that they are running 2.2GHz on a 0.25um process. And if they truly are a 32-bit CPU core, then they might be doing a 32-bit add in one cycle at 2.2GHz on 0.25um. This would be pretty impressive. But this is a house of "what-ifs", and I will be interested in the details.
 


<< erub, they said it's 32-bit CPU core, but they didn't say it's an IA32 CPU core. In fact, I highly doubt that it is x86 compatible (if it was, they would have mentioned it). I doubt it will run any of the standard benchmarks without recompiling. >>



So if that's the case, its not really that exciting is it? I understand that alot of work went into it, but the impact it would have on the current and near future PC market will be next to nothing.
 


<< So if that's the case, its not really that exciting is it? I understand that alot of work went into it, but the impact it would have on the current and near future PC market will be next to nothing. >>

I wouldn't say that. It's exciting. Just not to AT'ers. 🙂 But you are right, RGN The impact it will have on current and near-future PC market will almost certainly be 'next to nothing'.

To VLSI circuit designers, it's an interesting proposition and it could mean interesting products in the future.... if they can get this out in widescale deployment before process scaling makes their dynamic method less of a clear win... and if the power is as low as they seem to imply (which I doubt). The article makes sense in the context of EETimes (an electrical engineering periodical). It only makes in the Austin American-Statesman - the source of the Silicon Valley Times article - because the company is located in Austin.
 
well I thought that I had heard somewhere on AT that there were many chips out there that ran a lot faster than x86 architecture chips - x86 was designed to do a lot of things mediocrely (sp?) but nothing really outstanding. If this cannot run regular code, then is this a big deal at all, even in the semiconductor industry?
 


<< well I thought that I had heard somewhere on AT that there were many chips out there that ran a lot faster than x86 architecture chips - x86 was designed to do a lot of things mediocrely (sp?) but nothing really outstanding. >>

IA32-compatible CPUs are among the fastest out there currently - especially in clustered tpmC (transactions per second, a networking benchmark) and integer applications. Most RISC CPU's struggle to keep up with IA32 CPU's in integer code. So, IA32 CPUs are very fast for general purpose microprocessors.

But I'm sure I'm right. This is not talking about an IA32 compatible CPU - if it was more sources would be carrying the story, and it would be big news. Instead it's just interesting. It appears to be at the least, a publicity stunt, and at best a technology demonstration, but it's not an actual product.
 
&quot;Have to wait and see, because it could be another vapor chip such as the E2k, IDT/C6, Transmeta&quot;

Transmeta is not a vapor chip. It is very popular in mini-notebooks in Japan as is used by:

Sony
Fujitsu
NEC
Hitachi
Casio
Toshiba
Sharp

Microsoft has also chosen the Crusoe as one of its chip suppliers for the tablet PC. AOL and Gateway have also partnered with Transmeta for set-top boxes and other devices.

Corporations that do business with Transmeta
 


<< well I thought that I had heard somewhere on AT that there were many chips out there that ran a lot faster than x86 architecture chips - x86 was designed to do a lot of things mediocrely (sp?) but nothing really outstanding. >>

Yep, I wouldn't sell x86 so short...it may have a lot of architecture limitations (small # of registers, register/memory instruction format, stack-based floating point instructions, etc), but clever engineering can overcome most limitations. The main thing seperating higher-end RISC designs from x86 is larger caches, higher system bandwidth, and better scalability with multiple processors.



<< x86 was designed to do a lot of things mediocrely (sp?) but nothing really outstanding >>

The same could be said for any general purpose CPU. While current instruction sets support a variety of arithmetic instructions (did you know that x86 actually has tangent and arc-tangent functions?), adds/subtracts, multiplies, shifts, and logical instructions are used far more often than any other arithmetic instruction for general purpose CPUs. Specialized MPUs, such as networking and video processors, use more specialized pipelines and arithmetic operations. For example, I think 3D video processors make heavy use of matrix dot-product multiplications. Therefore, a video processor can have specialized dot-product hardware, which would take many more instructions to complete with a general purpose CPU using adds and multiplies.
 
I think this is very exciting news regardless of whether or not it's x86. If the chip can be acquired cheaply and a motherboard that takes standard pc components can be designed for it (IDE hard drives, DIMM memory, PCI expansion cards, etc) then this type of processor could make for a VERY fast Linux workstation. With all the source code available Linux doesn't rely on legacy compatibility nearly so much as Windows does. I'd love to have one of these.
 


<< Transmeta is not a vapor chip >>



You are right that Transmeta is not vaporware, but it might as well be. Transmeta promises extreamly low power consumption (&quot;it was promised to more than double the battery life&quot😉, which isn't so, and they also forgot to mention that it have to be underclock. Transmeta also forgot to mention that in real world performance it would add an aditional minutes to a laptop...not hours. And, stupid people didn't have a clue regarding computer bought it up, thinking that it is a superior solution to AMD/Intel...An underclock K6/Duron/Celeron combine with clock throtling, consume just as little power and is cheaper.
 
&quot;...Transmeta is not a vapor chip. It is very popular in mini-notebooks...&quot;

I agree with Jeffrey. The Transmeta Crusoe is a kickass CPU that *does* deliver on the low power consumption promise.

&quot;...Code Morphing software [has] the ability to adjust Crusoe?s voltage and clock frequency on the fly depending on the demands placed on the Crusoe processor by software. Because power varies linearly with clock speed and by the square of voltage, adjusting both can produce cubic reductions in power consumption, whereas conventional CPUs can adjust power only linearly (by only adjusting the frequency)...&quot;

Take a look at Sony's C1 &quot;Picturebook&quot;!

Fit Windows ME, Linux, or Windows 2000 Pro in your pocket at only 2.2 lbs...!

Running on a 667 Mhz CPU that is PASSIVELY COOLED! (no fan!!!)...

Awesome! 😀 😀
 
it gets faster as the morph-ware kicks in....

anyway, i'm sure they could use their process methods on x86 based processors, right?...
why couldn't they - it's an MPU like the rest...they could also use it on gpus, etc...

where's the details?
 
Would be *sweet* if Nvidia &amp; Transmeta teamed up to make a powerful GPU for mobile users that used these powersaving techniques 😀
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top