backtop


Print 84 comment(s) - last by saratoga.. on Feb 2 at 6:27 PM

More "Penryn" details emerge

Despite the plethora of attention Penryn received over the last few weeks, Intel's newest roadmaps put the processor launch for Q1'08.  This indicates the launch has not necessarily accelerated even though the initial tape-out proved extremely successful.

On the other hand, Intel's 2008 roadmap shows every segment simultaneously deploying 45nm products.  Like AMD's recent 65nm Brisbane launch, Intel guidance notes the processors will start shipping Q4'07 but the actual launch will come as a coordinated 2008 event.

The first Intel 45nm treatments will come from the quad-core Yorkfield and dual-core Wolfdale desktop processors.  Wolfdale has two physical cores on a single die and up to 6MB of L2 cache.  Yorkfield is then two Wolfdale dice on a single package. Also worth noting: Wolfdale ships with a 1333MHz front-side bus and Yorkfield ships with a 1066MHz front-side bus.  Chipset support will largely come from Bearlake-family that was previously disclosed on DailyTech.

Perhaps the most interesting thing about these two processors is the return of Hyper-Threading.  This, however, does not mean that Yorkfield will appear as eight logical cores, nor does it mean Wolfdale will appear as four logical cores. Intel's internal guidance on the subject specifically claims the processor will ship with Hyper-Threading, but will only utilize 4 threads.  On every Intel roadmap in the past, Hyper-Threading doubles the amount of listed threads in the guidance documentation.  Clearly, there is more of a mystery here still.  (Update: Please read the retraction below.)

"The official company policy is that our engineers have left the door open for Hyper-Threading, but we cannot confirm or deny any future plans for the technology," adds Intel Public Relations Manager Dan Snyder.

All Penryn cores also include Intel TXT, previously known as Intel LaGrande Technology.  TXT stands for Trusted Execution Technology and refers to the collection of devices.  The Trusted Platform Module, or TPM, is one component. DMA page protection is another. 

Alas, even if 2008 seems like a long time away for the 45nm platform, it's important to note that all Intel platforms will have 45nm SKUs in Q1'08.  Penryn, the family name for Intel's first generation 45nm consumer CPUs, also refers specifically to the 45nm dual-core mobile CPU.  Intel's current roadmap claims this processor will lead the Q1'08 mobile push with several low voltage models coming one quarter later.

For servers, Wolfdale will make an appearance as a dual and single socket Xeon.  It's been long-standing Intel policy to separate desktop, mobile and server chipsets into different products; Conroe was the Core 2 desktop CPU and Woodcrest, though physically nearly identical, was the Xeon counterpart.  Wolfdale as a server and a desktop CPU indicates the chips are electrically identical -- though each will likely receive different packaging for the different sockets. 

Yorkfield will not receive the same codenaming treatment as Wolfdale on the server. Instead, Harpertown will be the quad-core Xeon for two socket servers.  Yorkfield will still be the company’s single-socket quad-core Xeon offering.

Update 01/31/2007:  Channel sources have reached out to DailyTech to emphasize that the addition of Hyper-Threading to Penryn-family processors in 2008 is incorrect and the result of dated channel data.  My feelings and thoughts about the retraction can be read on my blog.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: Until there is proper software, HT is over rated
By TomZ on 1/30/2007 9:04:44 PM , Rating: 2
quote:
Unless you have a supporting O/S and properly written software HT is all marketing hype. As previously demonstrated, HT on Intel chips would slow most software performance until it was disabled.

Wrong. HT is useful in keeping an application running at 100% CPU from hogging the entire CPU. In my experience, that is more valuable than the few points of performance that is lost from having it enabled.


By Viditor on 1/30/2007 9:10:00 PM , Rating: 2
quote:
HT is useful in keeping an application running at 100% CPU from hogging the entire CPU

But how is this going to help in a multicore environment, especially on C2D?
I look forward to the benches, but at a guess I'd say that HT will be much less useful on a C2D chip...and keeping it disabled makes a bit more sense


By Nehemoth on 1/30/2007 9:29:06 PM , Rating: 2
Because HT and Multi Cores don't compete each other if not that they're complement each other.

For example in this scenario, HT will help that those MC's will always in use, course as someone said above the problem here is the FSB, but anyway i believe that this FSB will be fine until CSI appears, if someday will see the light at last.


By Viditor on 1/30/2007 9:45:00 PM , Rating: 2
quote:
Because HT and Multi Cores don't compete each other if not that they're complement each other

HT allows you to start 2 threads simultaneously, though you only have the resources to complete one...
MC also allows you to start multiple threads, but you have the resources to complete them all.
Add to that the ability of C2D to retire 4 ops/clock (though this never really occurs), and the only time that HT would help is when you have more threads that need starting than you have cores to start them...I can't think of many instances where this occurs.

With a single core, HT makes a lot more sense and certainly improves efficiency...but nowhere near as efficiently as MC does.

The long and short of it is that HT and MC DO compete...
For example, how will the scheduler choose between an HT (virtual) core and an actual second core? The second core is faster because it can complete the thread on it's own while the HT virtual core cannot.
In this case, using HT would actually slow things down...


By LittleMic on 1/31/2007 3:15:17 AM , Rating: 2
quote:

HT allows you to start 2 threads simultaneously, though you only have the resources to complete one...

That is where you are wrong : a single P4 core had enough ressources for more than 1 thread. That's why we've seen situations where HT increased performance.

This is even more true for Conroe that has a wider core.


By Griswold on 1/31/2007 6:58:24 AM , Rating: 2
The long pipeline of netburst made HT viable, parts of it were idling while others were busy - HT was just means to increase the load of the chip and thus its efficiency.

Now, if they would implement true SMT instead of this halfassed netburst approach, that would shed a different light on it. Until then, I'm with viditor and the fact that intel only considers this as an option that may never be used, supports it.


RE: Until there is proper software, HT is over rated
By Phynaz on 1/31/2007 1:10:55 PM , Rating: 2
What are you talking about? True SMT? What the hell is that?


By Griswold on 1/31/2007 2:18:45 PM , Rating: 2
So, you're telling me that HT is of the same quality as SMT in a, lets say, Power5 series processor or a Sun T1? No, dont answer me.

Compared to these two examples, Intels HT is a whackjob. Its fallout from the netburst design - it was a logical consequence of the design, but not remotely close to a "true" SMT design such as the Power5 or T1.

And as we can see in the update, there is no such thing in Penryn - for good reasons.


By saratoga on 1/31/2007 3:47:21 PM , Rating: 2
quote:
The long pipeline of netburst made HT viable, parts of it were idling while others were busy - HT was just means to increase the load of the chip and thus its efficiency.


Length and width are what matters. A long pipeline, or a wide pipeline are both good for HT. A long, wide pipeline is best.

quote:
So, you're telling me that HT is of the same quality as SMT in a, lets say, Power5 series processor or a Sun T1?


Define "quality". Both are SMT, though the T1's is very different then the P4's (which was more like the POWER5's).

quote:
No, dont answer me.


Huh?

quote:
Compared to these two examples, Intels HT is a whackjob. Its fallout from the netburst design - it was a logical consequence of the design, but not remotely close to a "true" SMT design such as the Power5 or T1.


Thats nonsense. Also, what is a "true" SMT design? I get the impression that you don't really understand how HT is implemented in the P4, based on your comparison to the T1.


By Viditor on 1/31/2007 7:06:16 AM , Rating: 2
quote:
a single P4 core had enough ressources for more than 1 thread. That's why we've seen situations where HT increased performance


No, it increased performance because it increases the efficiency of the CPU. By starting 2 threads at the same time, HT allows the CPU to have a partially run thread in the pipe at all times so that it doesn't have to wait for work (so to speak). Please note that I said complete one thread...
The problem you run into with HT and a multicore chip is that the operating system can't distinguish between a virtual core and a real core...



By LittleMic on 1/31/2007 9:54:47 AM , Rating: 2
quote:

The problem you run into with HT and a multicore chip is that the operating system can't distinguish between a virtual core and a real core...

Windows 2003 and linux have been doing this for a moment.

http://www.intel.com/cd/ids/developer/asmo-na/eng/...


By Viditor on 1/31/2007 9:54:13 PM , Rating: 2
It's true that affinity coding for the OS is becoming more widespread...but you (I think) are forgetting that the applications must be coded for this affinity as well...


By saratoga on 2/2/2007 6:27:55 PM , Rating: 2
Actually, no application support is required. The scheduler alone is all that is needed (since it assigns CPUs, not the apps themselves).


RE: Until there is proper software, HT is over rated
By TomZ on 1/30/2007 9:36:23 PM , Rating: 2
I agree, that the benefit of HT is less clear for multi-core. But I also think the performance hit for having it enabled is also overstated. I have a Presler EE, which is dual-core with HT, and I have HT enabled (four virtual processors). But I also tend to run quite a few apps at a time.


By Viditor on 1/30/2007 9:46:52 PM , Rating: 2
quote:
But I also tend to run quite a few apps at a time

That makes much more sense...but have you tested your speed while running single thread with HT on or off?
Or more telling, how about 2 threads?


RE: Until there is proper software, HT is over rated
By TomZ on 1/31/2007 9:33:39 AM , Rating: 2
No, I haven't taken the time nor the initiative to do that. I would be curious about the results, however. Can you suggest a good way to test that?


By saratoga on 1/31/2007 3:50:06 PM , Rating: 2
Your EE has more cache. More cache == less problems for the P4.

Theres not much point in testing HT with one thread. On a truely single threaded OS, HT and no HT are exactly the same. You need to actually have 2+ threads running before HT can be used. Otherwise the HT logic is just idle and you effectively have a non-HT processor. Sort of like how a dual core CPU run on Windows 98 is basically a single core CPU.


By TomZ on 1/31/2007 5:12:05 PM , Rating: 2
Regarding testing with one thread, etc., I'm not sure I get your point. I'm running Vista, and probably have dozens of threads actively running, plus maybe another 100 or more that are idle.


By Viditor on 1/31/2007 9:56:56 PM , Rating: 2
quote:
Can you suggest a good way to test that?


A fair question, and I'm sad to say that at this moment I do not (though if you post in the Forums, I would bet that someone like Duvie or one of the other gurus there could give you a good answer).


By Lord 666 on 1/31/2007 12:06:07 AM , Rating: 2
Wow, TomZ only has a Presler EE?

Very surprised you haven't upgraded to a C2DE or even quad yet.

You've previously said you use your rig about 70 hours a week and for your income, the new series of processors surely could help.

Honestly, why haven't you upgraded yet?


RE: Until there is proper software, HT is over rated
By TomZ on 1/31/2007 9:39:34 AM , Rating: 2
I don't think that spending the time or money now would make much performance difference. My processor has a WEI of 5.2, with all the other components of my system having WEI scores of 5.7-5.9. A CPU upgrade would surely increase the CPU score a bit, but probably not enough to be worth the time and money. I'll wait another year probably.


RE: Until there is proper software, HT is over rated
By mino on 1/31/2007 7:16:38 PM , Rating: 2
Do you REALLY think the WEI index is meaningfull way to measure performance by any single sub-score? Not to mention to consider their relation to the final score?

Go, read some AT before posting. May help a bit ;) .


By mino on 1/31/2007 7:23:01 PM , Rating: 2
A friend of mine once said:
"Well, IMO the WEI will replace 3DMark as an universal e-penis measurement tool."

Seems he was right.


By TomZ on 1/31/2007 9:03:08 PM , Rating: 2
Are you disputing my conclusion, my methodology, or just giving me grief for the heck of it? Are you saying that you feel that the WEI CPU score is a poor indicator of CPU performance? What are you saying exactly?


“So far we have not seen a single Android device that does not infringe on our patents." -- Microsoft General Counsel Brad Smith

Related Articles
Recent Intel Tidings, Retractions
January 31, 2007, 9:38 AM
Life With "Penryn"
January 27, 2007, 12:01 AM
Intel 45nm "Penryn" Tape-Out Runs Windows
January 10, 2007, 2:13 AM
AMD Announces "Brisbane" 65nm Processors
December 5, 2006, 1:27 AM













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki