The Sandy Bridge Preview
by Anand Lal Shimpi on August 27, 2010 2:38 PM ESTA New Architecture
This is a first. Usually when we go into these performance previews we’re aware of the architecture we’re reviewing, all we’re missing are the intimate details of how well it performs. This was the case for Conroe, Nehalem and Lynnfield (we sat Westmere out until final hardware was ready). Sandy Bridge, is a different story entirely.
Here’s what we do know.
Sandy Bridge is a 32nm CPU with an on-die GPU. While Clarkdale/Arrandale have a 45nm GPU on package, Sandy Bridge moves the GPU transistors on die. Not only is the GPU on die but it shares the L3 cache of the CPU.
There are two different GPU configurations, referred to internally as 1 core or 2 cores. A single GPU core in this case refers to 6 EUs, Intel’s graphics processor equivalent (NVIDIA would call them CUDA cores). Sandy Bridge will be offered in configurations with 6 or 12 EUs.
While the numbers may not sound like much, the Sandy Bridge GPU is significantly redesigned compared to what’s out currently. Intel already announced a ~2x performance improvement compared to Clarkdale/Arrandale, and I can say that after testing Sandy Bridge Intel has been able to achieve at least that.
Both the CPU and GPU on SB will be able to turbo independently of one another. If you’re playing a game that uses more GPU than CPU, the CPU may run at stock speed (or lower) and the GPU can use the additional thermal headroom to clock up. The same applies in reverse if you’re running something computationally intensive.
On the CPU side little is known about the execution pipeline. Sandy Bridge enables support for AVX instructions, just like Bulldozer. The CPU will also have dedicated hardware video transcoding hardware to fend off advances by GPUs in the transcoding space.
Caches remain mostly unchanged. The L1 cache is still 64KB (32KB instruction + 32KB data) and the L2 is still a low latency 256KB. I measured both as still 4 and 10 cycles respectively. The L3 cache has changed however.
Only the Core i7 2600 has an 8MB L3 cache, the 2400, 2500 and 2600 have a 6MB L3 and the 2100 has a 3MB L3. The L3 size should matter more with Sandy Bridge due to the fact that it’s shared by the GPU in those cases where the integrated graphics is active. I am a bit puzzled why Intel strayed from the steadfast 2MB L3 per core Nehalem’s lead architect wanted to commit to. I guess I’ll find out more from him at IDF :)
The other change appears to either be L3 cache latency or prefetcher aggressiveness, or both. Although most third party tools don’t accurately measure L3 latency they can usually give you a rough idea of latency changes between similar architectures. In this case I turned to cachemem which reported Sandy Bridge’s L3 latency as 26 cycles, down from ~35 in Lynnfield (Lynnfield’s actual L3 latency is 42 clocks).
As I mentioned before, I’m not sure whether this is the result of a lower latency L3 cache or more aggressive prefetchers, or both. I had limited time with the system and was unfortunately unable to do much more.
And that’s about it. I can fit everything I know about Sandy Bridge onto a single page and even then it’s not telling us much. We’ll certainly find out more at IDF next month. What I will say is this: Sandy Bridge is not a minor update. As you’ll soon see, the performance improvements the CPU will offer across the board will make most anyone want to upgrade.
Facebook
Twitter
RSS
200 Comments
View All Comments
AndreC - Saturday, August 28, 2010 - link
Hi there.. I`m currently a freelance 3D generalist.. and I was going to upgrade my old Core 2 Quad QX6700 with a Core i7 980X. But now i`m not that confident. Sandy bridge looks amazing, I was sad seeing the new socket for sandy bridge, it does not compell me to buy a new motherboard now... does anyone know if the 1366 socket will stick with the nex gen High end market? I dont want to shoot myself in the foot here.AndreC - Saturday, August 28, 2010 - link
Sorry dind`t read the last frase...Great Review btw. cheers
sdsdv10 - Saturday, August 28, 2010 - link
As noted in the Intel roadmap in the article, for at least part of 2011 they will be sticking with 1366 for the release of the Core i7 990X (to replace the 980X). However, after that the Intel performance platform will switch over to socket LGA-2011. Here is a quote from the articlea (page 3)."Original Nehalem and Gulftown owners have their own socket replacement to look forward to. In the second half of 2011 Intel will replace LGA-1366 with LGA-2011. LGA-2011 adds support for four DDR3 memory channels and the first 6+ core Sandy Bridge processors."
AndreC - Saturday, August 28, 2010 - link
Yeah.. as I said "Sorry dind`t read the last frase..." but thx anyway..It`s a shame to be always changing sockets, but probabily a necessity to evolve the technology.
Kaihekoa - Saturday, August 28, 2010 - link
Having the first chips target the mainstream market is a very smart move by Intel because that's where AMD makes it's money. I'm honestly not impressed by the performance numbers, but I am impressed by the overall performance, power consumption, and pricepoints for these next gen CPUs. What I'm really looking forward to is the performance segment of Sandy Bridge.mino - Saturday, August 28, 2010 - link
Intel's mainstream is not where AMD's is.Especially in 2011.
Ontario:
. . . CPU - above Atom, under everything else
. . . GPU - 5450/Sandy class
Lliano:
. . . CPU - 2C Sandy class
. . . GPU - 5650 class (at least 3x Sandy)
Bulldozer Desktop(8C):
. . . CPU - 4C Sandy Class
. . . GPU - discrete 5750+ class
So basically AMD's platform in the Intel's "mainstream" $200+ class will be a Bulldozer with discrete GPU. Aka AMD's high end stuff.
silverblue - Saturday, August 28, 2010 - link
Not sure I agree with that. From AMD's own figures, Bulldozer is significantly faster than STARS. It would be more realistic to expect Bulldozer to perform closely to Sandy Bridge, however we really need more benchmarks before we get a true idea. Bulldozer looks great on paper, but that's virtually all we have so far.In any case, you compared Bulldozer to "4C Sandy Class", which would be an 8-thread Sandy Bridge, and thus - at least relatively - high end. And I'm not getting into the core/module argument again... ;)
mino - Sunday, August 29, 2010 - link
What I wanted to point out is that Intel sees the 4C Sandy as a "mainstream" part.Reason being they are moving HUGE amounts (compared to AMD) of $150-$250 parts.
On the other hand, AMD sees the mainstream at $100-$200 and that is a Llano market.
For AMD, Zambezi is high-end that justifies discrete GPU.
And Yes, Bulldozer 8C should compare with 4C Sandy favorably, (it would mostly go to pricing).
tatertot - Sunday, August 29, 2010 - link
BD 8C is going to be up against 8C and 6C Sandy on LGA-2011 in the client space.mino - Monday, August 30, 2010 - link
am sure AMD WOULD like it that way. But no. Not really.8C Bulldozer versus 6C Sandy might actually be competitive.
However 32nm SOI is a new process so we might as well forget about 4GHz parts for now.
Also, Sandy 6C is Q4 part and 8C is most probably 2012 part.