backtop


Print 9 comment(s) - last by CrazyBernie.. on Feb 2 at 1:20 PM

My thoughts and the thoughts of others regarding the additional "Penryn" details

Quite honestly, I don't get much hate mail.  This week I did. Between Life with Penryn, my Hyper-Threading article and a few other roadmap leaks we've provided over the last few days, I am running out of time to individually address them all. Let me address a few common emails I've received.

With regard to Penryn being "delayed" -- several publications took up the story that Penryn is delayed, and that shipping it in 2008 is some kind of tactic to cover the weakness of the processor.  Several people emailed me, including writers at those publications, asking me why I didn't run with this story.  For the record, I did, it's right at the top of Welcome Back Hyper-Threading.  Secondly, is it really a delay? 

The roadmaps claim shipping in Q4 2007, with the launch date in 2008.  This is exactly the same phase AMD used to describe the ship date of 65nm Brisbane last year -- so I am a bit inclined to say nobody will see these processors until 2008.  Even then, is it delayed?  The date everybody had in their head was 2007, but almost every CPU in the last 15 years has taken at least a year from tape-out to launch date.  The Penryn tape-out was November 2006.

With regard to Hyper-Threading:  A slide to the roadmap supplied to DailyTech, available here, is one of the documents we used to verify the information in question.

However, I received a call today from someone running damage control claiming first generation Penryn will not have Hyper-Threading, and that the roadmap is incorrect.  Our source claimed "We will not discuss future products, but the DailyTech article that had references to Hyper-Threading on 45nm Penryn is incorrect.  The reference to Hyper-Threading was based on earlier data that has since been removed from Intel guidance."

We apologize for any confusion this may have caused.


Comments     Threshold


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

Better not to fly too high
By frent on 2/1/2007 12:23:23 AM , Rating: 2
Kubicki, you are wrong when you say "almost every CPU in the last 15 years has taken at least a year from tape-out to launch date":
1. A64 had a 10-months delay.
2. Xenon had a 8-months delay.

Notice also that Penryn ANTICIPATED a bit the roadmap, since it was able to boot all the systems.


a. K8L will arrive. When? July 2007, not before.
10-months since tape-out, like it was for A64.

b. Penryn will arrive. When? August/September 2007!
If tape-out had been in November 2006, why does in the Hell Intel wants to decelerate the roadmap it instead managed to ACCELERATE!?

I read a lot of FUD from little scared AMD zealots!



Let we talk about HyperThreading.
It was a GREAT thing. With HT, it was like having a multiprocessor. Quick response from OS and programs, loading time decreased, just like having a multi-CPU!.
I then sold my Intel rig and made the switch to an Athlon64 single-core, but I missed it a lot. Sometimes I get that crappy OS freezings for a couple of seconds.
So, again, why in the Hell you people have to say HyperThreading is crap??! Ignorance is crap, or prejudice, that is.
I only see a feature that would help squeeze 100% off the CPU.
Maybe it will be less useful since dual core will be the norm, but it isn't harmful.
Maybe there really won't be... or maybe there would be something similiar, who knows? Time will tell.


Sorry for my 6-yo English. That's the best I can do on 5.22 a.m.




RE: Better not to fly too high
By KristopherKubicki (blog) on 2/1/2007 8:24:57 AM , Rating: 3
quote:
1. A64 had a 10-months delay.


I'm not sure where you read that. K8 had its first tapeout in early 2002, and the processor launched in April 2003. I'm not as sure about Xenon's launch history, however.

quote:
a. K8L will arrive. When? July 2007, not before.
10-months since tape-out, like it was for A64.

AMD claims Barcelona was taped out in early August. The launch date is tennatively "Q3 2007." I don't really think we can pull dates together out of that.

quote:
b. Penryn will arrive. When? August/September 2007!

I really have no reason to believe this. It seems pretty apparent to me that Intel is setting up for a simultaneous launch of 45nm on all its platforms in 2008. Even if the chips are ready earlier, they aren't going to do a low volume launch.

Likewise, I don't really see the company canabalizing its August/September back-to-school season with a new processor. The company has spent *years* setting up its Santa Rosa launch schedule for notebooks, so you can be very assured there will be no mobile product until Q1'08.


RE: Better not to fly too high
By Viditor on 2/1/2007 2:03:11 PM , Rating: 2
quote:
AMD claims Barcelona was taped out in early August. The launch date is tennatively "Q3 2007."

I believe (at least it's what my sources tell me) that Barcelona launches the second week of June (Q2), however I guess it's possible that this will be the shipping date and not the availability date (which does mean a July availability).
Charlie D. is saying availability in June as well however...
http://www.theinquirer.net/default.aspx?article=37...

As for Penryn, Intel has published a Q1 08 availability for quite some time now...


RE: Better not to fly too high
By Kragoth on 2/2/2007 12:41:35 AM , Rating: 2
quote:
Let we talk about HyperThreading.
It was a GREAT thing. With HT, it was like having a multiprocessor. Quick response from OS and programs, loading time decreased, just like having a multi-CPU!.
I then sold my Intel rig and made the switch to an Athlon64 single-core, but I missed it a lot. Sometimes I get that crappy OS freezings for a couple of seconds.
So, again, why in the Hell you people have to say HyperThreading is crap??! Ignorance is crap, or prejudice, that is.
I only see a feature that would help squeeze 100% off the CPU.
Maybe it will be less useful since dual core will be the norm, but it isn't harmful.
Maybe there really won't be... or maybe there would be something similiar, who knows? Time will tell.


I think that you should investigate what HT is before blaming the non HT AMD chip for your decrease in performance. As it stands HT is purely a context switching device that enables the processor to have 2 processes held in the registers within the processor. Delays of 2 seconds as you are refering to have no relationship to HT. Context switching takes milli or nano seconds not 2 seconds. And technically HT would not decrease loading times of applications very much at all.

I think what you are arguing here would have a lot to do with the fact that when you changed from Intel to AMD you changed you BUS speed and this is where you may have suffered some decrease in performance relating to loading times.

There of some of us that are slightly bias towards AMD because of their architecture not purely because we are ignorant.

Sometimes its a good idea to research how much performance gain you get from HT before posting about how good it is. The best increase in performance HT gives you is on average about 3%. So it isn't all that fantastic.


By CrazyBernie on 2/2/2007 1:20:22 PM , Rating: 2
I realize my response is a bit late, but he said ALMOST.


By DallasTexas on 1/31/2007 2:44:56 PM , Rating: 4
Let's categoreze the 'nasty' mail you're getting and there is not need to apologize for your articles.

(1) Regarding "Penryn is delayed to 2008" and "Penryn problems is the reason" is coming from none other than AMD marketeers.

AMD's current strategy is FUD (Fear Uncertainty and Doubt). They have no solution to the Conroe onslaught until Barcelona and the only solution is FUD to delay QC purchases. The last thing AMD needs now is press on Penryn. They want you to write about "Penryn problems" to ensure Barcelona mindshare is not diverted to Penryn. Very simple.

(2) Hyperthreading. Who cares? I mean, really? HT is a great thing but multicore is what people want. HT enabled processors are just gravy for those applications that are HT aware. I feel the HT thing is a red herring.

In summary, you call it like you hear it and the spin-masters at AMD want you to help with Job #1 - which is stop the Core 2 promotion and focus on the future promise of barcelona.




Date correction?
By sdsdv10 on 1/31/2007 5:11:57 PM , Rating: 2
quote:
The Penryn tape-out was November 2007.


I believe you meant November 2006.




Northwood Take Two?
By JarredWalton on 1/31/2007 7:55:21 PM , Rating: 2
My bet is that Hyper-Threading is present on the cores but not fully validated and not enabled, just like it was on every 533 FSB up until... was it 3.06GHz P4 that enabled it first? Or like the EM64T stuff present in every Prescott but only enabled on the 5x1 and 6xx chips (and Pentium D, Celeron D 3x1, etc.) Anyway, it is highly unlikely that Intel will *ADD* HTT from scratch in the future, but not at all unreasonable to expect them to have disabled features present at Penryn launch that will be enabled later. Just my two cents, anyway - and screw the hate mail! ;)




By DeepThought86 on 1/31/2007 2:38:43 PM , Rating: 1
Looks like AMD will have a pretty decent window to reclaim server momentum for 1P, 2P and 4P then.




“And I don't know why [Apple is] acting like it’s superior. I don't even get it. What are they trying to say?” -- Bill Gates on the Mac ads

Related Articles
Update: Welcome Back Hyper-Threading!
January 30, 2007, 3:53 PM
Life With "Penryn"
January 27, 2007, 12:01 AM
Intel Samples "Penryn" 45nm Chips
November 28, 2006, 1:41 AM
















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