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
wut - Sunday, August 29, 2010 - link
Yep. I bet AMD is really wondering about that right now.greenguy - Saturday, August 28, 2010 - link
I bought basically what is an i5-750 based on Anand's review here. Or at least, the Xeon version with hyperthreading (needed ECC RAM).From what I can tell, you get about a 20%-30% improvement over the i5-750, with the same power consumption. That's pretty good. Not only that, you get some competent entry level graphics... which would have good open source drivers. That's somewhat exciting, though I wonder whether it would do multiple monitors. Any idea on that Anand?
Maybe I'll just stick to the cheapest Nvidia discrete cards I can buy, a couple G210s do the trick (to get 4 1920x1200 monitors). Unless Intel can make those G210s redundant, it represents just an incremental bump in performance, as the only thing that is of interest is the increase in CPU. One thing that is nice is that Intel is reputed to have the best open source support for their GPU drivers, which makes things really interesting now they are producing stuff that will compete with the entry level discrete market. It could be really good for Linux/BSD people like myself.
The other thing of interest for me is in the low power, low cost, high numbers of SATA connections space, with ECC. I wonder if Bobcat will have something there, as AMD don't seek to arbitrarily differentiate their markets like Intel does with the ECC RAM.
Also not really sure what the big thing is with the motherboards and same CPU. I tend to keep the same computer as a build. By the time you want to upgrade the CPU, there is invariably other stuff that needs upgrading, e.g. USB3, graphics, SATA, RAM, whatever. So you end up wasting the old parts for not that much benefit. Better to just re-purpose the old machine, and when you have enough money, buy the most performant parts that are still good bang for buck. A good example was the i5-750 about 8 months ago or so. So I don't fault Intel for this.
wut - Sunday, August 29, 2010 - link
Intel's going after the mid range market, where most of the money is. We'll have to wait and see how good AMD's Fusion mid range ends up being. Even if it catches up all the way and achieves performance parity so AMD can make more money by raising prices, Intel would have their newest gen on the market first. Fusion had better be really, really, good...Hrel - Sunday, August 29, 2010 - link
I'd say if that 2500K is 215 or less it'd be a fair buy. I'd still wait for the price to drop below 200 cause that's my absolute cap on a CPU. I am a little annoyed that it doesn't have hyper threading though, from a moral standpoint, I mean, from a raw material standpoint how much does adding hyperthreading cost? nothing! yeah, that's what I thought.Hrel - Sunday, August 29, 2010 - link
Those are some impressive integrated graphics. I've thought this for a while now, but we really don't need a card any lower than the HD5670, and maybe the 5650 in discrete graphics. Prefferably just the 5670 though. If mobo makers start setting aside a single DDR3 slot for the integrated GPU to use and dedicated GPU only memory, like a discrete GPU, so the integrated GPU doesn't have to share system RAM we really won't need low end graphics in laptops at all anymore.mino - Sunday, August 29, 2010 - link
Adequate.5450 is the LOWEST END card from 2008, A facelifted 4350.
And 780G, it FINALLY manages to out-pace, in 2011, is the mainstream part of 2008 too.
In 2011 there will be a 10W Ontario with 5450-class GPU on 40nm bulk ...
On the other hand it seems Intel is thaking the GPU side seriously. Finally.
But they are still where ATI/NV were in 2004 ...
LuckyKnight - Sunday, August 29, 2010 - link
I would have liked to have seen a better comparison when it comes to idle power consumption? How much has it improved since moving to a 45nm->32nm GPU?Also, has Intel addressed the Clarkdale issue of not outputting industry standard 24fps? (23.976 hz)
Miggleness - Sunday, August 29, 2010 - link
I was planning on purchasing an i5-760 in 2 weeks but looks like I'll have to settle for 2nd hand, low end parts instead and wait for the i5-2400's release.Great job Intel. I for one no longer have that much interest in overclocking when I have Turbo boost to compensate for that.
Hope we hear about the official pricing soon.
jfelano - Sunday, August 29, 2010 - link
Intel could have hit one out of the park with this one if it worked on existing s1155 motherboards....unfortunately it doesn't and it screws over everyone who bought into s1156 or s1366.....yet again.siberian 3 - Sunday, August 29, 2010 - link
Hi everyone i've read the preview and i am not so impresed by the performance of SB.The IGPis great but it make sense only for the mobile section of the pc not the desktop!!
From the preview i understand that this is not a real fusion product but an evolution off clarkdale and arandale products.
So i will wait for Llano to see what Amd has to offer!!!