afaik dual Turion with ddr2 will hit the market in q3 2006. yonah is comin in january, thus Intel will continue to dominate the mobile market.
Zebo
Elite Member
- Jul 29, 2001
- 39,398
- 19
- 81
Absolutly - I've been predicting victory for Intel in later half of 06' since I saw Pentium M on that asus/DFI board. If and when they bring it to desktop in desktop form it will have everything I like about AMD chips - effceint, low power, and fast. All that matters after that is price and how high they can clock.. Clock for clock they (merom/conroe) should be superior - but if they don't clock 2.8+ well AMD still has a chance - if they go over 3.2 it will be like nothwood C vs. A-XP again.
Which is less important than the P-M using less than half the power of a Turion ML and 30% less power than a MT in the mobile market.
http://forums.anandtech.com/messageview...atid=28&threadid=1744373&enterthread=y
http://babelfish.altavista.com/babelfis.../le-turion-dans-les-desktops/page4.php
Zebo
Elite Member
- Jul 29, 2001
- 39,398
- 19
- 81
Bogus test - AMD gets DDR2 you'll see the same. I forget who but I'm looking but - xbit or digilifemaybe- anyway when measured at the 12V loop with a current probe AMD turion and Pentium used about same power.. as usual Intel underating TDP and AMD overating. That's the only accurate way since boards and mem config are so different currently.
BTW you need some math lessons 99W is not half of 147W let alone less than half and neither is it 30% less than 124W.
Intel may PAPER launch Yonah as much as a week before Turion2 (DC Turion) comes out (more likely 1-2 days), but I'd bet money that availability of Turion2 will happen first.
AMD officially says dual core Turion in EARLY 2006
Acer Ferrari Turion2 Notebook
New stepping of Athlon64 X2 shows up with extra low voltage range (meaning its probably the new Turion2 stepping), 1MB L2 per core, running default speed (2.2GHz) at 1.1V, which a normal X2 uses for 1.0GHz under Cool n' Quiet, and overclocking to 2.8GHz at default 1.30-1.35V
http://www.xtremesystems.org/forums/showthread.php?t=78744
From your link, Dothan TDP is 27 watts, Turion MT is 25. How is 27 30% less than 25 (or half of 35)? And that's not even counting the facts that
a) the power consumption of the Turion includes the memory controller, which is instead consumed by the chipset on the Pentium M, and
b) AMD TDPs are the maximum that the boards need to be designed for, i.e. the peak the top-end CPU in the line might possibly use maxing it out with CPUBurn, while Intel rates TDP as power consumed under more normal conditions.
Even the paid-off Intel shill Tom's Hardware found the Turion to run substantially cooler than the Dothan in their tests: http://www.de.tomshardware.com/mobile/20051126/index.html
I was refering to the CPU. You compare the difference between idle and full load. The 2.2 Turion ML has a difference of 42W, the 2.2 Turion MT is 26W and the 2.13 P-M is 17W.
And going to DDR2 is not going to save 30W, which is the difference between identically configured Asus A6 series laptops, one with a P-M the other with a Turion ML running Doom 3 at the same frame rates.
Empirical measurements trump paper specifications. Notebooks based on the P-M uses significantly less power than Turion based notebooks under load.
Heat is a function of the cooling system. The Dothan uses less power in the CPU intensive tests, despite having a discrete video card, and was generally faster.
I don't understand how the difference between full load and idle is relevant...
Also, the Turion has the Northbridge onboard, thus skewing the numbers again...
The battery life as tested in the article I posted show that the PM is 25% more efficient when at max load, and 10% less efficient when at idle...
Remember that 95% of this is a function of the platform and not the CPU...
What's this about yonah paper launch after turion DC.... yonah is practically out the door right now.
Not surprising considering most are HIGHLY optimized for Intel processors. So think of it this way, Intel can barely win while getting a huge hand.
2006 looks pretty bad for Intel and their good ol' partner Dell. Intel essentially handed Dell as few of their worst quarters ever this year, expect even more, and even worse. Dell can hardly keep up with returns on all of the Prescott system meltdowns. Hence the third case revision this year...
It seems everyone is ignoring the fact that Intel has a significant lead in development as far as 65nm goes. Will it make a difference? Who knows, Intel struggled with the Pentium 4 desktop shrink to 90nm and that didn't help them much. Then again, if the roadmaps I've seen are accurate, you should see Yonah and other 65nm parts coming with about a 6 month head start on AMD. If Intel successfully transitions to 65nm that will make a huge difference in terms of pricing and performance. While architecture does make a big difference, in the past Intel's silver bullet has always been their work in shrinking processes.
AnandThenMan
Diamond Member
- Nov 11, 2004
- 3,991
- 627
- 126
It's the same old tired script. Intel is coming out with a new wunder-system "soon" to either get up to speed with AMD or wipe the floor with AMD etc. etc. Same old same old when it comes to a die shrink. Intel has proven one thing in the last 3 years. They are unable to produce anything NEW or competitive with AMD. Also, A die shrink does not change the underlying architecture.
People assume 2 things incorrectly.
1. Intel is going to show their true colours and outperform AMD "soon"
2. AMD is standing still just waiting for Intel to catch up.
Perhaps AMD just has better ideas, better mangagement, better engineers, and better long term vision. But most people can't seem to accept this and assume Intel will get out of their rut just because they are Intel so they have to.
People assume 2 things incorrectly.
1. Intel is going to show their true colours and outperform AMD "soon"
2. AMD is standing still just waiting for Intel to catch up.
Perhaps AMD just has better ideas, better mangagement, better engineers, and better long term vision. But most people can't seem to accept this and assume Intel will get out of their rut just because they are Intel so they have to.
It gives you a good indication of the CPU difference by cancelling out the different components.
Doesn't matter, it's factored in the differencing. The memory controller uses on the 915 uses a few watts at most, while the Turion ML uses over 20W more power than the P-M.
You're number 2 was answered by Viditor. It was the late release of Turion (and its newness to the notebook market) that hurt it. The Pentium M motherboard and notebook designers have had multiple years to refine the process. The AMD Turion is a clock-for-clock equal performer to the Pentium M. The battery life is more poor notebook creation process than AMD's Turion. Remember, AMD lists max TDP while Intel does average so they are about equal. The use of DDR2 helps the Pentium M but not significantly. If AMD matches Yonah's power number on the 90nm process that spells frightening news for Intel down the road in '06.
And what makes you say that? AMD has already demonstrated to the press their in-house 65nm CPU production (more than a month ago). Just because they aren't producing 65nm parts in volume doesn't mean they can't...it can also mean (and probably does) that they have determined that they will make a greater profit by tweaking the process and improving to mature yields before volume production. After all, they really have no pressure to rush 65nm out the door since their 90nm chips are already more than competitive in both temp/heat and performance...
How do you figure? Are you assuming that under full load the video and I/O function at the same level as they do at idle?
Where do you get the 20W number from?? Is this from your assumptions above?
I don't see your point...your saying that the Thinkpad Z60t has entirely different numbers, but of course it does as it has entirely different video! Also, for the comparison they set both laptops to the same settings...
"For the battery performance test, the power scheme was set to 'Portable/Laptop'. This activates either AMD's PowerNow! or Intel's Enhanced SpeedStep technology, which lowers CPU speed when not needed thus increasing battery life. Screen brightness and audio were both set to 50% and Wi-Fi & Bluetooth were turned on. Each test was repeated 3 times to ensure accuracy. Before each test was run, the laptop was rebooted and its hard drive defragmented. Each test was run 3 times, and the result is the average of the three runs"
The fact that they were so specific about every setting, and that they set EVERYTHING equal for the comparison, leads me to believe that the Laptop Logic test was very well done.
Yes, and because the P-M processor is the least leakiest x86 processor, and also the highest performing/watt, and has a power efficient supporting infrastructure, Turion can't compete with it.[/quote]
I disagree...
So what if they internally have 65nm? Do you have any idea how long it takes to qualify a part for mass production? Intel showed that Yonah was working for months already; getting a few samples for pre-production analysis is one thing, releasing millions of them for the mass market is quite another. Even by Anandtech's own reckonning, Intel has a significant lead in 65nm development (see the Yonah article posted this morning).
AMD is currently competitive with Intel's 90nm offerings, the new Yonah article is up and shows that Yonah significantly cooler and consumes less power than AMD's desktop solutions. Sure, AMD can pull off tricks like voltage reduction to reduce power output, but I doubt they'll be able to match Intel's 33% power advantage under load in this segment.
This isn't to say that AMD is doing poorly, they certainly aren't but 2006 isn't going to be all roses for them if they just sit on their hands. Intel has a significant advantage in terms of market share in the increasingly important notebooks space and with Yonah it doesn't look like it will change any time in 2006.
If you read the same article you mentioned here at AT, you'll note that 90nm AMD is still MORE than competitive with Intel at 65nm...and please note that you are talking about a voltage advantage of a mobile part over a desktop one...!
AMD has been gaining marketshare in mobiles faster than Intel since Turion's release...
They gained a 75% growth from Q2 to Q3...
Yes, because the test is done is a CPU intensive application that doesn't hit memory or video.
It's from numerous measurements which indicate that the P-M uses less than 1W per 100MHz.
My point is that the Z60t, compared to a slower, less equipped Turion model uses less power in all typical usage tests. So you can't use that one test of the Acer models to universally say that Turion has better idle power usage.
I will probably get an M2 A64 middle next year partly because I know it performs ok and partly because it looks like the M2 socket will take a 4 core A64 in it's later life so good upgradability. However that doesn't mean I think AMD is doing anything special. The model numbers I've seen look like there will be no clock speed increases in the near future, and the only performance boost is from DDR2. I doubt AMD can do any really large increases until they move from 90nm to 65nm which isn't going to happen any time soon. So basically other then DDR2 and giving a +5% to +10% performance boost there isn't much happening in desktop AMD land we know of at all next year.
Intel on the other hand in Conroe have a brand new architecture on 65nm due out in the second half of next year. Considering a 25 watt yonah chip reviewed by anandtech is very close in performance to a 95W A64 X2 (both have same 2 ghz clock) you've got to assume the desktop version (Conroe) where they have 95W+ to play with would blow the A64 away. It's not all rosy (still no on chip memory controller) but I suspect AMD are worried.
Intel on the other hand in Conroe have a brand new architecture on 65nm due out in the second half of next year. Considering a 25 watt yonah chip reviewed by anandtech is very close in performance to a 95W A64 X2 (both have same 2 ghz clock) you've got to assume the desktop version (Conroe) where they have 95W+ to play with would blow the A64 away. It's not all rosy (still no on chip memory controller) but I suspect AMD are worried.
do AMD have answer for Intel's *T? These technology will surely give intel an edge in corporate and businesses and will surely an advantage for their SIPP
Maximilian
Lifer
- Feb 8, 2004
- 12,604
- 15
- 81
Yes AMD should be worried. Theyve had what, a year and a half? two years? since the introduction of prescott and since Intels options started to suck, and they still have a small market share. They could lose what little they gained when some actual competition shows up, will you all still buy AMD if Intel has the performance crown? I think not, unless their prices are extremely high. Assuming conroe will be competition to the K8 architecture, which it should be, if it isnt then screw Intel.
Hard Ball
Senior member
- Jul 3, 2005
- 594
- 0
- 0
According to this roadmap, Q4 is when Merom/Conroe platform will come out. And in the same quarter, AMD will start supplying 65nm chips for the M2 platform. And at the same time S1207 will be introed, probably with FBDIMM and such; either this OR the opteron+ core in Q2 07 is probably the K8L rumoured by the Inq.
http://www.mikeshardware.co.uk/RoadmapQ406.htm
And if you look further down the road on the map:
http://www.mikeshardware.co.uk/Roadmap20XX.htm
K9 seems to be pushed back to Q4 of 07, over a year later than the original plan. And as we know, the Whitefield seems to be cancelled, in favor of the Tigerton. So it seems that Intel will have the performance lead from Q4 06 -- Q4 07. Socket M2 will likely be underwhelming, not really worth looking at. S1207 chips, very little is know about them, except for LGA, and possibly separate mem controllers fro separate cores, and the eventual host of Quad core chips; so we need to wait and see if the inital Socket F chips are really anything revolutionary in Q4 06, and if they can hold off Merom/Conroe until a year later.
http://www.mikeshardware.co.uk/RoadmapQ406.htm
And if you look further down the road on the map:
http://www.mikeshardware.co.uk/Roadmap20XX.htm
K9 seems to be pushed back to Q4 of 07, over a year later than the original plan. And as we know, the Whitefield seems to be cancelled, in favor of the Tigerton. So it seems that Intel will have the performance lead from Q4 06 -- Q4 07. Socket M2 will likely be underwhelming, not really worth looking at. S1207 chips, very little is know about them, except for LGA, and possibly separate mem controllers fro separate cores, and the eventual host of Quad core chips; so we need to wait and see if the inital Socket F chips are really anything revolutionary in Q4 06, and if they can hold off Merom/Conroe until a year later.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 384
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter