• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Why no higher clockspeeds?

Heat problems is what I was told, the current cooling solutions aren't good enough for the higher clock speeds. Switching the transistors on and off produces heat, the quicker you do it, the more heat is produced
 
Judgeneo said:
Heat problems is what I was told, the current cooling solutions aren't good enough for the higher clock speeds. Switching the transistors on and off produces heat, the quicker you do it, the more heat is produced
Click to expand...

Cooling isn't really such an issue with some of the 32nm chips, particularly with Sandy Bridge. You can basically hit the same stable overclocks with a mid-end and high-end cooler on a 2500k or 2600k (4.6-4.8).
 
the motorway analogy works perfect, its all about innovation, efficiency rather than raw clock-speed. in the real-world there will always be X amount of cars using the same stretch of road, only so many cars can occupy the same space at once so we get a build up of traffic, adding more lanes makes the road more efficient without resorting to adding more speed, so effectively you move more cars through the same space in the same amount of time.

same sort of thing applies to processor, only got so much electricity available, only got so much space available so making the best, most efficient use of the materials. the old analogy I used to a friend in the K8 and Pentium days was the Pentium is a dude with short legs, who takes ten small, fast steps to cover a distance and the K8 was a taller bloke, with longer legs covering the same distance in less steps, doing the same for less.

also there is only so fast a transistor can switch so that is a huge limitation, combined with the excessive heat generated by this rapid switching makes it impractical to keep ramping up the speed. so until these limitations are overcome (Graphene replacing Silicon maybe?) don't think we will see any dramatic jumps in the frequency of processors, just more innovation, more efficient use of power and more efficient multi-tasking, which is better to be honest.
 
ian54 said:
But you wouldn't design software that way any more, that's my point, so clockspeed isn't as important as parallelism.
Click to expand...

Whether you would design software that way any more doesn't prevent the situation where some applications could benefit from it though. Some older software (games in particular) benefit from high clockspeed. There was a thread on these forums a few months ago where somebody wanted an extremely fast system for single threaded applications, where basically latency was key, it smelt like it was intended to be used for trading or something like that.

Essentially as I said, I'd like to have my cake and eat it, I want more clockspeed not just more cores! As the number of cores increases we'll also start to get get diminishing returns unless software gets very smart, because the more cores you have the more evenly distributed you need the demands to be.

topdog@OC said:
Roll on 2010+ and still it seems "only we're just now noticing". Development cycles take a long time to change, but there's been time enough for a couple of full cycles in this time at least.
Click to expand...

To be fair in the OP I did start to talk about noticing it prior to that when I upgraded 3 times all at the same clockspeed - although it is kinda noticing the obvious! :)

So single threaded apps are still way faster on a modern processor than they were on a a P4.
Click to expand...

Again this is something I conceded to in the OP - but again I want to have my cake and eat it! A P100 was faster than a DX4-100 yet they still doubled the clocks, a P2-233 was faster than a P233MMX yet they still nearly doubled the clocks etc (ok not exactly fair comparison due to chipset differences but you get the point).

edit: happy to let this one go now as people have started explaining what the technical limitations are with transistors which is what I was after, I just wasn't happy with this idea that there was 'no need' for higher clockspeeds due to modern processors being faster per clock and having more cores.
 
Last edited:
Regarding the 1ghz thing, I remember a friend I was in Halls with at uni (1998-9) who was studying computer science told me that 1ghz shouldn't be possible because of some health and safety reason (I can't remember what) when it hit that frequency, that it was gonna start emitting dangerous waves or something :)
 
even wiki get's it right though. Moore's law has been applied to other areas in the past, such as clock speed, but Moore's law (ie, what Moore actually said) as nothing to do with anything other than the doubling of transistor count.

blurgh lol
 
You must log in or register to reply here.
Back
Top Bottom