IDF Sandybridge overclocking - 4.9Ghz

[ydnas7]

Various reports are saying the following
Sandybridge 'K' part was demonstrated overclocked at the IDF
speed was 4.9Ghz
it was on air
had some voltage increase
was stable, doing a loop of cinebench
outperforms a magny cours in cinebench
part price is 300euro or less
not saying which model number it was.

one of the links http://apcmag.com/Content.aspx?id=6203

 

[VirtualLarry]

At the end of that video, if you scroll back a frame or two, you get a clear shot of what the CPU is, that was supposed to be under NDA according to the audio of that video.

Core i7 2600K
4 Cores/8 Threads @ 3.40 Ghz.

 

[Idontcare]

Ugh, I kinda wish they hadn't done that...we all know how Intel speed-demo's work out.

So basically we just got shown the absolute cherry-picked of cherry-picked sandy's to have left the fab to date and the take home message here is that no one is going to get above 4.9GHz on air in their own rig save for maybe a pony show between DrWho and Fugger.

(sure the statistics are that this is one sample, but we know which part of the distribution tail this one sample was taken from...)

I was really thinking they'd be getting 5.2-5.5GHz on air with fully matured 32nm. Guess that will be water territory then.

 

[IntelUser2000]

They totally fail on the NDA part. Intel spokesman clearly asks them to cover it up but the video is released raw haha. I wouldn't be surprised if they aren't invited to those special events next time.

Sandy Bridge is 25% faster per clock on Cinebench R11.5 compared to Bloomfield. I think that kind of gain might be more common than most think. I also noticed from other sites that the K series are specifically targeted at consumers, including overclockers. That means those chips are the ones that have higher GPU specs and less business oriented specs like Vpro. Just a guess though.

 

[soccerballtux]

Ugh, I kinda wish they hadn't done that...we all know how Intel speed-demo's work out.

So basically we just got shown the absolute cherry-picked of cherry-picked sandy's to have left the fab to date and the take home message here is that no one is going to get above 4.9GHz on air in their own rig save for maybe a pony show between DrWho and Fugger.

(sure the statistics are that this is one sample, but we know which part of the distribution tail this one sample was taken from...)

I was really thinking they'd be getting 5.2-5.5GHz on air with fully matured 32nm. Guess that will be water territory then.

do you remember the article where Intel talked about how the transistor switching time actually became a little bit slower at 32nm, compared with 45nm, but they were able to compensate by increased pressure on the parts of the silicon lattice that made up the channel? I don't think we'll be getting to higher frequencies very soon. Personally, I'm fine with that; we just need more apps to optimize on utilizing multiple cores (Firefox, I'm looking at you).

 

[IntelUser2000]

They didn't say it was slower, they said if they didn't implement that it could have been slower. It's kinda like saying if the Core i7 didn't have L3 cache it would have been slower than Penryn, but there is no Core i7 that lacks L3 cache, so it doesn't matter.

 

[khon]

For a processor that people were panicking about not being able to overclock, I'd say 4.9GHz on air is pretty good.

If I could hit 4.5GHz on my own setup I'd be very pleased.

 

[Idontcare]

They totally fail on the NDA part. Intel spokesman clearly asks them to cover it up but the video is released raw haha. I wouldn't be surprised if they aren't invited to those special events next time.

Sandy Bridge is 25% faster per clock on Cinebench R11.5 compared to Bloomfield. I think that kind of gain might be more common than most think. I also noticed from other sites that the K series are specifically targeted at consumers, including overclockers. That means those chips are the ones that have higher GPU specs and less business oriented specs like Vpro. Just a guess though.

25% IPC boost is crazy amazing...especially when that comes on top of Bloomfields already impressive IPC.

You think this boost is basically coming from the addition of that L0 uop $?

Or are we looking at a special compile of Cinebench that takes advantage of some of the new ISA extensions?

do you remember the article where Intel talked about how the transistor switching time actually became a little bit slower at 32nm, compared with 45nm, but they were able to compensate by increased pressure on the parts of the silicon lattice that made up the channel? I don't think we'll be getting to higher frequencies very soon. Personally, I'm fine with that; we just need more apps to optimize on utilizing multiple cores (Firefox, I'm looking at you).

http://en.wikipedia.org/wiki/Strain_engineering

It is actually yet another one of those major "pros" to the lengthy list of performance advantages that are intrinsic to doing a disposable gate (also called replacement gate or "gate last") integration scheme.

Don't get me wrong, my rant above wasn't to say that 4.9GHz on air isn't great, means we'll probably be seeing some 6GHz water-cooled rigs in the forum. Hmmm...6GHZ...

 

[IntelUser2000]

I would say the 3 biggest contributors are cache changes, additional load unit, and the uop cache. Bigger OoO buffers and branch prediction would help too, but maybe not as much. They probably lowered instruction latency in overall, but again a small gain.

 

[Nemesis 1]

God almighty .

NDA is laughable / We have proof see what it says . LOL . watch the the test. It shows 8 threads running . It could only be 1 SB .

 

[Borealis7]

i cant see the video at work but i wonder...what HSF were they using that fits an LGA1155 M/B? and what were the temps @4.9GHz

 

[Nemesis 1]

Beings this is about O/C the K series . I thought I would show this . My daughter gives a cd of this with ever system she sells . Using this hardware. The young man does a great job. I pretty sure the SB k series will work even better than these . With its turbo advances. This is pretty cool stuff.

http://www.youtube.com/watch?v=6s1n3eTl-0g&feature=channel

 

[khon]

Hard to see exactly, but it looks like it scores 9.xx in that benchmark.

That's very nice for a 4C/8T processor, but it's not a 25% IPC improvement over Bloomfield.

 

[Idontcare]

Beings this is about O/C the K series . I thought I would show this . My daughter gives a cd of this with ever system she sells . Using this hardware. The young man does a great job. I pretty sure the SB k series will work even better than these . With its turbo advances. This is pretty cool stuff.

http://www.youtube.com/watch?v=6s1n3eTl-0g&feature=channel

That was cool, I didn't know Asus put out a nice mobo like that.

Does anand do mobo reviews anymore? I actually don't check the frontpage all that often these days.

 

[Flipped Gazelle]

That was cool, I didn't know Asus put out a nice mobo like that.

Does anand do mobo reviews anymore? I actually don't check the frontpage all that often these days.

1 or 2 each month, plus the occasional round-up.

Not always "performance" motherboards, either - sometimes niche products, like ITX.

 

[GrumpyMan]

Does anand do mobo reviews anymore? I actually don't check the frontpage all that often these days.

Lol....I thought mods read the whole complete site to be up to date on everything and flown up to Anand's castle once a month to talk about it. :sneaky:

 

[Nemesis 1]

I think maybe IDCs remark was a dig guys.

 

[GrumpyMan]

I'm sure it was. Just kidding around a bit.

 

[Idontcare]

Lol....I thought mods read the whole complete site to be up to date on everything and flown up to Anand's castle once a month to talk about it. :sneaky:

It's a lair actually, as in underground...

(just teasing ya with the rolly eyes, good post, funny :thumbsup

 

[GrumpyMan]

I'm just glad they let you out of the Video forum to get some fresh air once in awhile...lol. Sorry, didn't mean to hijack the thread.
Someone post something brilliant and informative about Sandybridge o/c please, since I am no help obviously. Thank you.

 

[extra]

they did show the cooler used, it was a big heatpipe one, but not as big as a noctua or something lol. No doubt this was a completely cherry picked sample, but still . . . on air.

This means we'll probably see 4.5ghz on air become common, and maybe 5ghz on water become commonplace. Yes pwease.

 

[Martimus]

they did show the cooler used, it was a big heatpipe one, but not as big as a noctua or something lol. No doubt this was a completely cherry picked sample, but still . . . on air.

This means we'll probably see 4.5ghz on air become common, and maybe 5ghz on water become commonplace. Yes pwease.

What it is telling me is to temper my expectations on the Bulldozer clock speed. If Intel has a better HKMG process than GF, and still isn't seeing much higher clocks than they got from 45nm, I am not sure how the GF process will do.

I believe that the SOI and perhaps better low-k might help mitigate this, as well as the difference in design (BD appears to be designed for high clock speeds). Even so, my expectation that we would see a 4GHz stock SKU from AMD by the end of 2011 is sufficiently tempered.

I am not an expert on this, but if IDC seems to think this is low, then I know to expect even less from GF due to the inherently weaker gate first HKMG implementation.

 

[ilkhan]

Even 4.5Ghz on air for average chips would be impressive to me. Q8200@2.3 -> 2500k@4.5 plus IPC improvements...wow. After core maximizer for better using multi-core, my app (supreme commander) is still limited by clock speeds.

 

[OCGuy]

Wow. Two weeks ago people were saying that you can't even overclock SB, and now I may possibly have a 4.5ghz air-cooled SB???

I don't think I will be able to hand them my money fast enough at launch.

 

[Idontcare]

Wow. Two weeks ago people were saying that you can't even overclock SB, and now I may possibly have a 4.5ghz air-cooled SB???

I don't think I will be able to hand them my money fast enough at launch.

If you like, you can alleviate some of that $-flux bottleneck by getting a head-start on the transfer and send it to me, "c/o CASH"

What it is telling me is to temper my expectations on the Bulldozer clock speed. If Intel has a better HKMG process than GF, and still isn't seeing much higher clocks than they got from 45nm, I am not sure how the GF process will do.

I believe that the SOI and perhaps better low-k might help mitigate this, as well as the difference in design (BD appears to be designed for high clock speeds). Even so, my expectation that we would see a 4GHz stock SKU from AMD by the end of 2011 is sufficiently tempered.

I am not an expert on this, but if IDC seems to think this is low, then I know to expect even less from GF due to the inherently weaker gate first HKMG implementation.

It's low for my lofty expectations, but not silly low. If all they could muster was 4GHz OC demo then that would have been silly low. I'd have been impressed with 5.3-5.4GHz on air for cherry-picked sample. Its not that I'm dissapointed with 4.9GHz, I'm just not impressed.

Regarding BD and clockspeeds...we can't make those comparisons here, the BD architecture could be clockspeed friendly (like P4 and Power7) and we could see 5GHz stock locks (just an example, but IBM is doing that with 45nm) or it could be like Core 2 Duo and be really IPC strong but not so crazy about scaling to high clocks.

Process tech can become the clockspeed limiter as can architecture, but both are different for AMD with BD and glofo 32nm so we can't really say anything about our expectations of BD's clockspeeds.