Tech Question (Daz?)

[The Boosh!]

quote from another forum i use

http://cgi.ebay.co.uk/DUALCORE-INTEL...hippingPayment

Opinions please.

Although i don't know why people automatically assume that if a core 2 duo processor is 2.0ghz then it is actually 4.0ghz because that is wrong.

is he right? i always thought if you had a 1.7 ghz dual core it ran at 3.2 ghz:quiet: if hes right can you explain why

 

[Daz]

I'm trying to find an article which can explain it better than me..

 

[The Boosh!]

thanks maan

 

[Daz]

In simple terms, it doesn't really mean 2Ghz == 4Ghz.
It means, two threads can run at 2Ghz each at the same time, in parallel.
Multithreaded applications could in theory use both cores at the same time so "in theory" maybe.. but, a lot of apps aren't, so they'll only use one core in reality.

I can't believe there isn't a guide about it somewhere on the internet going into the detail - I'll keep looking, it's bugging me now. as I'm sure my explaination is a bit dodgy

 

[The Boosh!]

i understand ish thankyou

 

[Daz]

Ah.
http://www.geekstogo.com/2007/07/24/do-you-need-a-quad-core-cpu/

 

[Daz]

So technically, if two threads of the same program are maxing each core, then you would be close to 4Ghz in theory for overall processor usage, hence the advert is right.

It's one of those which sounds much better by doubling it, but in the real world, a single 4Ghz processor would be better than a dual core 2GHz (or even quad core 2GHz) if the application you are testing out is single threaded

 

[Mike_182]

Consider it like this:

Twin engined car - yes it makes it faster, but it doesn't make it twice as fast. Depends from what point you're comparing it to, and how efficiently it's done.

The more complicated explanation involves a conversation including Hyperthreading technology, shared vs. discrete cache etc., which can get very boring. But basically, no, you get 2x1.7Ghz processors, not an effective 1x3.4Ghz (sorry to correct your mathematics as well!)

As Daz linked...

 

[The Boosh!]

so does the usage on the processor vary. like it never actually runs at 4ghz unless your caining programmes?

 

[Mike_182]

So technically, if two threads of the same program are maxing each core, then you would be close to 4Ghz in theory for overall processor usage.

It's one of those which sounds much better by doubling it, but in the real world, a single 4Ghz processor would be better than a dual core 2GHz (or even quad core 2GHz) if the application you are testing out is single threaded

Not entirely true - as Intel quickly worked out with the Pentium D. Consider that a Core2 1.8Ghz can quite easily outpace a 3Ghz Pentium D, even in singlethreaded applications. Pipelines, cache and all sorts of other design elements make clock-for-clock comparison near pointless on any processors of different design models.

AMD have been playing on this for years, with Athlon XP 2000+ being clocked much slower at 1666Mhz. This was due to it's design being more effective in places, allowing it to bench similarly to a Pentium 4 2Ghz. Unfortunately educating the public is hard work, which is why they took the easier approach of giving a 'comparitive' speed than an actual speed on the packaging - otherwise consumers would take the one with the bigger number as it sounded better.

 

[Daz]

If you're running 100 processes, then you will be running the processor each core of the processor at 2.0Ghz.

Say you have a JCB on running on a computer at 4Ghz, single core.

Daz is a workman, who can dig half as quick as the JCB.
You run Daz on a dual core CPU, and you get two of him.. both running at 2.0Ghz. Do Two Daz's Beat a JCB? Not really.. cus' one would be digging one thing, the other Daz would be digging another thing.. whereas the JCB is digging one thing on his single core 4Ghz. So, Two Daz (2x2.0Ghz) isn't equal to one JCB (4Ghz).

Hahahaha. WTF. I need to get out more.

 

[Chrisjk]

Basically what you want is a cyrix 2 333mhz processor with 16 bit capabilities.

 

[Daz]

Not entirely true - as Intel quickly worked out with the Pentium D. Consider that a Core2 1.8Ghz can quite easily outpace a 3Ghz Pentium D, even in singlethreaded applications. Pipelines, cache and all sorts of other design elements make clock-for-clock comparison near pointless on any processors of different design models.

Gets very complex, quickly.

As said, you can only compare like for like at the Ghz level from the same CPU family really.. but, it makes for some good Saturday night foruming

 

[The Boosh!]

If you're running 100 processes, then you will be running the processor each core of the processor at 2.0Ghz.

Say you have a JCB on running on a computer at 4Ghz, single core.

Daz is a workman, who can dig half as quick as the JCB.
You run Daz on a dual core CPU, and you get two of him.. both running at 2.0Ghz. Do Two Daz's Beat a JCB? Not really.. cus' one would be digging one thing, the other Daz would be digging another thing.. whereas the JCB is digging one thing on his single core 4Ghz. So, Two Daz (2x2.0Ghz) isn't equal to one JCB (4Ghz).

Hahahaha. WTF. I need to get out more.

Now i understand. your such a boff :rasp:

 

[Daz]

You could argue that the two Daz's are digging more than the single JCB.. which is how it generally works