• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

32nm when?

cannondale06

[H]F Junkie
Joined
Nov 27, 2007
Messages
16,180
sorry if this is obvious but do any roadmaps show a time frame for i7 going to 32nm? the only thing I have noticed is a 6 core cpu that will come out 1st half 2010 that will be the i7 successor. surely there is a shrink to 32nm in between.
 
clarkdale/arrandale (dual desktop/mobile) Q4/2009
Theres been a flood of info lately.
There may or may not be a 32nm quad, there may or may not be duals without iGPU. We just dont know yet.
The 6-core you mentioned is gulftown, and will replace bloomfield as the s1366 offering (may or may not have a 32nm quad for s1366).
 
I'm not optimistic it will make this year though.
Well, the process will definitely be ready (Intel publicly demo'd their first 32nm chip recently, and it's pretty much ready to go). Whether or not they actually release their 32nm chips is a different story though.
 
Yeah that was my focus, given the recession, theres no point in pushing the boat out.
 
Another larger shovel with which to bury AMD with....

I'm sorry... but you set up that sentence so as NOT to end with a preposition... yet you did it anyway!... I need to find an even larger shovel with which I can hit you on the head with.

I'm hoping I can hold onto my e8400 until overclockable 32nm steppings are released. :)
 
Yeah that was my focus, given the recession, theres no point in pushing the boat out.

Actually there is a bigger push to do that with the recession. Intel has recently focused on getting more fabs up to speed on 32nm and closing up shop on larger processing nodes that are no longer as competitive to make in terms of dies/profit per wafer. The quicker they get 32nm ramped the quicker they can see the increased margins from it. R+D into 32nm is already a sunk cost at this point, might as well go ahead and use it now since it's working.

Intel has always had a headstart in this area compared to other fabs, but this time it might be an especially key advantage given the state of the economy. AMD will have to use more materials and get less cores per wafer out of those materials and also try to compete with low prices and margins on those cores. If AMD's transition to 32nm is not going more smoothly than the 45nm ramp then 2H2010 and into 2011 may be a very bad time for them.
 
Another larger shovel with which to bury AMD with....

I sure hope not. With no competition prices will certainly go up and performance inprovements will slow down. It would be good for the shareholders though (increased profits).
 
Actually there is a bigger push to do that with the recession. Intel has recently focused on getting more fabs up to speed on 32nm and closing up shop on larger processing nodes that are no longer as competitive to make in terms of dies/profit per wafer. The quicker they get 32nm ramped the quicker they can see the increased margins from it. R+D into 32nm is already a sunk cost at this point, might as well go ahead and use it now since it's working.

Intel has always had a headstart in this area compared to other fabs, but this time it might be an especially key advantage given the state of the economy. AMD will have to use more materials and get less cores per wafer out of those materials and also try to compete with low prices and margins on those cores. If AMD's transition to 32nm is not going more smoothly than the 45nm ramp then 2H2010 and into 2011 may be a very bad time for them.

Agreed. Moving to a smaller process may sound like big dollars at first, but increases profits as more dies can fit per wafer, given the smaller thermal envelope, smaller heatsinks too for retail pkgs!
 
Actually there is a bigger push to do that with the recession. Intel has recently focused on getting more fabs up to speed on 32nm and closing up shop on larger processing nodes that are no longer as competitive to make in terms of dies/profit per wafer. The quicker they get 32nm ramped the quicker they can see the increased margins from it. R+D into 32nm is already a sunk cost at this point, might as well go ahead and use it now since it's working.

Intel has always had a headstart in this area compared to other fabs, but this time it might be an especially key advantage given the state of the economy. AMD will have to use more materials and get less cores per wafer out of those materials and also try to compete with low prices and margins on those cores. If AMD's transition to 32nm is not going more smoothly than the 45nm ramp then 2H2010 and into 2011 may be a very bad time for them.

They still have to sell them though!
 
They still have to sell them though!

I think that's why Intel chose to go with the Clark/Arrandales to start 32nm out with for a quarter or two. That's a completely new product that will appeal directly to the low-cost market. They will be cheap to produce and cheap to buy. Should sell like hotcakes to laptop makers and OEMs with a (hopefully) stronger economic situation near the end of the year. I think Intel can eek out a slightly higher margin if they sell their dual-core chips with an IGP onboard than if they were to package both separately.

If they had decided to launch 32nm with the i7 follow up which itself has sluggish sales that kind of slow adoption rate would probably have continued. The 'dales are a nice change of pace and this will extend the timeframe that the 45nm i7 is on top of the market so they can sell through any surplus inventory they might have before Gulftown takes the throne.
 
If the economy is improving, I agree that is a decent strategy.
I'm a pretty optimistic chappy but I am also quite realistic (keeps me in check :))
With the best optimism, I still dont have much hope that things will be better by the year end.
I believe everyone will have tightened their belts more by then sadly.
Its better not to continue on this line of discussion as it will derail the thead and its also pretty depressing.
my 2p.
 
They'll probably have the 32nm ready before we can even imagine... Just secretly holding it until AMD releases something...

I would. :)
 
Considering the only people they'd trump by releasing it sooner than necessary, would be themselves.

Why shoot yourself in the foot. Give it time to mature in testing scenarios and get the Board builders time in house to thoroughly test the new goodies.

That's my feelings on it at least. I mean its pretty well been shown in every aspect, that nothing compares to the i7 at this point, and I doubt AMD can pull enough rabbits outta their rectum to produce such a beast in the near future.

Do I wish they could, yes, do I see it as feasible, not at this point.
 
They'll probably have the 32nm ready before we can even imagine... Just secretly holding it until AMD releases something...

I would. :)

Well, they already do have working 32 nm CPUs, which they have demonstrated to the press.
They have also said that ramping up yields is going better than expected, faster than both 65 nm and 45 nm did. So it looks like they're ahead of schedule.
But as said, they can't release it too soon, because they'd shoot themselves in the foot if they get stuck with 45 nm stock. They are already binning their CPUs very conservatively, and they've delayed the introduction of mainstream Nehalem parts.
 
blog_amd_roadmap.jpg


blog_westmere_roadmap.jpg
 
Gulftown on x58 chipset in 2010+

woohoo! Glad I chose i7. Core Duo/Quad owners are screwed and will need to update mobo + cpu + mem to go 32nm. i7 owners will only need to buy a new cpu to stay on top.

AMD is dying and will need to come up with a surprise. It's roadmap isn't looking purty compared to Intel's. :(
 
I thought I read somewhere that AMD is working on an 8 core CPU. I guess not if that map is accurate.
 
I thought I read somewhere that AMD is working on an 8 core CPU. I guess not if that map is accurate.
They just demoed a six core 45nm CPU. It'll be a 4/8way socket part (opteron 8xxx series)
Intel has an 8-core (beckton, 45nm, s1567) coming out, well, this year; and a 6-core coming out H1-2010 (gulftown,32nm, s1366)

What we're going to see is clarkdale (32nm) laughing at AMDs dual cores, lynnfield laughing at deneb/tri-cores, and beckton/gainestown laughing at the 2/4/8-way market. Then we get to 32nm shrinks in 2010 and they just laugh harder.

Id say, grab an i7 now for a gulftown 32nm upgrade, and then an arrandale based laptop come early Q1-2010. The arrandale will run aero just fine and have plenty of CPU for most tasks, let the gulftown run the render/edit farm and gaming. Clarkdale or 32nm Atom will make a perfect home server chip.
 
when did all the processor lines become an "X-dale"
duals are -dale
quads are -field
sexi (6) are -town
octo (8) are -ton

from what we've seen. -ton may or may not hold up, we've only seen beckton. -town is also just one (gulftown) but seems more realistic than -ton.
 
Gulftown on x58 chipset in 2010+

woohoo! Glad I chose i7. Core Duo/Quad owners are screwed and will need to update mobo + cpu + mem to go 32nm. i7 owners will only need to buy a new cpu to stay on top.

AMD is dying and will need to come up with a surprise. It's roadmap isn't looking purty compared to Intel's. :(
I see that as a good thing... It'll have time to mature as well as be cheaper. I am probably going to skip i7 all together. My OC'd Q9300 and E8500 seem to work well and are dirt cheap to build a system around right now. I generally stay one step behind anyway... I get more for my money that way.
 
woohoo! Glad I chose i7. Core Duo/Quad owners are screwed and will need to update mobo + cpu + mem to go 32nm. i7 owners will only need to buy a new cpu to stay on top.


Some of the screwed Duo/Quad owners have had their setups a lot longer than you have had your i7 (getting a lot of use from them). And if they do go buy a new motherboard/cpu/mem, they will likely get a lot more choices with better prices and features/performance that you don't have (USB 3 for example)
 
Better prices? 32nm is going to be high-end, even more-so. Production costs will decrease but I doubt retail prices will lower for something that is labeled high-end.

What I'm saying is: it will be a lot cheaper for i7 owners to upgrade to 32nm. There's no arguing that. If you bought a Q9550 (and up) system recently, you didn't quite make the right decision..
 
What I'm saying is: it will be a lot cheaper for i7 owners to upgrade to 32nm. There's no arguing that.

Of course it will

you already got reamed on the motherboard and DDR3 :p


but yes, cheaper prices also. 32nm non "dale" cpus will be high end at first but so are current i7 cpus. The price of the cpus will be the same or lower than what i7 started at. Take that and cheaper/better/wider selection of motherboards and ram and you get "better prices"
 
and from what I have seen of how Intel works, there is a LOT higher chance that six core 32nm cpus will launch for cheaper prices instead of "the same" prices at quad core i7. It is a volume business. Your i7 was bought when it was just introduced. i7 is about one percent of Intel's sales going to the consumer market. They don't expect it to be that way in 2010. The way you increase sales a lot is by lowering prices.



If you bought a Q9550 (and up) system recently, you didn't quite make the right decision..

Possibly, but you would have to ask somebody who bought that. I bought my E7200/motherboard and ram for less than you likely paid for your ram alone.
 
Better prices? 32nm is going to be high-end, even more-so. Production costs will decrease but I doubt retail prices will lower for something that is labeled high-end.

Wasn't the E8400 'high-end' when it came out? It was MSRP'd at a very cheap price (although retailers reamed the customers for awhile).


What I'm saying is: it will be a lot cheaper for i7 owners to upgrade to 32nm. There's no arguing that. If you bought a Q9550 (and up) system recently, you didn't quite make the right decision..

As another already said, you've already been bent over for the RAM and MOBO. My Q9300 setup that I recently put together ran about $350 tops for mobo, ram, cpu and Ninja2. It'll do what I need for a year and then I could look at getting i7, RAM, and mobo for far cheaper than what you'd pay now. Being behind the curve can really help your wallet. My e-peen is big enough :)
 
you already got reamed on the motherboard and DDR3 :p
OK, Im just plain tired of this shit already.
Yes, DDR3 costs more than DDR2.
Live with it.
Its cheaper than DDR2 was 2 years ago, and under $20/GB already. We wouldn't killed for that price level a few years ago. So just stop with this BS. Nothing says you need DDR3-2000 for an i7, in fact with triple channel getting anything above DDR3-1333 is just plain stupid. So just stop.
this isn't directed at you, the quote is just a convenient example.
 
q6700 is enough for me right now, but if I happen to have a job when this comes out and maybe even back to 5 day work weeks I may bite just for the novelty/e-peen.
mobos should be more reasonable and ddr3 even cheaper.
 
resurrection :D


so is P55 dead end or will there be a transition to 32nm on those platforms to? also arent there supposed to be some different chipsets for coming for 1156?
 
I don't see no reason why 1156 can't/won't get 32nm in the future.
They're already getting Clarkdale 32nms in Q4 2009 (was it?)

I'm assuming you're awaiting the 32nm quadsbut that won't be for a while seeing 1156 just launched and 32nmGulftown is expected before that.
 
I don't see no reason why 1156 can't/won't get 32nm in the future.
They're already getting Clarkdale 32nms in Q4 2009 (was it?)

I'm assuming you're awaiting the 32nm quadsbut that won't be for a while seeing 1156 just launched and 32nmGulftown is expected before that.
yeah it might be a while before the 32nm quads get here but I was just trying see all my possible options. I dont really need to upgrade my cpu now as I just dont see that being cost effective since besides the cpu I would be needing mb, memory and new cpu cooler too. I probably will do a gpu upgrade soon and then just wait for more info on whats happening with i5/i7 going to 32nm.
 
clarkdale/arrandale will both on [H]55, which is P55 with the graphics chip connection. Lynnfields will run on H55 fine. As always, first adopters will get screwed (no chance at the duals, albeit anyone buying now is getting a quad instead).
And sandy will have 32nm quads on s1156, but thats a ways out.
 
Last edited:
Is it certain that the lack of graphic chip connection will prevent the Clarkdales from working, rather than just making that one feature unavailable? I've no need for an IGP in this setup that I'm looking for for the P55, but if there's no future 32nm Quads that will work on it - well, that'd be a big bummer.
 
no, clarkdale will probably work on P55 without graphics, it just won't have graphics.
 
Back
Top