AMD Piledriver rumours ... and expert conjecture

Page 21 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Reynod

Reynod

Administrator
Mar 26, 2007
14,921
16
42,215
We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...
 
Sort by date Sort by votes
Wrong company. IBM.

Think about: IBM's triumphant return to the consumer desktop, with its brand new X86 based Cell processor. Now THAT would be something revolutionary. Plus, they get to expand into the GPU market, which will allow them to have a complete hardware platform for the first time since they lost control of the PC market.
Click to expand...

Actually I heard that one too. Right - IBM lost most of their PC marketshare during the '80s, to the likes of Compaq, HP et al. So they jettisoned CISC in favor of RISC, which is basically what Cell and PowerPC is, IIRC. But modern x86 is basically RISC-like too, once you get past the x86 legacy front-end. At least that's my understanding.

Anyway, considering that IBM sold off their PC division to Lenovo some 16 years ago, long before ATI was up for sale, it would be quite a jump for IBM to be interested in buying ATI back in 2005.
 
Upvote 0 Downvote
I find this comment to be some what false how can Intel keep scaling down x86 but Amd can't with their APU? Plus its way more possible for Amd to work with arm then Intel working with them.


SO I guess Intels no longer a target? Is he basically saying screw the CPU market? Because last I checked thats where they can make the most money.
Click to expand...

I think this means screw the high-end market. Amd can fight off Intel with their new APU's. They should also team up with arm.
Click to expand...

This would be a mistake. They will never gain market share that way, especially if they don't compete in the high end server market.

Phenom II was/is K10 and not K8..
to your earlier posting.
Click to expand...

Phenom was K10 but K10 was a "evolution" of K8. BD on the other hand is a new arch. Kinda like SB vs Nehalem. SB was based heavily on Core 2 with a IMC while SB is a major change from that.

One thing I will add on AMD's not competing with Intel, its more of an announcement on how stupid their marketing team was on trying to make people believe that BD would outperform SB-E.

One must hope that AMD doesn't drop out of the desktop cpu race alltogether, if they do, well, bye bye cheap cpus, seems to me Intel has no desire to make thier cpus faster as IB is the same clock speed as SB.
Click to expand...

No one wants them to. I think though that its their way of taking the attention off of the BD flop. There is no way AMD can not compete with Intel. Dropping the DT is not possible either. ATI cannot keep AMD afloat alone.

I guess we will see. The new CEO thinks they wont have to compete with Intel, they will. As I said before, I think Intel will push into the smartphone/table market before AMD will, especially if they keep using the same IGP as current phones are and get to 22nm with advanced power gating I think Atom will be a very decent mobile part.

Of course its all to be seen, but still I think AMD has a bit to go before Fusion hits the UMD market. Although it would be cool if you could get a smartphone and choose the CPU and GPU. Have a nice quad core CPU and ATI GPU. Yep that would rule.
 
Upvote 0 Downvote
why is amd still using around 9xx pin count, i think if amd can increase the pin count then it can increase its cpu's performance

for example
compare i7-39xx and sb, they have huge difference in pin count and also in perfornance

daily life example
consider a stadium with 10 gate and another one with 20 gate . which can alow more people to go in and out when they both are completely filled

so more pin count can make speed up I/O
are you agre with me
 
Upvote 0 Downvote
Hmmmm... I think where the 39xx and 2700 are similar they are using the same number of cores and where they differ is highly threaded workloads where the 39xx is using its extra pair. Not sure the number of IOs enters into it.
 
Upvote 0 Downvote
i am no certain but i7 39xx and 2700 are similar in some features and design and have a huge performance difference and that rings a bell that pin count can make I/O and processing faster
Click to expand...
the 3900 has 50% more cores thats why there is a big performance difference, per core there is very little difference
 
Upvote 0 Downvote
rumor of trinity (apu with piledriver cores) coming in march, new llanos in january.
http://www.xbitlabs.com/news/cpu/display/20111129235646_AMD_to_Start_Production_of_Desktop_Trinity_APU_in_March_Document.html
Click to expand...

Advanced Micro Devices plans to initiate manufacturingf its next-generation code-named Trinity accelerated processing units (APUs) for desktops in March, 2012. Initially, the company will target low-power systems and sometimes in May in plans to start manufacturing of high-performance next-gen Fusion A-series APUs.
Click to expand...

IIRC Intel says it has already begun producing Ivy Bridge, and it won't be available to us until March or April next year. So if AMD/GF is similar, my guess we won't see Trinity until at least 4 months later.

And from one of the comments from your link:

Hmm, AMD said earlier, that Trinity will be in shops at begginning of the year. So this statement 'begin production in March' looks like FIRST DELAY OF TRINITY.
Click to expand...

Not sure if it's true that AMD had Trinity scheduled for the "beginning" of 2012 however.

And that "20% higher" CPU performance compared to Llano's Star cores seems to be another result of BD's shared front end - great for mobile if the TDP is reasonable, but not for desktop where performance rules.
 
Upvote 0 Downvote
not really.
are you a chip designer or engineer.?
are you certain more pins means better performance.?
think about it and maybe research (Google) more before you answer.
I will just say this, LGA1366.
where does that fit in your hypothesis.?
Click to expand...

MOAR [strike]COARS[/strike] - er, PINZ!! 😀

IIRC over half the pins (or ball-grid connectors) on modern CPUs are power & ground, I think mainly for shielding purposes since the CPU operates in the GHz range. All those tiny connections would otherwise be tiny antennae and probably broadcast whatever you were doing with your CPU (Online banking? pr0n?? 😀) all over your neighborhood.
 
Upvote 0 Downvote
why is amd still using around 9xx pin count, i think if amd can increase the pin count then it can increase its cpu's performance

for example
compare i7-39xx and sb, they have huge difference in pin count and also in perfornance

daily life example
consider a stadium with 10 gate and another one with 20 gate . which can alow more people to go in and out when they both are completely filled

so more pin count can make speed up I/O
are you agre with me
Click to expand...
Your 10 gate stadium vs 20 gate stadium analogy works, if they are all being used. Having more pins allows more data to be transferred from the mb to cpu, but if the cpu isn't using it, there is no point.(no pun intended)
 
Upvote 0 Downvote
Advanced Micro Devices plans to initiate manufacturingf its next-generation code-named Trinity accelerated processing units (APUs) for desktops in March, 2012. Initially, the company will target low-power systems and sometimes in May in plans to start manufacturing of high-performance next-gen Fusion A-series APUs.
Click to expand...

IIRC Intel says it has already begun producing Ivy Bridge, and it won't be available to us until March or April next year. So if AMD/GF is similar, my guess we won't see Trinity until at least 4 months later.

And from one of the comments from your link:

Hmm, AMD said earlier, that Trinity will be in shops at begginning of the year. So this statement 'begin production in March' looks like FIRST DELAY OF TRINITY.
Click to expand...

Not sure if it's true that AMD had Trinity scheduled for the "beginning" of 2012 however.

And that "20% higher" CPU performance compared to Llano's Star cores seems to be another result of BD's shared front end - great for mobile if the TDP is reasonable, but not for desktop where performance rules.
Click to expand...
yeah, the timeline and the delay seem likely. earlier (before this article came out) i though i saw trinity's entry in wikipedia entry change from q1 2012 to h1 2012. i'd always wondered if i saw right nor not.
trinity's cpu performance improvement seems a little small considering the 'newness' of the architecture and the smaller 28 nm process.
 
Upvote 0 Downvote
Definitely pr0n!
Click to expand...


mega-icon-smiley-thumbs-up.jpg
 
Upvote 0 Downvote
Upvote 0 Downvote
1.2 Nillion transistors makes a lot more sense now out of the performance that they have been getting with the product ... does it not?

This would tend to suggest (unfortunately) that no amount of tuning or spins is going to improve the performance of the product other than improvements in clockspeed.

Unfortunately as it has a significant IPC disadvantage then a redesign is needed.

Infact they should probably just ditch it and spend time on Llano ... it seems a more promising product.





 
Upvote 0 Downvote
1.2 Nillion transistors makes a lot more sense now out of the performance that they have been getting with the product ... does it not?

This would tend to suggest (unfortunately) that no amount of tuning or spins is going to improve the performance of the product other than improvements in clockspeed.

Unfortunately as it has a significant IPC disadvantage then a redesign is needed.

Infact they should probably just ditch it and spend time on Llano ... it seems a more promising product.
Click to expand...
I agree. Bulldozer is simply just not good. I also will say, that for all the die space that Llano uses on graphics, it preforms very well cpu wise compared to PII x4 965.

http://www.tomshardware.com/reviews/amd-a8-3850-llano,2975-15.html

Also note how it is clocked .5 ghz lower in all the tests.


EDIT: In fact someone should do some math and find out how BD preforms clock for clock compared to Llano.
 
Upvote 0 Downvote
Okay im going to use Toms photoshop test for the comparison.

Llano (A-8 3850): 3.26 minutes * 4 cores = 13.04 minutes/core. 13.04 minutes*2.9Ghz = 37.816 minutes(rounded to 4 digits)/1 Ghz
Bulldozer (8150): 1.24 minutes * 8 cores = 9.92 minutes/coire. 9.92 minutes*3.6Ghz = 35.712 minutes(rounded to 4 digits)/1Ghz


Now consider: Llano has no L3 cache, is using more than 1/3 of its chip dedicated to graphics, and has a TDP of 100w
Bulldozer has no on die graphics, more transistors, and has a TDP or 125w


anyone else think Llano would have been good as just a cpu with a few tweaks?

Edit: fixed my calculations
 
Upvote 0 Downvote
Okay im going to use Toms photoshop test for the comparison.

Llano (A-8 3850): 3.26 minutes * 4 cores = 13.04 minutes/core. 13.04 minutes*2.9Ghz = 37.816 minutes(rounded to 4 digits)/1 Ghz
Bulldozer (8150): 1.24 minutes * 8 cores = 9.92 minutes/coire. 9.92 minutes*3.6Ghz = 35.712 minutes(rounded to 4 digits)/1Ghz


Now consider: Llano has no L3 cache, is using more than 1/3 of its chip dedicated to graphics, and has a TDP of 100w
Bulldozer has no on die graphics, more transistors, and has a TDP or 125w


anyone else think Llano would have been good as just a cpu with a few tweaks?

Edit: fixed my calculations
Click to expand...

(Photoshop use graphics)
i think so
remove dedicated graphics can make it a better cpu than bd
 
Upvote 0 Downvote
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom