AMD Readies Phenom II Processors for Mobile Computers for May Launch.

[ajaidev]

http://www.xbitlabs.com/images/news/2009-12/amd_danube_chips_specs.gif

"Advanced Micro Devices plans to release its next-generation mobile platform code-named Danube in May, 2010, a source familiar with the company’s plans said. The new platform will at last feature processors based on AMD’s latest micro-architecture, which may make the new chips more competitive...."

http://www.xbitlabs.com/news/cpu/display/20091217191244_AMD_Readies_Phenom_II_Processors_fo r_Mobile_Computers_for_May_Launch.html

The phenom II N930 looks yummy :p: too bad i am more interested in congo/ION+culv platform more...

[Wesker]

TDP's are still way too high compared to Intel's Core 2 Duo line.

The P920 and P520 look interesting, but the lack of a turbo like feature really hurts the 1.6GHz P920 in apps that only make use of 1 or 2 cores.

[informal]

TDP's are still way too high compared to Intel's Core 2 Duo line.

Core 2 Duo is a dual core MPU... 45W for a Deneb class Quad Core core working @ 2.3Ghz is really really good.

[ajaidev]

i7-720QM is 1.6ghz and has a TDP of 45W top part has 55W tdp.

Somehow i am expecting the x920 BE to cost less than the i7 720QM. But the other cpu's will have a huge fight with mobile Core i5's...

[gallag]

Will the cache hurt them?

[ajaidev]

P520 has a misprint these cpu's dont have any L3, lack of cache will hurt performance specially in Games but not much in multimedia IMO.

Intel vs AMD mobile edition:-

Core i3-330M vs Turion N530
Core i3-350M vs Phenom II N620
Core i5-430M vs Phenom II N620
Core i5-520M vs Phenom II N620
Core i5-540M vs Phenom II X620 BE
Core i7-620M vs Phenom II X620 BE
-----------------------------------
Core i7-720QM vs Phenom II X920 BE

[Wesker]

Core 2 Duo is a dual core MPU... 45W for a Deneb class Quad Core core working @ 2.3Ghz is really really good.

As ajaidev mentioned; Clarksfield (non XE) also operates in a 45w TDP envelope, and has the ability to turbo up to 3.06GHz (Core i7 820QM) if all the cores aren't being taken advantage of. Furthermore, both the 720QM and 820QM feature Hyperthreading, with the 820QM also featuring a full 8MB of cache (720QM features 6MB of cache).

The only bummer about the Core i7 is that manufactures have so far restricted it to high end (typically ~17") laptops--also, they seem to be obsessed with pairing the Core i7 with Nvidia "guess the G80 derivative" GPU's.

[gallag]

it says in the article that they do not have any L3? That does not sound right.

Quote
"It should also be noted that next-generation mobile processors by AMD lack third-level (L3) cache. With relatively low clock-speeds (1.60GHz – 2.30GHz for quad-core and triple-core chips and 2.30GHz – 3.10GHz for dual-core processors), the new processors will hardly be able to offer truly extreme performance, however, will rise performance bar for AMD-powered laptops."

[zalbard]

Pretty nice numbers, actually.
I'd like to see some power consumption tests for the platform as a whole, though. :up:

[ajaidev]

Yep nice numbers, not much interested in the dual cores tough but the quads do look nice. The tri cores seem not that good a choice no sense in dragging a part of the cpu part in a laptop.

Phenom II N930 does seem the best of both worlds i expect it to cost much less than x920 BE and cost along the lines of $230-250 per chip.

Sweet

[Nedjo]

Yep nice numbers, not much interested in the dual cores tough but the quads do look nice. The tri cores seem not that good a choice no sense in dragging a part of the cpu part in a laptop.

Phenom II N930 does seem the best of both worlds i expect it to cost much less than x920 BE and cost along the lines of $230-250 per chip.

Sweet
'm very interested in A II X2 620BE @ 3.1 GHz ;)

[blindbox]

Core 2 Duo is a dual core MPU... 45W for a Deneb class Quad Core core working @ 2.3Ghz is really really good.

I totally didn't see this. Quadcore for notebooks, FINALLY!

[Tao~]

Any pricing estimates ?
Particularly interested because I dont like Arrandale's (expectedly) crappy Intel graphics chip.
But clock for clock AMD takes a beating, so a 2.6-2.7 GHz chip will be needed to compete with Intel's 2.4 Ghz one,thus almost cancelling the TDP due to no IGP.
N930 looks VERY promising but will need to be priced right(~$200). That and Mobility Radeon-5 series and we will RIOT Intel :p

[Manicdan]

anyone else feel they are missing a very key chip? the 45W tricore, that one should probably perform near the best since its clocks will be much higher than a quad so it wont feel so horrible when using programs that dont multithread well. while at the same time has the extra core for programs that do.

[gallag]

I totally didn't see this. Quadcore for notebooks, FINALLY!

I think the first quad for notebooks was in dec 2008, The q9000 2ghz 6mb cache 45w tdp, I have one in a acer 8935g. There is also the q9100, Q9200 and the QX9300.

[saaya]

quadcore in a laptop, rofl... puh-lease... :rolleyes:
the 1% who DO need a quad need perf and they will def go for a mobile 32nm quad with 8 threads from intel instead...

the P520 looks nice... above 2ghz is fast enough, and 25W is good... especially if its real tdp and now acp, so actual average tdp will be around 18W.

what i dont get is why there are no sub 2ghz 15W chips...
cause culv and atom is something amd COULD compete with very well...

[Meaker]

Can't compare core 2 and Phenom II TDPs at this level. You can compare it with i7 however:

Core 2 duo + NB + SB

Phenom II + single chip

[vietthanhpro]

BE= Black edition ! :rofl:

[Boissez]

About bloody time... AMD really should focus more on their laptop parts - we've had 45nm AMD parts for over a year now but they've been sorely MIA in the mobile scene although they could have been competitive with Intels current line-up. :shakes:

[naokaji]

Needs L3 cache... (at least for some of the models) we've already seen from reviews of the AMD desktop chips that the ones with less or no L3 suffer a lot from the lack of cache.

[generics_user]

Needs L3 cache... (at least for some of the models) we've already seen from reviews of the AMD desktop chips that the ones with less or no L3 suffer a lot from the lack of cache.

it's around 3% slower for dual cores and 5 on quads

great review with one test-rig + same clockspeeds + DDR3 by Computerbase.de (http://www.computerbase.de/artikel/hardware/prozessoren/2009/test_amd_athlon_ii_x4_620_athlon_ii_x3_420/)

i don't think that they suffer alot from the missing L3 cache; it's more like a win situation because amd is able to cut the die size by almost 50% and power consumption gets lowered too (around 5-10W)

it's great to see that amd is finally on par in mobile performance/watt; together with their superior chipsets they might be able to steal some market share from intel :up:

[terrace215]

it's great to see that amd is finally on par in mobile performance/watt

They aren't, and by May, when these are due out, they will be further behind.

[grimREEFER]

a 25w quad core!
damn that is nice!

although im not sure what 25w in amd tdp is like in intel tdp

[ajaidev]

a 25w quad core!
damn that is nice!

although im not sure what 25w in amd tdp is like in intel tdp

Its better thats wht it is.

AMD's TDP means MAX a CPU can draw where as Intels TDP is avg of the measures taken in some tests.

To be clear AMD's TDP is not really possible to achive in normal day to day activities and Intels TDP is totally achievable and can even be exceeded.

[generics_user]

They aren't, and by May, when these are due out, they will be further behind.

considering the fact that athlon II is on par with penryn 3m in terms inpoerformance/clock, their TDP is the same for the given clock ranges and that intel mobile chipsets suck compared to amd ones i think that they are on par; and core i3 isn't going to bring a huge performance leap over penryn ;)

[Boissez]

considering the fact that phenom II is on par with penryn 3m in terms inpoerformance/clock, their TDP is almost the same for the given clock ranges and that intel mobile graphics suck compared to amd ones i think that they are on par; and core i3 isn't going to bring a huge performance leap over penryn ;) fixed for inacuracies

[Manicdan]

i think that amd is a little late to the game, i was hoping once we saw C3, we saw mobile chips within a month. intel can release a 32nm counterpart for every 45nm mobile chip amd has.

[gallag]

Lets face it, When these launch in may 2010 they will finally be competitive with (or with the lower ipc than desktop ph2's maybe even slower than) what Intel launched in December 2008. Its the truth.

[accord99]

To be clear AMD's TDP is not really possible to achive in normal day to day activities and Intels TDP is totally achievable and can even be exceeded.
Reviews with power measurements disagree with this claim. If anything, it's the other way around, it's far more likely for an AMD processor to reach its TDP than an Intel processor.

[Tao~]

Predicting market position for May '10 is not possible, given we havent yet seen how much ground Arrandale 'extends' .

Given the choice of a 2.4 GHz Core i5 Arrandale dual and a 2.0 Ghz PhII Quad at same TDP, I'd gladly take AMD.

And Mobile quads are very much necessary, they havent pervaded just because of TDP/cost limitations.

Intel has taken the 'false quads' (aka SMT x 2 core) route, AMD might go for real quads, rest will be clear as the time draws near. :up:

[gallag]

Reviews with power measurements disagree with this claim. If anything, it's the other way around, it's far more likely for an AMD processor to reach its TDP than an Intel processor.

People still like to live in the p4/k8 days when this would have been true, Intel deserves flack for p4 though so its all good.

[FlanK3r]

nice to read! Whau, BE in notebooks

[gallag]

would be pretty cool if the BE overclockes well, If amd does some good combos with new 5xxx graphics cards and ocing cpu company's like alien ware will snap them up, A lot of marketing value in the first overclocker notebook.

[Hornet331]

Its better thats wht it is.

AMD's TDP means MAX a CPU can draw where as Intels TDP is avg of the measures taken in some tests.

To be clear AMD's TDP is not really possible to achive in normal day to day activities and Intels TDP is totally achievable and can even be exceeded.

Sry but thats not true anymore, just go to amd/intel and see what they write about TDP and how it is defined:

AMD:

TDP. Thermal Design Power. The thermal design power is the maximum power a processor can
draw for a thermally significant period while running commercially useful software. The
constraining conditions for TDP are specified in the notes in the thermal and power tables
http://support.amd.com/us/Processor_TechDocs/43375.pdf

Intel:

TDP should be used for processor thermal solution design targets.
TDP is not the maximum power that the processor can dissipate. TDP is based on
measurements of processor power consumption while running various high power
applications. This data set is used to determine those applications that are interesting
from a power perspective. These applications are then evaluated in a controlled
thermal environment to determine their sensitivity to activation of the thermal control
circuit. This data set is then used to derive the TDP targets published in the processors
datasheet. The Thermal Monitor can protect the processors in rare workload excursions
above TDP.
http://www.intel.com/Assets/PDF/designguide/318675.pdf (from the older xeon design guides since there it is written clearer)

Both companies use now applications to determin TDP, and both can be exceeded.

@ Topic
Nice to see a move on the mobile quad front, maybe leads to bit price preasure for intel there. :)
BE on notebook front is nice, it keeps the XE in check.

[naokaji]

it's around 3% slower for dual cores and 5 on quads


Depends on the review I guess, xbitlabs got a bigger difference here (http://www.xbitlabs.com/articles/cpu/display/athlon-ii-x4-630_11.html).

But looking at the power consumption difference with the desktop parts it becomes rather clear why AMD choose to cut the L3 on the mobile chips.

[ajaidev]

Sry but thats not true anymore, just go to amd/intel and see what they write about TDP and how it is defined:

AMD:

http://support.amd.com/us/Processor_TechDocs/43375.pdf

Intel:

http://www.intel.com/Assets/PDF/designguide/318675.pdf (from the older xeon design guides since there it is written clearer)

Both companies use now applications to determin TDP, and both can be exceeded.

@ Topic
Nice to see a move on the mobile quad front, maybe leads to bit price preasure for intel there. :)
BE on notebook front is nice, it keeps the XE in check.

kk still AMD's case is MAX and Intels case is AVG, this is in it self very different values even for the same chip "MAX vs AVG"


Depends on the review I guess, xbitlabs got a bigger difference here (http://www.xbitlabs.com/articles/cpu/display/athlon-ii-x4-630_11.html).

But looking at the power consumption difference with the desktop parts it becomes rather clear why AMD choose to cut the L3 on the mobile chips.

If true then these Phenom II's may be based on propus rather than deneb.

[generics_user]

Depends on the review I guess, xbitlabs got a bigger difference here (http://www.xbitlabs.com/articles/cpu/display/athlon-ii-x4-630_11.html).

But looking at the power consumption difference with the desktop parts it becomes rather clear why AMD choose to cut the L3 on the mobile chips.
xbitlabs states up to 10%; CB is 5% on average :up:

it's nice to see some proper mobile parts from amd and i can't wait for some reviews; my laptop is starting to show its age and i need a replacement in the next 6 months :(

[Oliverda]

Sry but thats not true anymore, just go to amd/intel and see what they write about TDP and how it is defined:

AMD:

http://support.amd.com/us/Processor_TechDocs/43375.pdf

Intel:

http://www.intel.com/Assets/PDF/designguide/318675.pdf (from the older xeon design guides since there it is written clearer)

Both companies use now applications to determin TDP, and both can be exceeded.

@ Topic
Nice to see a move on the mobile quad front, maybe leads to bit price preasure for intel there. :)
BE on notebook front is nice, it keeps the XE in check.

Read this (http://www.semiaccurate.com/forums/showpost.php?p=7949&postcount=86)

[zanzabar]

i7-720QM is 1.6ghz and has a TDP of 45W top part has 55W tdp.

Somehow i am expecting the x920 BE to cost less than the i7 720QM. But the other cpu's will have a huge fight with mobile Core i5's...

i think that is wrong, the tdp at stock with no turbo is 45W, at with no turbo that chip is slow the advantage of it is that with turbo is can go to 2.8ghz 2 threads and 2.2 or 2.4 with 4 threads and the tdp should be much higher like 65-70+W i would guess. and with the i parts intels tdp is equivalent to amds acp as its the average high load and not the 100% max that the chip can draw through software (intels tdp is closer to the max than amds acp to its max though)

[Hornet331]

Its nice what he writes, but thats not what amd writes in there own technical design specifications. :p:

Also he post kinda nonsense for intels TDP. ACP(AMD)=TDP(Intel) is a load of marketing BS.

Cause if you look at al the power consumption test out there, there are only 2 possibilities.

A) Intel drastically overstates ther TDP, since ACP=TDP by AMDs defenition, and still they consume nearly the same for the same Power bracket. (125W TDP P2

B) AMDs statement that ACP=TDP(intel) is plain wrong.

I wont comment further on this in this thread since its only derailment. This topic has been discussed so many times already and just always turns out the same way. :p:

[zanzabar]

i wasnt saying that they are the same in functionality just in idea. and the turbo comment stands, intel dosnt rate turbo into the power ratings and thats the main thing, as without turbo the i7 mobile is slow

[generics_user]

these chip don't have a ACP rating at all so i don't understand the point of your discussion

intel and amd have their own way of doing TDP rating wand we have no possibility to verify which one is max and which one not was we'd need a tool capable of measuring CPU power draw which isn't possible right now so discussing which tdp is real and which one is wrong is pointless...

amd still has ACP AND TDP ratings for their opteron line; acp is just a way to show real world (average) power draw @ 100% load while tdp shows the highest possible power draw possible

imo intels TDP is somewhere between ACP and TDP but as i said no one except for amd + intel really knows how they measure tdp so it's just my personal opinion

[Manicdan]

i really just want to see a new AMD mobile quad benchmarked, i remember hearing about how hot i7 laptops ran, and i think some people said they would underclock themselves cause of it.

[Monkeywoman]

ive never seen a sony vaio with an AMD cpu, i hope they start with these though, id like one of those quads

[tajoh111]

These can't come soon enough.

AMD needs much faster chips in the mobile space. I hope these rock hard because this is the segment AMD is getting most butchered in.

It too bad that these chips did not come out sooner those. ULV processors are taking over and high end mobile chips are becoming less and less popular and thus, less profitable.

[ExodusC]

These can't come soon enough.

AMD needs much faster chips in the mobile space. I hope these rock hard because this is the segment AMD is getting most butchered in.

It too bad that these chips did not come out sooner those. ULV processors are taking over and high end mobile chips are becoming less and less popular and thus, less profitable.
This is exactly what I was going to say.

I'd say what AMD needs is CPUs that draw significantly less power. You rarely, if ever, find AMD laptops that get better battery life than a similar Intel laptop. The mobile Core 2 Duo line is fast, and have a higher power draw to performance ratio compared to AMD's lineup.

Intel's CULV line of processors draw something like 10W of power, and are still faster than many AMD mobile CPUs.

AMD will continue to get slaughtered in this segment unless they push out a low power CPU than can still perform around where the current Core 2 Duos perform at.

Then again, everything will likely change with moble i3/i5. Things will get kicked up a notch. I'm terrified of seeing what a CULV i3/i5 could do (if Intel decides to sell something along this line).

[naokaji]

i think that is wrong, the tdp at stock with no turbo is 45W, at with no turbo that chip is slow the advantage of it is that with turbo is can go to 2.8ghz 2 threads and 2.2 or 2.4 with 4 threads and the tdp should be much higher like 65-70+W i would guess.

Turbo works nothing like on desktops where it's permanently on and on all cores, on the mobile chips its only as intel intended it to be, a temporary speed increase on a limited number of cores when not all are fully loaded to boost single threaded app perfomance, so the tdp should stay within the normal numbers as it won't kick in when you have all cores fully loaded.

[ajaidev]

This is exactly what I was going to say.

I'd say what AMD needs is CPUs that draw significantly less power. You rarely, if ever, find AMD laptops that get better battery life than a similar Intel laptop. The mobile Core 2 Duo line is fast, and have a higher power draw to performance ratio compared to AMD's lineup.

Intel's CULV line of processors draw something like 10W of power, and are still faster than many AMD mobile CPUs.

AMD will continue to get slaughtered in this segment unless they push out a low power CPU than can still perform around where the current Core 2 Duos perform at.

Then again, everything will likely change with moble i3/i5. Things will get kicked up a notch. I'm terrified of seeing what a CULV i3/i5 could do (if Intel decides to sell something along this line).

Hum not so sure about i3/i5 changing the game, most i3's and i5's "mobile" will perform around the x3's and i am sure that the x4's will beat them "multi thread only" If you want dual cores the Phenom II N620 will most likely be around the performance of most i3's and i5's given the 400mhz - 500mhz advantage the AMD chip has.

So if Intel does not release a 3 core part you can be sure that top of the list the i7 quad core will come and after that the x4's from AMD and not the i3/i5...

As for the GMA at best its performance will be at most around 9300 or ION it cant be more at default IMO "remember we are talking about the mobile part"

[grimREEFER]

if there is no ipc improvement from the current athlon neo, that 25w quad will actually not really outperform the current intel 25w dual cores in multithreaded apps actually. going by the cinebench scores, the current single core neo gets about 1300, factor in boost from having 4 cores, and it will score about 4500-5000ish, like a p8600.

[generics_user]

if there is no ipc improvement from the current athlon neo, that 25w quad will actually not really outperform the current intel 25w dual cores in multithreaded apps actually. going by the cinebench scores, the current single core neo gets about 1300, factor in boost from having 4 cores, and it will score about 4500-5000ish, like a p8600.

athlon neo is brisbane based (stepping G2) so it's slower than 3 year old windsor cores ;)

those new cores are athlon II (shanghai/deneb/regor) based so IPC should go up ALOT (20-30%) compared to neo and previos turions :up:

[informal]

athlon neo is brisbane based (stepping G2) so it's slower than 3 year old windsor cores ;)

those new cores are athlon II (shanghai/deneb/regor) based so IPC should go up ALOT (20-30%) compared to neo and previos turions :up:

Exactly,Deneb/Propus to the rescue^^.

[hyc]

Eh? Aren't these the same family as Turion II?

[ajaidev]

Eh? Aren't these the same family as Turion II?


no it is not, if you want you can call the K10/K10.5 also a K8 variant but if so i7 is also a variant of C2D??

[dartaz]

Those mobile tri-core and Quad-core Phenoms don't have L3 cache and they have low stock clock speed

I think even higher clocked hyperthreaded dual core can easily outperform most of these processors. Go and check the new thread called "First proper review of i3/i5", the i3 @ 2.8GHz often beat E7400 2.8GHz by %35-%50 in applications

And TDP is not very good measurement of power consumption. Just wait for reviews for Arrandale to see the actual power consumption

[ajaidev]

Those mobile tri-core and Quad-core Phenoms don't have L3 cache and they have low stock clock speed

I think even higher clocked hyperthreaded dual core can easily outperform most of these processors. Go and check the new thread called "First proper review of i3/i5", the i3 @ 2.8GHz often beat E7400 2.8GHz by %35-%50 in applications

And TDP is not very good measurement of power consumption. Just wait for reviews for Arrandale to see the actual power consumption

Go and see that thread again "The thread was mine" i have added the scores for Athlon II 620 and athlon x3 425 and it looks like i was thinking i7 on top followed by x4 and then x3/i5/i3 lastly x2.

Oh ya no mobile i3/i5's are suppose to be 2.8ghz, highest i5 is 2.5ghz but with 3ghz turbo.

[dartaz]

Go and see that thread again "The thread was mine" i have added the scores for Athlon II 620 and athlon x3 425 and it looks like i was thinking i7 on top followed by x4 and then x3/i5/i3 lastly x2.


Yea, it seems than Core i3 2.8GHz did better than Athlon II X3 425 (2.7GHz) in multi-threaded applications. In single-threaded and dual-threaded applications the difference should be much bigger

By the way, the mobile Phenom X3/4 stock clock speed are much lower than Arrandale. So, I expect the difference to be even bigger



.

Oh ya no mobile i3/i5's are suppose to be 2.8ghz, highest i5 is 2.5ghz but with 3ghz turbo.

Which is much higher than these mobile Phenom X3/x4

By the way, the fastest mobile dual core (Arrandale) is Core i7-620M which has stock speed of 2.66GHz (3.33GHz turbo)

[ajaidev]

Yea, it seems than Core i3 2.8GHz did better than Athlon II X3 425 (2.7GHz) in multi-threaded applications. In single-threaded and dual-threaded applications the difference should be much bigger

The mobile Phenom X3/4 stock clock speed aremuch lower than Arrandale. So, I expect the difference to be even bigger




Which is much higher than these mobile Phenom X3/x4

By the way, the fastest mobile dual core (Arrandale) is Core i7-620M which has stock speed of 2.66GHz (3.33GHz turbo)

Did you even read what i wrote "i7 will be top and followed by x4 and then x3/i3/i5" i7 no matter dual or quad it will be on top but will be expensive...

You are talking about dual cores and i am talking about i3/i5 not i7!!

Also i dont know where you read about turbo but 4 core @ turbo does not happen in mobile cpus, the default freq will be 2.5ghz which is 200mhz more than the 2.3ghz of AMD's top mobile x4 and the difference between i3 @ 2.8ghz and x4 620 is 200mhz.

[h0bbes]

As far as i know, the three mobile quads i7 processors can utilize "turbo mode" when not using all four cores

[Chock]

Got i7? :)

http://i210.photobucket.com/albums/bb51/ChockNorris/SAM_0020.jpg

[hyc]

no it is not, if you want you can call the K10/K10.5 also a K8 variant but if so i7 is also a variant of C2D??

You're talking total rubbish. There is no such thing as K10 or K10.5. There's "AMD Family 10h" and Turion II belongs to that family, same as Phenom / Phenom II. The previous line, Turion Ultra, was mostly a K8 but even that was "AMD Family 11h", not K8.