HardOCP is going to preview Bulldozer on the 16th of July:
http://hardforum.com/showthread.php?t=1612784
Maybe now we can change the name back to 25 days t0 Bulldozer again

Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
HardOCP is going to preview Bulldozer on the 16th of July:
http://hardforum.com/showthread.php?t=1612784
I'm not comparing Llano to SB, I'm just wondering if there's something other than a difference in clock causing the rather large gap in TDP between low-end Llano and the higher-end models as they're all running the same silicon as far as I can tell.
interesting info
http://www.theinquirer.net/inquirer/news/2080551/amd-demonstrates-core-bulldozer-serverCHIP DESIGNER AMD chose the International Supercomputing Conference (ISC) to finally demonstrate a working Bulldozer system.
At AMD's ISC stand one could find several 2U and 4U servers built with older Opteron chips, but it was a 1U pizza box server made by Supermicro that housed two 16-core Bulldozer chips running live demonstrations of POVRay. This is the first time that AMD has publicly displayed its next generation Opteron processor, codenamed Bulldozer.
The chaps manning AMD's stand said that Bulldozer still has a Q3 2011 launch date and, judging by the fact that it has started to display working machines, we can assume that timeframe is not too optimistic. Asked whether AMD will be coming up with a Llano style Opteron featuring an accelerated processor unit (APU), AMD told The INQUIRER that "an Opteron APU still at least two years off".
Coming back to the present, Supermicro's 1U dual socket Bulldozer server packed in 32 cores, with space for an accelerator card. AMD said that other vendors are managing to put four sockets into a 1U server, resulting in a very impressive 64 scalar processing cores in just 1U.
Initially the chaps at AMD were not keen for us to photograph its naked server, however with The INQUIRER's winning smile the lid was lifted on the machine. One thing that surprised us was how cool the chips were running. We were able to touch the heatsinks with our bare hands. Given that this is a 1U server where cooling capabilities are stretched to the limit, that is a mighty impressive showing from AMD.
AMD was not willing to talk frequencies, cache sizes, cost or any other details without signing an NDA at this point. Although it wouldn't release exact thermal design power figures, AMD said that those will be the same as its Magny Cours Opterons'.
The 16-core Bulldozer chips should once again put pressure on Intel's Westmere EX processors, however Intel told The INQUIRER that it will be producing Sandy Bridge Xeons by the end of the year. And for AMD's Bulldozer, the litmus test will come not with Intel's Westmere EX line but next year, up against the Sandy Bridge Xeons. µ
Because by the time i put everything i wanted on it that's how much it cost.
Cannot believe these are still not on the market. Was interested to see how they were going to fair versus an SB build but if I had of waited I would have been pretty miffed by now![]()
Cannot believe these are still not on the market. Was interested to see how they were going to fair versus an SB build but if I had of waited I would have been pretty miffed by now![]()
The GPU in the LLano is vastly more capable than the one in Sandy Bridge ... it was frequently acheiving 50-200% higher frame rates in tests. Hence the difference. Also, Intel's TDP vis a vis AMD's is misleading. Intel's tends to be best possible case, AMD's worst case. Hence why in battery life tests (particularly under heavy load) the mobile LLanos actually outperformed significantly or matched similar Sandy Bridge stuff. Judging by LLano, AMD's 32nm process is pretty damned good.
Cannot believe these are still not on the market. Was interested to see how they were going to fair versus an SB build but if I had of waited I would have been pretty miffed by now![]()
Getting it right and getting it on time usually should go hand in hand.
No point getting it right, late, then being trumped by another chip maker. If you abstract brand loyalty why would you intentionally wait for BD?
Question is this, take a phenom 2 six core add two cores to it, make it 15-20% more efficient clock per clock and then ask the question is this going to be competite against Intels SB and IB?
I think the answer is this:
In windows and multithread apps it will probably be competitive (Possibly including games which scale up to 8 cores)
But in apps which are more focused on raw core speed, it will probably still be behind the SB.
(Incidently I was dreaming about BD and some weird variations of an iphone 3 last night)
If BD mobo dont support PCI Ex 3, then Im wating for the x79.
If im going to upgrade, im going to upgrade to tech which will allow me to upgrade to pci ex 3 gpus etc etc etc....
(Oh and none of this 8x 8x pci ex lane nonsense)
Are they bringing out a new CPU laptop range based on BD?
According to AMDs slides it can split 256bit AVX instructions over the two cores in a module, but I have no clue what that means in real terms to be honest. Anyone shed any light on that part? Or is that nothing to do with running threads?
IIRC it means that the FPU can do one 256-bit instruction, or two 128-bit instructions simultaneously.
edit: http://www.xbitlabs.com/news/cpu/display/20101026234515_AMD_Calls_New_FPU_Flex_FP_Defends_Dual_FMAC_Approach.html