Intel Moves Nehalem Launch Forward

Nehalem owns. Samples at high speeds have been available for awhile... which seemed odd considering how far off release was. Perhaps they pushed up launch because things are going unusually smoothly? And just to embarrass AMD I guess, heh.
 
unless you're a serious gamer or number cruncher you will have zero need for this CPU...
 
Also don't forget that you will need at least 3 DIMMs to populate the memory controller.

What I heard is that the memory controller doesn't need to run in triple channel mode, so you could use it with 1 or 2 DIMMs (just as 1 DIMM works in current dual channel Intel chipsets). Ofcourse you won't get the full performance, but it's a good starting point, you can upgrade both the memory capacity and performance later.
 
I would suggest buying on the tick. The Tock is when they take an existing process and create a new architecture. The Tick is when they take a proven architecture and move it to a smaller, more efficient process (i.e. 55nm-45nm as with Penryn). This will generally lower prices because they can produce more die per wafer. As we saw with the e8400 and e 8500, which were faster, more power efficient, ran cooler, and were still around the same price as the previous gen.

While Penryn was a nice improvement over Conroe, you're forgetting the huge leap that it was from Presler to Conroe (the previous 'tock'). A sinple mid-end Conroe like the E6600 could easily outperform any Pentium ever built, including the Extreme Editions. So they redefined the complete CPU pricing. What was high-end before the 'tock' was low-end after the 'tock', and the rules of price/performance were rewritten. Nehalem will do this again, although probably in a less dramatic way than Conroe did, because the Core2 architecture hasn't quite run out of steam yet in the way that Pentium 4/D did.

Aside from that, Penryn prices have been 'artificially inflated', so we've not actually reaped much of the benefits of the lower production cost. Best example is the Q6600 which is still by far the most bang-for-the-buck quadcore.
 
I don't think the mainstream chips will be available until 2009.
The Q9450 didn't come out until 4 months after the QX9650.
 
Interesting, but I am on the 4 year upgrade plan. I built my system last year. I figure next year I get a dirt cheap 45nm quad and use that two more years. Then 2011 I start looking at what looks like another good 4 year setup.

Glad I no longer suffer from bleeding edge-itis. :D
 
Well I was planning on upgrading before the Spring 09' semester so it looks like it might work out well. My E6600 has held up very well over the last 2 and change years. I just hope that some decent motherboards get released that don't cost a ton of $buck.
 
Supposedly they'll be launching sub 200 dollar motherboards, of good quality, at launch. Which will be a change from their, as of conroe generation, current trend of releasing good quality motherboards in the +$200 dollar region.

Only time will tell how well they hold up to this promise.
 
Is that true? The board will not boot with less than 3 DIMMs? That seems like a bit of an oversight by Intel.

As far as I know that's not true. It can run in single or dualchannel... in fact they plan to release two lines of chips, one with dualchannel controllers and the other with triple channel.
 
I never meant to say that 3 channel memory was required to boot, but why would anyone buy a $500 + MB/CPU and then skimp out on a $75 memory module? I had not heard of anything about the 2.66 Ghz Bloomfield having only 2 channel memory. Wiki shows it to have all 3 channels. If it still has 3, I plan to buy the 2.66 and a watercooling setup and try to beat the clock speed that the 2.93 can get on air.
 
I had not heard of anything about the 2.66 Ghz Bloomfield having only 2 channel memory.

I've heard a rumor to that effect, but only once. Unless I hear it again from more places, I wouldn't put too much stock in that rumor.

I agree though; if I were to go through the trouble of getting a Bloomfield CPU and a X58 board, I'd rather get 3 sticks of cheaper, slower memory than 2 sticks of uber-premium memory. The third channel will more than make up for any insignificant difference in the individual memory sticks.
 
Oh, but what Scali2 said is true though; there will be a whole line (actually two) of dual-channel Nehalems. Bloomfield, that's coming out soon, is triple-channel. Lynnfield, on another socket, will only be dual-channel though. Dual-core Havendale, on the same socket as Lynnfield, will not only utilize dual-channel RAM, but have integrated graphics directly on the CPU (as opposed to the chipset).
 
Nope, Bloomfield will be Socket1366. Apparently it is slightly larger than the 775 socket.

Also known as Socket B. If it is indeed larger than 775 then get excited for new heatsinks. Larger contact area is good.
 
I never meant to say that 3 channel memory was required to boot, but why would anyone buy a $500 + MB/CPU and then skimp out on a $75 memory module? I had not heard of anything about the 2.66 Ghz Bloomfield having only 2 channel memory. Wiki shows it to have all 3 channels. If it still has 3, I plan to buy the 2.66 and a watercooling setup and try to beat the clock speed that the 2.93 can get on air.

You must have missed the sarcasm.

Anyways, to get optimal performance out of the board you'd obviously want to populate all 3 channels. However, this will (with the current/projected prices of DDR3) increase the initial cost of going Nehalem.
 
Well I was planning on upgrading before the Spring 09' semester so it looks like it might work out well. My E6600 has held up very well over the last 2 and change years. I just hope that some decent motherboards get released that don't cost a ton of $buck.

I'm with you on that one, I'm still running my E6600 I bought on launch day. In my adventures with the chip its had a great life. I will be upgrading to bloomfeild at or near launch depending on the motherboards. Heres hoping ddr3 prices drop at least another $50 around launch. ^^
 
Back
Top