• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

unlocked llano's available for pre order!

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
This is most interesting because the "Husky" cores in Llano have more IPC than the "Thuban" cores in Phenom II, which has more IPC than Bulldozer cores. So if an unlocked version can reach ~4GHz these should outperform any bulldozer offerings

Not possible, the cache system is gimped in Llano which negates IPC gains by a pretty large amount.
 
This will interest me too. I wonder if I could live with one of these (with overclocked IGP) for a LAN party rig? 😕 Would be sweet to show up with my rig VESA mounted on the back of my monitor.

Like those guys that show up with their damn iMacs. I mean, yeah, you can mock them about it but they walk in with a bag in one hand and the mac in the other. When its time to go, same thing, no three/four trips ferrying crap back and forth...
 
i'm more curious of how high the cpu will oc. deneb/thuban stops 4.2 at the highest ive seen (according to threads/sigs)

It will be interesting to see how these pan out 🙂

You would think 4.4 Ghz.....on the relatively new 32nm process.

Then maybe 4.8 Ghz as GF perfects the node?
 
I'm not really seeing how Llano is considered faster clock for clock?

http://www.anandtech.com/bench/Product/399?vs=80

The IPC improvements are mostly negated by the lack of L3
or in other words, the IPC improvements compensate for the lack of L3
also notice the 2 in that link are roughly 3.4% appart MHz wise, which correlates closely to the performance difference. So pretty close ultimately when not relying on cache loving benchmarks/apps.

matching the performance per clock of the X4 with half the cache is the idea
more with less

also note that the L2 on llano is larger than the L2 on the phenom
this could have a large effect at higher clocks for apps that can "fit" in that L2
 
Last edited:
Like those guys that show up with their damn iMacs. I mean, yeah, you can mock them about it but they walk in with a bag in one hand and the mac in the other. When its time to go, same thing, no three/four trips ferrying crap back and forth...

That's what a notebook computer is for (which I have a Clevo M980NU). And I already make only one trip for LAN parties. Got a handle for my mini ITX rig and a handle for my monitor, plus a bag for keyboard/mouse/etc.
 
I assume they're doing this because they announced that 45nm CPUs are End of Life. They need something to fill the void since people aren't going to buy BD.

And I assume they need to kill 45nm due to what they said about 32nm / 45nm tool sharing cutting into volume & impacting Llano. They are selling all the Llanos they can make (mostly for notebooks, of course,) and this can give them an upside in volume AND an increase in gross margin.
 
???

How so?
I'm not really seeing how Llano is considered faster clock for clock?

http://www.anandtech.com/bench/Product/399?vs=80

http://www.anandtech.com/bench/Product/399?vs=105

Check the bench above ... an Athlon II x4 with 2mbs of cache can do the same as the 3850 at lower clock speeds, or in some cases beats llano with less cache AND lower clocks ... the IPC gains show themselves nowhere so here's how I see it. If we're lucky these cores are so weak 1mb of cache per core will not hinder performance, but the better bet is that llano will behave like its closer relative the Athlon 2 when you overclock it. The higher your core clock gets the more of a bottleneck the lesser cache becomes.
 
Last edited:
It will be interesting to see how these pan out 🙂

You would think 4.4 Ghz.....on the relatively new 32nm process.

Then maybe 4.8 Ghz as GF perfects the node?

Without knowing the targeted clockspeeds when they shrunk the stars core, we really can't say whether they intended to shrink it with entitlement to >4GHz clocks.

It comes down to xtor density, the same reasons a very dense sram layout will require lower clockspeeds, or the GPU logic. If AMD intended for the upper-edge of the clockspeed envelope to be around 3GHz for Llano's cores then the xtor density would have increased, making for a smaller, easier to yield, die at the expense of being a chip that simply can't clock super high.
 
i dont know where i saw it but i believe it was said that amd compensated for the lack of L3 by doubling the L2 on llano. idk, its not really important to me either way. i just want a 3870k regardless unless it cant reach 4ghz oc.
 
i dont know where i saw it but i believe it was said that amd compensated for the lack of L3 by doubling the L2 on llano. idk, its not really important to me either way. i just want a 3870k regardless unless it cant reach 4ghz oc.

Superpi is largely a cache test, and it's showing ~5% faster per clock on Phenom II. I'm sure there are other apps that run faster on the Llano's design. I don't think there can be definitive answers on Llano vs. PhenomII. Some applications will prefer the 45nm design and some will prefer the 32nm design. They are different, but close in performance.

In the end, I think the comparison to the 45nm parts is somewhat moot, since they are discontinuing the 45nm parts in order to gain manufacturing efficiency and increase their volume. Soon the only AMD options will be Llano on FML or Bulldozer on AM3.
 
Last edited:
Back
Top