Intel's Sandy Bridge Architecture Exposed
by Anand Lal Shimpi on September 14, 2010 4:10 AM EST- Posted in
- CPUs
- Intel
- Sandy Bridge
Multiplier-only Overclocking
There's no new information on Sandy Bridge overclocking at this point (although it's looking increasingly likely that there will be a reasonably priced K-series SKU for those users who want the flexibility to overclock without spending $1000). I've included the overclocking text and roadmap from our Sandy Bridge Preview below if you're interested in seeing what Intel has planned.
It wasn’t until the Pentium II that Intel started shipping multiplier locked CPUs. Before then you could set the multiplier on your CPU to anything that was supported by the line, and if you had a good chip and good enough cooling you just overclocked your processor. Intel’s policies changed once remarking, the process of relabeling and reselling a lower spec CPU as a higher one, started to take off.
While multipliers were locked, Intel left FSB overclocking open. That would be an end user or system integrator decision and not something that could be done when selling an individual CPU. However, ever since before the Pentium III Intel had aspirations of shipping fully locked CPUs. The power of the enthusiast community generally kept Intel from exploring such avenues, but we live in different times today.
Two things have changed Intel’s feelings on the topic. First and foremost is the advent of Turbo Boost. So long as Intel doesn’t artificially limit turbo modes, we now have the ability to run CPUs at whatever clock speed they can run at without exceeding thermal or current limits. We saw the first really exciting Turbo with Lynnfield, and Sandy Bridge is going to expand on that as well. On the flip side, Intel has used Turbo as a marketing differentiator between parts so there’s still a need to overclock.
The second major change within Intel is the willingness to directly address the enthusiast community with unlocked K-series SKUs. We saw this recently with the Core i7 875K and Core i5 655K parts that ship fully unlocked for the overclocking community.
The K-series SKUs, these will be more important with Sandy Bridge
With Sandy Bridge, Intel integrated the clock generator, usually present on the motherboard, onto the 6-series chipset die. While BCLK is adjustable on current Core iX processors, with Sandy Bridge it’s mostly locked at 100MHz. There will be some wiggle room as far as I can tell, but it’s not going to be much. Overclocking, as we know it, is dead.
Well, not exactly.
Intel makes three concessions.
First and foremost we have the K-series parts. These will be fully unlocked, supporting multipliers up to 57x. Sandy Bridge should have more attractive K SKUs than what we’ve seen to date. The Core i7 2600 and 2500 will both be available as a K-edition. The former should be priced around $562 and the latter at $205 if we go off of current pricing.
Secondly, some regular Sandy Bridge processors will have partially unlocked multipliers. The idea is that you take your highest turbo multiplier, add a few more bins on top of that, and that’ll be your maximum multiplier. It gives some overclocking headroom, but not limitless. Intel is still working out the details for how far you can go with these partially unlocked parts, but I’ve chimed in with my opinion and hopefully we’ll see something reasonable come from the company. I am hopeful that these partially unlocked parts will have enough multipliers available to make for decent overclocks.
Finally, if you focus on multiplier-only overclocking you lose the ability to increase memory bandwidth as you increase CPU clock speed. The faster your CPU, the more data it needs and thus the faster your memory subsystem needs to be in order to scale well. As a result, on P67 motherboards you’ll be able to adjust your memory ratios to support up to DDR3-2133.
Personally, I’d love nothing more than for everything to ship unlocked. The realities of Intel’s business apparently prevent that, so we’re left with something that could either be a non-issue or just horrible.
If the K-series parts are priced appropriately, which at first indication it seems they will be, then this will be a non-issue for a portion of the enthusiast market. You’ll pay the same amount for your Core i7 2500K as you would for a Core i5 750 and you’ll have the same overclocking potential.
Regardless of how they’re priced, what this is sure to hurt is the ability to buy a low end part like the Core i3 530 and overclock the crap out of it. What Intel decides to do with the available multiplier headroom on parts further down the stack is unknown at this point. If Intel wanted to, it could pick exciting parts at lower price points, give them a few more bins of overclocking headroom and compete in a more targeted way with AMD offerings at similar price points. A benevolent Intel would allow enough headroom as the parts can reliably hit with air cooling.
The potential for this to all go very wrong is there. I’m going to reserve final judgment until I get a better idea for what the Sandy Bridge family is going to look like.
Facebook
Twitter
RSS
62 Comments
View All Comments
JoJoman88 - Wednesday, September 15, 2010 - link
To me the problem is that instead of me overclocking without reguard to TDP, now Intel will do the overclocking for me, but it will be within the TDP that Intel thinks is best. Will this not just kill the after-market cooler makers with an almost locked TDP,and to some degree high end memory maker with a locked BCLK.This will change how overclocking is done from now on unless AMD keeps things as they are and forces Intel from going down this road.
gvaley - Wednesday, September 15, 2010 - link
It's true that the CPU will turboboost within the CPU's TDP, but exactly how much it will turboboost (how much bins it will gain) will depend on how well the CPU is being chilled. So having a better (read: aftermarket) cooler will allow you to take the best of your CPU's turboboost.shodanshok - Wednesday, September 15, 2010 - link
Hi Anand,while I expect the ring bus to provide great performances, I doubt that it don't impact die size and power consumption in respect to the Nehalem/Westmere L3 organization.
Let me explain...
From my internal test, the Nehalem/Westmere L3 cache seems accessed by a four 64 bit channels (one per core). At 3 Ghz L3 cache, it translate in a maximum of 24 GB/s per core, or 96 GB/s for 4 cores. This cache organization seems confirmed by the tests at Techreport (on a i975X, SANDRA's L3 cumulative bandwidth is at about 60 GB/s: http://www.techreport.com/articles.x/18581/5) and Xbitlabs (EVEREST single-core L3 bandwidth of about 20 GB/s: http://www.xbitlabs.com/articles/cpu/display/intel...
So, on Nehalem/Westmere I do not expect 4 x 256 wires, but only 4 x 64 wires (more or less).
Now, lets examine SB...
We have 4 x 256 bit bus (4 indipendent rings) that runs around the L3 cache, for a total of 1024 wires. So, we have a lot of wires, that needs to be powered. These wires, in turn, need additional die space, and to me seems the main reason why most models will have "only" a 6 MB L3.
What do you think about? It is possible to ask Intel something about the Nehalem L3 cache organization and/or about the decision to equip most SB models with 6 MB of L3 cache?
Thanks.
Casper42 - Wednesday, September 15, 2010 - link
Knowing what you do about Nehalem EX and SNB on socket H2, any speculation on what we can expect from the Socket B2/R chips when they finally arrive sometime next year?I am mainly thinking of Northbridge/QPI and PCIe Lanes as compared to DMI used on the Mainstream parts discussed in this article.
I waited and waited for Westmere Core i7 to become "cheap" and thought the 970 was going to be my chip of choice @ $550. When they released it at $900 (you could already find 980Xs for less) it pretty much killed my plans to upgrade.
So now I am basically debating on do I build a high end H2 or wait for the enthusiast version to arrive instead?
My understanding from seeing the server roadmap is there will be Socket B2 and Socket R with the differences between them mainly consisting of memory channels and # of PCIe Lanes. I have also read that both will support PCIe 3.0 whereas H2 will continue to use 2.0.
Add all these changes up and I am also hopeful we will see USB3 on the Enthusiast platform as well since it will have an additional 3-6 months to mature.
So any ideas/insight you have here would be awesome.
linkages - Thursday, September 16, 2010 - link
With the price of LCDs dropping, I am noticing that more and more consumers have more than one display for their mainstream machines. Has Intel said anything about how many displays the onboard graphics will be able to push? Have they said anything about what tech they are going to use ie. display port, HDMI, DVI-D, something else?I can see myself getting a new SB machine sometime in Q1 2011 but I run at least 2 monitors at all times ( need the real estate for the type of work that I do ). I don't play many games but having the video decode/encode is important to me since I do tend to do some videoconferencing now a days.
The last thing I would like to know is if Intel is going to do the right thing with the drivers for their graphics. Will we humble linux users finally have a graphics driver that does not suck. Will Intel finally open source the driver so that the community can keep it updated and optimize it for X?
chukked - Thursday, September 16, 2010 - link
Hi Anand,thanks for the review, you addressed everything but left virtualization :(
which processors support vt-x and vt-d ?
iwodo - Friday, September 17, 2010 - link
x264, the best h.264 encoder there is, produce better quality video and similar speed when using "ultrafast" setting. And with 2 / 4 Core we could even transcode 2 - 4 video at the same time.The hardware encoder inside SandyBridge is not that speedy. While i could scarifies quality for speed. But Power VR's VRE Core manage 1000fps +, @ 400fps the encoder is like a waste of die space.
Intel could have further tuned the x264 for Sandy Bridge for speed and just release it with their drivers. If the hardware encoder aren't giving many times the increase in speed, then what is the point? They may as well have added extra 6 EU for GPU inside.
A Link to someone's blog posting some figures.
http://lee.hdgreetings.com/2010/09/intel-cpu-vs-nv...
Wolfpup - Wednesday, September 29, 2010 - link
Pretty disappointing. I'm sure AMD's glad though!10-30% improvement-obviously that's great, but not as big as their previous tocks if I'm remembering right, and not much different from what "ticks" like Penryn did...I know Penryn was like a 10% boost minimum over Conroe...
I'm guessing it's because they're wasting effort and die area on a worthless GPU. I *HOPE* no one on this site, no power users are going to be using that thing. (Well, okay, for a tiny notebook or something maybe...)
JumpingJack - Wednesday, September 29, 2010 - link
I don't believe you are remembering correctly.Conroe was the first tock, and certainly, it was a major leap over the P4 line. But it's bloodline was actually derived from P6, which was carried through from Banias, Dothan, then Yohna. The improvement over Yohna was in the 10-20% IPC range.
Then came Penryn the tick, which was on average only 5%, http://www.anandtech.com/show/2306/3
Then came the tock, which was Nehalem. In single threaded performance, it was roughly another 5-10% over Penryn, but in multithreaded -- again, clock for clock, it had leaps of performance, around 20-40% again. http://www.anandtech.com/show/2658/20
The tick of Nehalem was Westmere, now Westmere did not launch a quad core part so it is hard to find a clock for clock, but in single threaded performance -- roughly the same as Nehalem, factoring out any turbo advantages...
Now SB, a tick, with another 10-30% across the board both single and multithreaded, depending on workload.
Of course, the GPU is not so worthless, it is indeed challenging low end GPUs -- no doubt Llano will offer up strong GPU performance, but for the majority of the market SB is perfectly fine.
Indeed, AMD is not too happy here, at least I would suspect. On the CPU side, Intel will crush anything AMD has in the same market segement where SB resides... GPU, surely AMD will crush SB. On the CPU front, AMD is already 20-40% behind Nehalem clock for clock, core for core, SB just extends that another 10-30%.
gundersausage - Tuesday, October 26, 2010 - link
i7-950 vs i7-2500K... So which will be faster and a better gaming chip? anyone?