AMD - Bulldozer Chitchat

[shaithis]

Looks like the same specs as the 3280 HE wiki has had listed as available from Q4 2011.......also 65w

[scaryjim]

No wonder given its paltry clockspeed.

If you have highly parallelised simple tasks to run then a low speed high core-count CPU is exactly what you want. Great for virtualisation and a variety of other server-side uses. it fits very nicely with AMDs server roadmap, alongside their purchase of SeaMicro. I'd expect to see 45W and 35W variants cropping up before too long, down in the 1.8GHz kind of area...

[Terbinator]

If you have highly parallelised simple tasks to run then a low speed high core-count CPU is exactly what you want. Great for virtualisation and a variety of other server-side uses. it fits very nicely with AMDs server roadmap, alongside their purchase of SeaMicro. I'd expect to see 45W and 35W variants cropping up before too long, down in the 1.8GHz kind of area...

Of course, the point i was trying to get at is that its not really braking any boundaries/come as a surprise given the spec and resulting TDP.

Is it?

[scaryjim]

Hmmm, first 65W native oct-core CPU? If you accept bulldozer as native oct-core, anyway

Hmm, my wiki-fu tells me it dosn't exist? I can find bulldozer-based 4200 and 6200 series CPUs on sockets C32 and G34, including 65W & 35W TDP oct-core processors....

Ah, it's in "future" AMD processors, along with a couple of processors that are incorrectly identified as 45W 8-core, which are actually 45W quad cores. So I guess it's notable as a low power native oct, although AMD have already outdone it on "proper" server platforms.

I dunno - I think in terms of pushing a full multi-core CPU architecture into a very low power threshold it's pretty impressive, but perhaps not completely groundbreaking. Depends on the price you pay for it I guess: could be a very cheap way to get heavily-threaded performance: it should just drop into any AM3+ board, which can be had for < £40 still (AFAIK)...

[CAT-THE-FIFTH]

Someone is going to get into trouble:

http://www.overclock.net/t/1225201/8...#post_16638592

Some comments on the markets for the CPU:

http://www.overclock.net/t/1225201/8...#post_16641753

[badass]

If you have highly parallelised simple tasks to run then a low speed high core-count CPU is exactly what you want. Great for virtualisation and a variety of other server-side uses. it fits very nicely with AMDs server roadmap, alongside their purchase of SeaMicro. I'd expect to see 45W and 35W variants cropping up before too long, down in the 1.8GHz kind of area...

It would be if the architecture wasn't so incredibly slow for virtualisation compared to Xeons http://www.anandtech.com/show/5553/t...-for-servers/6

[scaryjim]

It would be if the architecture wasn't so incredibly slow for virtualisation compared to Xeons

That'd be a suitable conclusion if that review actually had any price comparable processors in it It has neither an E5-2630 nor an Opteron 6274, which is the midrange server point where the two processor ranges coincide, price-wise. They've attempted to produce those CPUs by disabling and downclocking higher-spec CPUs, but that's never going to give an exact representation.

And looking at the linked chart, the Opterons are hardly incredibly slow - they turn in 93.5% of the (mocked up) roughly price equivalent Xeon's throughput performance. They're slower, yes, but incredibly? 7% slower in benchmarks might not even be noticeable during typical usage (I note the response times detailed in the power testing section of the review are far shorter - and far closer to each other - than those in the benchmarking section, indicating that the benchmarking section does *not* represent a typical usage pattern). It seems to me you're being a little unfair in your assessment...

[CAT-THE-FIFTH]

The virtual E5-2630 which was created has 33% more L3 cache than the actual E5-2630 too(with a 100MHZ clockspeed decrease).

Also the 2.5GHZ Opteron 6280 has been released which replaces the 2.3GHZ Opteron 6276 AFAIK.

[CAT-THE-FIFTH]

Some details about Trinity from AMD:

http://blogs.amd.com/fusion/2012/03/...9D-generation/

A Trinity laptop listing has appeared in Russia:

http://tpshop.ru/position/106124/

[scaryjim]

Interesting. It would appear that the Trinity performance testing by AMD labs used DDR3-1600, whereas the Llano testing used DDR3-1333. So the 56% graphics performance improvement will in part be down to the increased memory bandwidth available ... sneaky AMD...

[CAT-THE-FIFTH]

Funnily enough most Llano laptops seem to have 1333MHZ RAM!! I only realised this when looking for a new laptop for a friend. What a load of fail.

[scaryjim]

otoh if the Trinity reference platform uses DDR3-1600 then presumably that will encourage more laptop manufacturers to go for it with their retail designs ... one would hope, at least!

[Biscuit]

I upgraded the ram in my netbook the only day... its only a E-450 based thinkpad but i got corsair 1600Mhz Vengence anyway.... overkill much

[scaryjim]

overkill much

probably, for an 80shader iGPU, on a chip with a single channel memory controller that only officially support DDR3-1333! Is it actually running at 1600MHz or is it stuck at 1333 due to IMC restrictions?

[Biscuit]

Not even sure its running at 1333, il have to check when im back home. I did notice the timings it was running were quite low though, like cas6.
It was like £5 more and i figured it worth it if RAM Is set to get way more expensive in the coming months, if not me for a friend or family.

[watercooled]

if RAM Is set to get way more expensive in the coming months

It is?