Posted on 08/24/2010 11:53:03 AM PDT by Ernest_at_the_Beach
Youve sent in your questions and weve begun to sort through them to pull out the best. There were plenty of common themes that were arising, so well be grouping some of the bigger categories together. I am going to tackle some of the easiest ones first because some of the more technical questions will need to go to the engineers.
Well handle this blog in four rounds, with 5 questions each.
Lets get started.
There has been some confusion among those in the tech community regarding the actual CPU architecture, with modules and cores being explained differently by different people. Waffle911
Yes, there has definitely been some confusion about modules and cores. Modules are only our way of laying out the subcomponents of the processor. You will not see us market modules as they are largely invisible to everyone but the designers. Operating systems, for instance, will enumerate the integer cores, seeing a 16-core AMD Operton processor (currently codenamed Interlagos) as 16 cores, not 8 modules. Modules do impact the way that certain CPU features are addressed a discussion of which well save for a later date but in general we will focus on cores and not modules. The reason that we have modules is to help cut down on a lot of redundant circuitry in the processor. With multiple cores there is lots of duplication and this eats up die space and increases power draw. There are areas within the processor that can be shared because there is no major impact on performance, and other areas that should not be shared because they create bottlenecks.
You will never see a spec sheet with modules called out. Modules will not have a marketing name, they will only be Bulldozer modules. In reality, modules will only matter to the designers. Since we went out with Bulldozer information very early we focused on the shared architecture and talked at the module level (it is still far too early to be sharing die shots .) Because of this the two most misunderstood theories became a.) the module was the whole processor and b.) the module was somehow equal to one core.
When we talk about cores we will always be using the most agreed upon definition of cores the integer logic. Today most workloads are integer with a much smaller portion being floating point. This is why we focused on integer cores as the most logical way to define a core.
Each integer core will be able to run one software thread, and these threads can all be done simultaneously, unlike an SMT-type technology that lets two threads share one core. You typically find SMT technology on processors with much lower core counts, and its shared nature can create bottlenecks, even resulting in negative throughput in some cases.
As for core counts, here is what we have committed to at this point:
What are the virtualization advantages of Bulldozer relative to current AMD and Bulldozer time-frame Intel architectures? Muzaffer Kal
Well, to begin with, the competition has not revealed anything about their virtualization features in that timeframe so I will stick with AMD comparisons.
**************************************THIS IS AN EXCERPT************************************************
Since I haven't clicked the excerpt FR box...let me add the easily clickable link to this Blog here:
Well, I have to admit, this technical jargon is way beyond my comprehension, but I will say, there’s NO WAY this bulldozer is better than the one that ran over Rachel Corrie.
That one ought to be in the Bulldozer hall of fame.
1. ...Enthusiasts Website: ---AMDs Bulldozer and Bobcat Processors Previewand
2. ...Deeper into design ---AMD Discloses Bobcat & Bulldozer Architectures at Hot Chips 2010
Bobcat is aimed at lower end netbooks to compete with Intel's ATOM.
Bobcat is going to have an integrated Graphic engine on the chip...something Intel has NOT done well at all....
That is going to be a WILD market with a LOT of different possible microprocessors under the cover....
Bocat ----Q4 2010.....
Bulldozer --- Mid 2011....maybe like late Q2 2011....
The Sockets...well G34 and G32 are the ones for Servers....
Speculation --- AM3+ for Desktop and Workstations....
mmmm... processor advancement
Good stuff
************************************EXCERPT *************************************
The x86 core (Bobcat) of AMD Fusion APU Ontario will be based on Bulldozer architecture? Fabio Mendes
Actually, these are different designs. The upcoming Ontario processor will be based on the Bobcat core, which has a different core architecture than Bulldozer. There have been some that have made the assumption that a Bobcat was just a scaled down Bulldozer, but they are, in fact, different. Im sure that between the two there are similarities and some small sub-components that are shared, but you wont see the modular design of Bulldozer in Bobcat.
Hope there’s some eventual trickle down to lower Intel i7 hexacore prices :) Then it will be time for an upgrade on my main pc :P
And get the yield to make it available....
So you are Intel all the way ...I would guess.
***********************************EXCERPT**************************************
JC August 24, 2010
About the differences between Bulldozer and Bobcat I can inform that AMD have posted a nice video on the two architectures at http://www.youtube.com/watch?v=VIs1CxuUrpc It also does a great job at showing the differences mentioned in the first answer.
No never been ‘religious’ over processors, right now I have an older AMD X2 in my main home PC, work PC has a quad core Xeon and laptops are intel too. So not too picky on the CPU as you can see :). Just that the current best of the crop seems to be the i7, if I’m upgrading want it to last a while again, but its just quite expensive right now, including a new motherboard in the mix as well, would cost more than most complete PCs for just the CPU and M/B :)
Intel will have a fusion like product before AMD.
Intel was doing Modules 4 years ago.
More than that, I can’t say.
Heck, I want a 32 core CPU with a graphics card that can get me 50k Points Per Day doing Folding@Home without my electric bill jumping an extra $100 per month and my office feeling like a Swedish sauna at 120F !!!
OpenCL for CPU Scales Applications Easily on Four Six-Core AMD Opteron Processors
AMD OpenCL multi-core CPU demo at Siggraph Asia 2008
and from the comments.....
*********************************EXCERPT*****************************************
fireuser3d | February 03, 2009
The first public demonstration of OpenCL functionality was by AMD's FirePro / Firestream team at Siggraph Asia 2008. The particle & fluid simulation shows how OpenCL can extract high performance parallel computing out of GPUs, DSPs and multicore CPUs. Basically the idea is that you can write up your core computational code in OpenCL and suddenly you scale to whatever processors are available. Initially the demo only uses one core of a Dragon-based system (quad-core Phenom II). As the additional cores are enabled the simulation compute time is cut in half!
**********************************AND*****************************************
Modules, huh? It would be cool to order a CPU with four integer units per core and no floating-point units if you rarely do floating point. This is true for much of business processing. Conversely, scientific types could order a chip with four floating-point units per core and no integer units. Give an option to ditch the SIMD unit in favor of doubling your other units.
Each would still be able to do the other type of calculation, just a lot more slowly.
Electronista has a promo video about these, article sez Apple may be among first customers. Also sez that these won’t ship for 18 months, which will mean two gens of Intel by then, as well as one or two next-gen game systems, and who knows how many GPU iterations. Interesting that info about the programmable integrated GPUs...
http://www.electronista.com/articles/10/08/24/amd.bulldozer.to.match.sandy.bridge/
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.
@sonofhendrix AMD have published their OpenCL implementation (ATI Stream SDK), and it supports both their ATI GPUs and CPUs (where non-AMD models are not officially "supported"). nVidia also have an OpenCL implementation for their graphics drivers, but it is wasteful with CPU. CUDA is fairly mature, but OpenCL is far more portable.