Yonah article here on Anandtech

[fbrdphreak]

The Napa platform refresh will add Merom in 2H06. Merom will be pin compatible with Yonah and targeted towards thin & light form factor. Merom is targeted to support Windows Vista and EM64T functionality. Chipset will remain the same. Merom will feature 4MB L2 Cache, dual cores, Virtualization, Enhanced Intel Speedstep, EM64T, and the XD bit. High end Yonah chips should support Virtualization and the rest, but no EM64T. Merom will also be on a 667MHz FSB. Merom should feature architectural enhancements to make it more powerful than Yonah clock for clock, but I have no details on this yet.

Any other questions?

 

[malG]

Originally posted by: SynthDude2001
The things I worry about at this point are the viability of Yonah as a desktop CPU

I thought Yonah was supposed to be a notebook CPU?

 

[rstrohkirch]

Originally posted by: malG

Originally posted by: SynthDude2001
The things I worry about at this point are the viability of Yonah as a desktop CPU

I thought Yonah was supposed to be a notebook CPU?

re-read his post

 

[SynthDude2001]

Originally posted by: rstrohkirch

Originally posted by: malG

Originally posted by: SynthDude2001
The things I worry about at this point are the viability of Yonah as a desktop CPU

I thought Yonah was supposed to be a notebook CPU?

re-read his post

:thumbsup:

I'm quite aware that it's a mobile CPU. However, I am not in the market for a laptop and very well may never be. So it only (potentially) interests me in reference to the desktop segment.

 

[Avalon]

I didn't think it did half bad. I particularly like the very low power consumption. That's one cool running chip!

I also wish Anand would see if he could overclock the chip any, motherboard permitting. Would be nice to see how the chip scales.

 

[fbrdphreak]

Originally posted by: Avalon
I also wish Anand would see if he could overclock the chip any, motherboard permitting. Would be nice to see how the chip scales.

He is doing a Part II!

 

[Leper Messiah]

meh. Not bad intel, you kinda finally caught up to AMD. 'cept your proc is 6 months too late and is probably going to be uber expensive.

If it OC's to the heavens though I may bite the bullet.

 

[Avalon]

Originally posted by: fbrdphreak

Originally posted by: Avalon
I also wish Anand would see if he could overclock the chip any, motherboard permitting. Would be nice to see how the chip scales.

He is doing a Part II!

I'll be looking forward to it.

Leper: According to Anand, it costs Intel the same to make a Yonah @ 65nm as it does a Dothan @ 90nm, so we may actually see some decent prices on the lower models.

 

[mamisano]

Well, the Dual Core Turions will use DDR2 which should also lower power usage vs. DDR.

And what is this about Intels excuse for excluding x86-64? Vista is supposed to ship Q3 of 2006 and yet a mobile chip selling in Q1 will not be able to run the 64-bit version?

I guess the memo that Microsoft is moving forward towards dropping 32-bit did not make it to Intel's designers.

 

[DarkKnight69]

I dont think that Intel expected to walk out and crush the x2 with this laptop chip. However, this does show us that Intel is heading int he right direction. I would say that this is a impressive chip considering it's very low power and lack of onboard mem controller.

I think this chip will give us an idea of what they can do with memron. Intel knows they screwed up with netburst, and they know that they did well with Dothan.

If this chip O/C's well and runs cool, you might have a competitor to x2. But I wont hold my breath, i will wait for memron.

But this will make a kick ass lappy chip.

 

[fbrdphreak]

Originally posted by: mamisano
I guess the memo that Microsoft is moving forward towards dropping 32-bit did not make it to Intel's designers.

Since when is MS dropping 32-bit? There will be 32-bit & 64-bit versions of Vista, and probly the post-Vista OS. When at least 90% of computer owners are 32-bit only, you don't leave them in the dark.

Originally posted by: Leper Messiah
meh. Not bad intel, you kinda finally caught up to AMD. 'cept your proc is 6 months too late and is probably going to be uber expensive.

If it OC's to the heavens though I may bite the bullet.

You mean Intel is furthering its lead over AMD. Intel has been leading the mobile market for quite a while and AMD still has a long ways to catch up. Not as much in tech (but let's face it, Turion isn't exactly as advanced as the Pentium M is), but in market share & availability.

A lot of you are looking at Yonah as Intel's next competitor against the X2; it is just not so. Yonah is designed to the next generation mobile chip, offering performance never before seen in a notebook. Intel doesn't want (and nor will it ever) Yonah to compete with the X2 or other desktop CPU's; Yonah is going right up against Turion and the forthcoming dual core Turion. Wait for Intel's next desktop dual core CPU and then compare it to an X2 (Is it Conroe? I haven't kept up on upcoming desktop CPU's). I bet it won't have near the power consumption of Yonah and probably over AMD still.

 

[AkumaX]

hah, i can't believe so many of you were disappointed!! i honestly thought they did quite well considering i wasn't expecting ANYTHING from Yonah. In almost all the benches, most of the time it was 830 who? (i still cant believe it won in one)

power consumption was something i expected, you know Intel has something with these mobile cpus

i can't wait for next-next generation Intel to compete with today's AMD!!

 

[kamranziadar]

Looks like AMD is going to see some competition from INTEL. Low power consumption means less heat, it might be a good overclocker too.

Lets see when Turion will get dual core.

 

[Darth Farter]

ppl in the comments I have a statement which does not seem addressed and makes me feel like there is inconsistency whithin the article.

it is regarding the power consumption on the last page:

"While the Yonah and Athlon 64 X2 systems consumed relatively similar power at idle, Yonah hardly eats up any more power under full load. In fact, a 2.0GHz Yonah under 100% load consumes less power than an Athlon 64 X2 3800+ at idle"

well, yonah's platform uses DDR2 1.8V ram right?

well there should be a tangible difference in using DDR2 (1.8V) or DDR1 (2.6V) of the AMD X2 platform...

yet they focus mainly on the cpu's in the article (ok for performance, but hey on power consumption it should matter and ppl don't seem to realize it in the comments with some of their statements)...


if you think I'm wrong, please state why...

 

[Lyfer]

Looks promising, performance IMO is on par with the X2-3800 (except for the Video encoding tests). Intel should just clock this puppy higher and sell it as a desktop CPU. Far better CPU than what they to have offer for Socket T.

 

[dexvx]

Yonah matches X2's clock for clock. But lets extrapolate power consumption of Yonah:

Since its marked as "T" its TDP is 25-44W.

At Load: 144 -108 - 36W difference from the X2 3800+. The X2 3800+ has a TDP of 89W, but thats at maximum possible. Typical load consumption is about 75% of max. 89W * .75 = 66W. 66W - 36W = 30W.

2Ghz Yonah DC has about a 30W load consumption.

Originally posted by: Darth Farter
well, yonah's platform uses DDR2 1.8V ram right?

well there should be a tangible difference in using DDR2 (1.8V) or DDR1 (2.6V) of the AMD X2 platform...

if you think I'm wrong, please state why...

Power = Current * Voltage. Simple example, while processors voltage gets increasingly lower (the original Pentium operated at 5V compared to current P4's at 1.3V), the current pumped through gets higher. It would be erraneous to say the Pentium Classic sonsumes more power just because it operates at a higher voltage. In fact, it consumed about 10W of power compared to around 100W for current P4's.


---


About Turion vs Dothan. Their performance is about the same, clock for clock. Except the Turion cannot keep up with the Dothan performance/watt ratio. Turions consume about 50% more power at load. Dothans have a much longer battery life compared to similarly equipped Turions.

Celeron-M vs Sempron64-M. As the latest AT reviw indicates, a lowly 1.4Ghz Celeron-M can match a 2800+ Sempron64-M 1.6Ghz. And it still dominates in battery life.


---


Edit: One thing to note is that the Yonah has 2MB L2 total shared between 2 procs. Dothan has 2MB L2 for 1 proc (and faster L2).

 

[Zap]

Originally posted by: AkumaX
i can't wait for next-next generation Intel to compete with today's AMD!!

:laugh:

I read the article and felt that while it may make a powerful notebook chip, it isn't up to snuff on the desktop side of things.

Regarding performance, they were comparing 2GHz to 2GHz but counting total amount of cache the Yonah has twice the amount. Shouldn't they have gotten a Toledo core x2 and lowered the multilpier to 10x? That would have raised power usage, but also raised performance, plus I feel that it is a better match. Half the cache would be the Celeron M (or Celeron Y) version of Yonah.

Regarding power draw, did they take into account power saving features for the idle power draw? AMD has Cool and Quiet for desktops. I don't recall any desktop solutions for mobile Intel processors that were able to take advantage of Enhanced Speedstep. If the idle current draw is comparing full GHz Yonah to Cool and Quiet AMD, then Yonah becomes more impressive. If it is comparing full GHz to full GHz, then it doesn't look quite as good because Yonah already should draw less power using a smaller process than the AMD chips. If it is comparing full GHz, then the AMD should be even better at idle (and indeed even at low CPU usage) because the CPU can run slower and thus take even less power. That is, unless some motherboard manufacturer implements Speedstep into their desktop offerings.

Overall, not super impressed. AMD may be able to offer as low power draw (at least in idle) if it magically could produce a 65nm chip today.

 

[dexvx]

Originally posted by: Zap
I read the article and felt that while it may make a powerful notebook chip, it isn't up to snuff on the desktop side of things.

No, but you can tweak it for desktop usage. I have my Dothan Pentium-M at 2.7Ghz stock cooling on air. Some people with better HS/F's get 3Ghz on air. You will never see that in notebooks. Nobody knows how high Yonah will clock, but given its low power consumption (unlike Prescott's higher 100+W default), it looks like it has a lot of breathing room.

Originally posted by: Zap
Regarding performance, they were comparing 2GHz to 2GHz but counting total amount of cache the Yonah has twice the amount. Shouldn't they have gotten a Toledo core x2 and lowered the multilpier to 10x? That would have raised power usage, but also raised performance, plus I feel that it is a better match. Half the cache would be the Celeron M (or Celeron Y) version of Yonah.

Sempron performs so well because the K8 architeture is not cache dependant. Toledo vs Manchester at same clock speeds (X2-4600 vs X2-4800) wont have more than a 5% performance delta in most applications

Celeron Yonah's will have 1 core disabled. So it'd be like a Dothan.

Originally posted by: Zap
Regarding power draw, did they take into account power saving features for the idle power draw? AMD has Cool and Quiet for desktops. I don't recall any desktop solutions for mobile Intel processors that were able to take advantage of Enhanced Speedstep. If the idle current draw is comparing full GHz Yonah to Cool and Quiet AMD, then Yonah becomes more impressive. If it is comparing full GHz to full GHz, then it doesn't look quite as good because Yonah already should draw less power using a smaller process than the AMD chips. If it is comparing full GHz, then the AMD should be even better at idle (and indeed even at low CPU usage) because the CPU can run slower and thus take even less power. That is, unless some motherboard manufacturer implements Speedstep into their desktop offerings.

It is with C&Q, unless you're implying that the total system draw at idle is around 40W (assuming processor is always at full), which is just rediculous.

I doubt Anand is using a production motherboard. He's probably using a pre-production testing board that has the proper features enabled. Mobo manufacturers for desktop boards dont even have to use SpeedStep; because at full load it still consumes less power than a X2-3800 at idle.

 

[Leper Messiah]

Originally posted by: Avalon

Originally posted by: fbrdphreak

Originally posted by: Avalon
I also wish Anand would see if he could overclock the chip any, motherboard permitting. Would be nice to see how the chip scales.

He is doing a Part II!

I'll be looking forward to it.

Leper: According to Anand, it costs Intel the same to make a Yonah @ 65nm as it does a Dothan @ 90nm, so we may actually see some decent prices on the lower models.

Ooh. If intel doesn't screw us over, it might actually be worth it to buy one of these and stick it under a big arse HSFU. Can't wait to see the OC results though. Hurry up Wesely and crew!

 

[Hacp]

Originally posted by: dexvx
Yonah matches X2's clock for clock. But lets extrapolate power consumption of Yonah:

Since its marked as "T" its TDP is 25-44W.

At Load: 144 -108 - 36W difference from the X2 3800+. The X2 3800+ has a TDP of 89W, but thats at maximum possible. Typical load consumption is about 75% of max. 89W * .75 = 66W. 66W - 36W = 30W.

2Ghz Yonah DC has about a 30W load consumption.

Originally posted by: Darth Farter
well, yonah's platform uses DDR2 1.8V ram right?

well there should be a tangible difference in using DDR2 (1.8V) or DDR1 (2.6V) of the AMD X2 platform...

if you think I'm wrong, please state why...

Power = Current * Voltage. Simple example, while processors voltage gets increasingly lower (the original Pentium operated at 5V compared to current P4's at 1.3V), the current pumped through gets higher. It would be erraneous to say the Pentium Classic sonsumes more power just because it operates at a higher voltage. In fact, it consumed about 10W of power compared to around 100W for current P4's.


---


About Turion vs Dothan. Their performance is about the same, clock for clock. Except the Turion cannot keep up with the Dothan performance/watt ratio. Turions consume about 50% more power at load. Dothans have a much longer battery life compared to similarly equipped Turions.

Celeron-M vs Sempron64-M. As the latest AT reviw indicates, a lowly 1.4Ghz Celeron-M can match a 2800+ Sempron64-M 1.6Ghz. And it still dominates in battery life.


---


Edit: One thing to note is that the Yonah has 2MB L2 total shared between 2 procs. Dothan has 2MB L2 for 1 proc (and faster L2).

The "domination" in battery life is due to the whole platform Put the celeron M with x200 graphics and see how well it does in terms of battery life

 

[Hacp]

Originally posted by: fbrdphreak

Originally posted by: Furen
If you guys remember GamePC's very old Turion review

OMG don't pay any credence to that thing. It was written by a company trying to sell their own product!

Yonah is really kind of where I expected it. mArch improvements have yielded a slightly more efficient design (hence Yonah keeping pace with/exceeding Dothan @ +.13GHz clock), dual core performance is there but the full performance is traded off for power saving. For people who use laptops exclusively (kind of like me, except I do it only 'cuz I'm never at my desk), this will be nice to improve performance. But from what I'm hearing, the default power state is single core only on battery anyway. Yonah is a great step in the right direction, but until I can see dual cores yielding acceptable battery life I won't rejoice. Personally I'm waiting for merom

Exaclty how I feel. I'm sure Intel could have come up with a product that had a larger bus and higher clocks to beat the AMD desktop chips, but that would kill the point of notebooks wouldn't it? Yonah right now is a notebook chip, and IMO, power consumption is the most important part of notebooks. Those of you dissapointed in Yonah have been expecting something from yonah that was totally rediculous.

With the power of dual core processors in the hands of Intel notebooks, it should be interesting how the TUrion dual core can compete, not in terms of performance but in terms of power ussage and battery life.

 

[Zap]

Originally posted by: dexvx
No, but you can tweak it for desktop usage. I have my Dothan Pentium-M at 2.7Ghz stock cooling on air. Some people with better HS/F's get 3Ghz on air.

Yes, I'm quite aware of the overclocking potential of the Dothan chips. However, A64 chips are capable of those speeds as well. Power draw would go up on both platforms.

Originally posted by: dexvx
It is with C&Q, unless you're implying that the total system draw at idle is around 40W (assuming processor is always at full), which is just rediculous.

I'm not implying anything. I asked, "did they take into account power saving features for the idle power draw?" I ask that because I wanted to know. The author of the article would know what settings he used more than either of us. Why would it be rediculous for an x2 CPU to draw 40W at idle? I don't know how much power a CPU draws at full MHz while idling. Do you? I don't think an A64 CPU would draw the same 89W at idle as at full load even if Cool and Quiet were turned off.

 

[Markbnj]

I guess the memo that Microsoft is moving forward towards dropping 32-bit did not make it to Intel's designers.

Maybe the memo that Intel isn't ready yet didn't make it to Microsoft's designers .

 

[dexvx]

Originally posted by: Hacp
The "domination" in battery life is due to the whole platform Put the celeron M with x200 graphics and see how well it does in terms of battery life

For some reason I doubt the X200 alone is the reason why the Pentium-M dominates. The Turion has a higher TDP than the Pentium-M. But the integrated graphics is also a double-bladed sword... people here lament on how poor performing Intel is in 3d games. However, they fail to realize it is optimized for battery life.

Originally posted by: Zap
Yes, I'm quite aware of the overclocking potential of the Dothan chips. However, A64 chips are capable of those speeds as well. Power draw would go up on both platforms.

No my point was that if you're putting the chip in a laptop, you are severely restricted in cooling options, and thus clock speed.

Originally posted by: Zap
I'm not implying anything. I asked, "did they take into account power saving features for the idle power draw?" I ask that because I wanted to know. The author of the article would know what settings he used more than either of us. Why would it be rediculous for an x2 CPU to draw 40W at idle? I don't know how much power a CPU draws at full MHz while idling. Do you? I don't think an A64 CPU would draw the same 89W at idle as at full load even if Cool and Quiet were turned off.

This is a preview article... while he did not say C/Q was on, I think its a given that Cool and Quiet is on for every test system. Without C/Q, the idle/load consumption delta is rather small.

 

[stateofbeasley]

There are a number of things wrong about these assumptions people make.

First, Microsoft is not moving towards "dropping 32-bit." Windows Vista has concurrent x86 and x64 builds. Most people will not be able to tell the difference. I have not read any mention of the next Microsoft Office being an x64 program either. Lack of 64-bit will be inconsequential for all but power users.

Second, microprocessors are not designed and brought to market in a matter of months or a year. Yonah has been taped out since fall 2004 and been undergoing testing and tweaking since then. The design specs were probably finalized before Banias (original Pentium M) became a shipping product. In the late 2002 early 2003 timeframe, there was no indication that x64 would gain the momentum it did.

I guess the memo that Microsoft is moving forward towards dropping 32-bit did not make it to Intel's designers.