AMD Trinity & Richland APU Discussion

[vickybat]

re: AMD Trinity & Richland APU Discussion

I don't even think they are in a very poor positition. 8-core Phenom II on 32nm would have steamrolled i7-2600k in multithreading. If they stop making wrong choices, they'd be fine to go.

I don't think its that easy to pull off. Costs and tdp will go sky high. Then i would say add two more cores to 2600k and see it steamroll the competition.

I guess it doesn't work that way. Adding cores isn't like applying butter on bread. Even intel hasn't made a proper 8 core retail chip yet.

 

[Skud]

re: AMD Trinity & Richland APU Discussion

What he tried to say is that instead of wasting energy and resources on Dekstop BD, they should have concentrated on a 32nm 8-core Phenom chip (II or III or whatever), which looks achievable. It's not applying butter on bread for sure, but this is the same company which have given us world's first quad-core below 100USD.

 

[skeletor]

re: AMD Trinity & Richland APU Discussion

I don't think its that easy to pull off. Costs and tdp will go sky high. Then i would say add two more cores to 2600k and see it steamroll the competition.

I guess it doesn't work that way. Adding cores isn't like applying butter on bread. Even intel hasn't made a proper 8 core retail chip yet.

Plain and simply wrong assumption.

Adding more on Skud's post, 8 core Phenom II on 32nm would have been ~220 sq. mm. Quad-core Llano which has 50% of die size devoted to GPU is ~220 sq mm again (CPU part is ~110 sq mm). The Llano cores on 32nm based off Phenom II are very power efficient. Just a much wiser alternative it would have been.

We have Bulldozer with 320 sq mm die size, TDP sky rocketing due to unproven architecture, and not even performing. Phenom II X8 on 32nm would have been the wiser choice as Bulldozer for desktops wasn't even fixed.

Intel has already added two more cores to i7-2600k and seling it as SB-E. Pricing matters. They haven't priced it at $300 anyways.

 

[ranjitsd]

re: AMD Trinity & Richland APU Discussion

May be they must improve their product quality.my cousins 2nd amd went kaput
trinity is 125w and ivy is 77w for long run amd guy will be paying more then intel guy

 

[skeletor]

re: AMD Trinity & Richland APU Discussion

^^

Something on-topic at last. Trinity TDP isn't even out. Plus, Trinity's TDP contains a potent GPU.

On the mobile side, people are expecting 50% decrease. Won't happen so for desktop though.

 

[sukesh1090]

re: AMD Trinity & Richland APU Discussion

ranjitsd,
look at llano's power consumption and sb's power consumption,llano power consumption is less than and sb and llano is even having a potent gpu in it which can handle most of the games even DX 11 and take a look at the graphical performance of SB.SB is not a true fusion they are like giving two partitions in your tiffin box rather than selling them separately.but llano is true fusion where gpu can take the work of cpu which it can easily complete with lesser time.....

 

[skeletor]

re: AMD Trinity & Richland APU Discussion

^^

now definition of "true fusion" will also vary. Some might not even call Llano as "true fusion"

I think Sandy bridge to be as much a fusion as Llano to be. It's the same die after all. Your tiffin box analogy is fine for first generation Core i3. But then even Sandy/Llano are not "true" fusion as some might say.

May be they must improve their product quality.my cousins 2nd amd went kaput
trinity is 125w and ivy is 77w for long run amd guy will be paying more then intel guy

ever thought how much you have paid more by going for Pentium 4 rather than Athlon 64?

thats why intel rules, buyers need performace

was this the reason you bought Pentium 4?

 

[Joker]

re: AMD Trinity & Richland APU Discussion

I don't think its that easy to pull off. Costs and tdp will go sky high. Then i would say add two more cores to 2600k and see it steamroll the competition.

I guess it doesn't work that way. Adding cores isn't like applying butter on bread. Even intel hasn't made a proper 8 core retail chip yet.

seeing llano's trackrecord..8 core phenom ii would have been easier and better than bulldozer.

it is not bread and butter...lol...every1 knows it. amd should have released BD for servers and stuck with phenom ii "tock" for desktop. bulldozer fiasco would have been avoided.

shrinking is easier than a new arch altogether on a new node. bread and butter vs french toast.

 

[sukesh1090]

re: AMD Trinity & Richland APU Discussion

@ico,
i know that llano is not complete fusion but i will say a bit better fused than SB for at least.but i think first generation core series didn't have any thing like gpu with it.it was just cpu.isn't it?
@ranjitsd,
about your cousin's AMD processor going kaput is not a usual problem.those are some rare cases which may or even may happened with some people with Intel processors.if you see like that then my cousin is still using athlon 64 processor from last five years and he never took his system to any service man because of any problem.so they are just rare cases and processor going kaput may be due to a lot of factor may be the mobo is not correct or psu etc.,

yes it is true that shrinking phenom was a better option but it looks better option now because BD failed to give what expected but AMD really didn't know this thing back then when they were working on BD.at that time they never would have thought that BD is going to fail,even we were excited by looking at its architecture.the architecture AMD designed is good but they failed implement that correctly.in every work planning and then implementing it both are very important.if one of them fails then the whole thing will fail.

 

[Joker]

re: AMD Trinity & Richland APU Discussion

Intel p4 2. 4 ht,intel 865 gbf,1 gb ddr ram,seagate 200gb hdd,priya smps 400
w,prime focus atx cabinet,samsung 17 crt

may be his cousin used Mona SMPS 400 and....cousin of priya...i.e. mona wasn't as good as priya. :

 

[skeletor]

re: AMD Trinity & Richland APU Discussion

Shall we now move onto Trinity. What's been done is done. Let's not go off-topic now.

 

[ranjitsd]

re: AMD Trinity & Richland APU Discussion

ever thought how much you have paid more by going for Pentium 4 rather than Athlon 64?
My pc is 7year old without any problem check my smps.my cousins amd athlon is dead and his phenom 2.5ghz is with service guy
was this the reason you bought Pentium 4?

I didnt buy it my dad bought it cost 28k with apc ups and hp all in printer and iam in udipi in cousins house and he has 9500 gt without pc and crysis 2 and dues ex er in his external hard disk this is the only reason i hate amd

 

[skeletor]

re: AMD Trinity & Richland APU Discussion

ever thought how much you have paid more by going for Pentium 4 rather than Athlon 64?
My pc is 7year old without any problem check my smps.my cousins amd athlon is dead and his phenom 2.5ghz is with service guy
was this the reason you bought Pentium 4?

I didnt buy it my dad bought it cost 28k with apc ups and hp all in printer and iam in udipi in cousins house and he has 9500 gt without pc and crysis 2 and dues ex er in his external hard disk this is the only reason i hate amd

sorry to be rude, but you don't even know the right cause.

Most probably it's the PSU which is dead or the motherboard because of a bad PSU.

Read this: *www.thinkdigit.com/forum/power-sup...89-power-supply-blacklist-thread-newbies.html

If it died despite you having a fine PSU, then it is just bad luck.

 

[ranjitsd]

re: AMD Trinity & Richland APU Discussion

may be his cousin used Mona SMPS 400 and....cousin of priya...i.e. mona wasn't as good as priya. :

zebronics dude and iam waiting trinty bcoz intel is out of reach i dont want spend 10k on cpu with gpu lower than llano.

 

[Skud]

re: AMD Trinity & Richland APU Discussion

Last one: this one sounds better:-

AMD taps servers for future growth - ZDNet Asia News


I don't think they are going to leave the PC market altogether, may be just some "repositioning" to focus on more lucrative markets.

 

[Jripper]

re: AMD Trinity & Richland APU Discussion

Some relief that. :S It would totally suck if amd stops making proccy's for the regular users and just focuses on the server market.

 

[ranjitsd]

re: AMD Trinity & Richland APU Discussion

sorry to be rude, but you don't even know the right cause.

Most probably it's the PSU which is dead or the motherboard because of a bad PSU.

Read this: *www.thinkdigit.com/forum/power-sup...89-power-supply-blacklist-thread-newbies.html

If it died despite you having a fine PSU, then it is just bad luck.

psu is fine cant convince him to buy decent psu pc died due to high voltage from lightning

 

[vickybat]

re: AMD Trinity & Richland APU Discussion

Plain and simply wrong assumption.

Adding more on Skud's post, 8 core Phenom II on 32nm would have been ~220 sq. mm. Quad-core Llano which has 50% of die size devoted to GPU is ~220 sq mm again (CPU part is ~110 sq mm). The Llano cores on 32nm based off Phenom II are very power efficient. Just a much wiser alternative it would have been.

We have Bulldozer with 320 sq mm die size, TDP sky rocketing due to unproven architecture, and not even performing. Phenom II X8 on 32nm would have been the wiser choice as Bulldozer for desktops wasn't even fixed.

Intel has already added two more cores to i7-2600k and seling it as SB-E. Pricing matters. They haven't priced it at $300 anyways.

Hmmm what you say makes sense in your perspective not in amd's. Yes phenom 2 had a proven architecture and they had implemented its 32nm avatar in llano as you say is 110 sq mm.

But phenom 2 belonged to a previous gen era and its single core performance was always slower than nehalem and here we have sandybridge. Simply adding cores wouldn't have been enough because apps also needed to be heavily threaded in order to utilize core resources properly. What if the additional cores would have to sit idle in maximum cases?

We don't need a true 8 core processor yet. An 8 core phenom 2 would have been ahead of its time as far as core count goes.

Therefore amd designed bulldozer. They never build it thinking it would fail. If you keep on adding cores without changing architecture, it would surely look like the old gigahertz wars.

Amd needed to take a new step and they did. It didn't live upto the hype but atleast we expect them to learn from their mistakes and improve the design in piledriver and future iterations.
Offcourse trinity will have piledriver cores and i think we should take their approach positively.

Last offtopic post from me. Trinity only from now on

 

[skeletor]

re: AMD Trinity & Richland APU Discussion

Pretty much missing the point again. Post edited and PMed.

 

[Skud]

re: AMD Trinity & Richland APU Discussion

Lets bring the thread on track:-

AMD to Start Production of Desktop "Trinity" APU in March - Document - X-bit labs


*www.xbitlabs.com/images/news/2011-10/amd_trinity_advantages.jpg


Four 65W families, A10-5700, A8-5500, A6-5400 and A4-5300 and two 100W families A10-5800 and A8-5600 have been planned. Besides, Llano would be refreshed with unlocked A8-3870K and A6-3670K (note the 'K' ) as well as A8-3820, A6-3620 and A4-3420. There's also a disclaimer of sorts:-

The 20% speed improvement represents AMD's projections "using digital media workload" and actual performance advantage over currently available Fusion A-series "Llano" vary depending on the applications and usage models.

In all probability, Piledriver should hit the market before this. Or will this be the testing ground for Piledriver?

On a side note, it looks like GloFlo has sorted out the production issues of 32nm wafers.