45 / 32nm single core

pr2thej

Distinguished
Sep 25, 2008
1,352
0
19,290
I was on my way to work this morning, and was thinking about ripping.org and the fact that a single core currently holds the overclock record by a lonnngggg way. Bearing in mind the chip in question was 65nm its logical to assume that a single core based on more efficient tech would clock higher.
I know the market isnt going this way, it wont be a big seller etc but I dont know much about what it takes to produce a new chip.
I would certainly buy it at a reasonable price just for kicks.
A quick google shows that AMD wont make any more single cores, and Intel's Silverthorne is a 45nm single-core designed for low power usage instead of high clocks.

Coupla questions - can you see that single core record ever being broken? (8.2GHz)
If not how long do you think it will be before we see any chip break it (dual, quad etc)? It seems the trend is to bring chips out that can manage a reasonable 4GHz OC and then throw on a bunch of cores which reduces the speeds back down as seen with Dual / Quad releases. So the 6 (8?) core chips will run at stock of 2GHz ish and we are back to where we started!


/ramble
 
I suppose that AMD or Intel could make a single core platter (disabled dual core) for bragging rights, could be seen as a useful maketing gimmick.

It would be very expensive, but the majority of the work is already done, and they are stockpiling so its not like they are actually selling what they produce, so stopping the 'machine' to change over to a different print wouldn't actually cost them anything, they are already employing the people doing the work and there is no lost capacity that is revenue earning.
 
An 8.2 Ghz P4 still doesn't have a great IPC because of the crap design.

A 1.5Ghz P4 is about on par with a 900Mhz Athlon.

So a 4.92 Ghz Athlon would waste it on most tasks.

It is all about IPC.