Mini 1.5 Core Solo vs. Dell 2.8 GHz P4HT - WHAT?

[gkarris]

Hey, what's faster?

At church for recording the sermons, we had a Dell 2.8 GHz with Cubase and a pro audio box. The machine is no longer under warranty and has started acting up. I took the opportunity to get a Mac Mini with OS X instead.

We got a Mac mini with 1.5 Core Solo with 1 Gig RAM. I used it yesterday (Sunday) to record the sermon (usually about 40 minutes in length). I exported the audio mixdown and it took the Mac Mini less than 1/2 the time to export the AIFF file than the Dell.

What's up? A machine with a 1.5 GHz processor takes less time than the 2.8 GHz one???

 

[Eidorian]

What's up? A machine with a 1.5 GHz processor takes less time than the 2.8 GHz one???

Shorter instruction pipelines and a more efficient processor in general. Any clue on which 2.8 GHz Pentium IV?

 

[Queso]

That's a good example of the megahertz myth. A Core Solo processor is a generation ahead of a Pentium, so can accomplish much more at lower clock speeds.

 

[gkarris]

Shorter instruction pipelines and a more efficient processor in general. Any clue on which 2.8 GHz Pentium IV?

The Dell's over 2 years old already, so whatever was out back then...

 

[Eidorian]

The Dell's over 2 years old already, so whatever was out back then...

Sounds like a late Northwood or early Prescott then.

Core Solo would slam Northwood. I'm not sure about Prescott though.

 

[4God]

Yeah, and don't forget about bus speed. I'd bet that dell was running 333 or 533 Mhz front side bus as compared to the 667 on the mini. Memory speed may play a minor role in that as well.

 

[gkarris]

Yeah, and don't forget about bus speed. I'd bet that dell was running 333 or 533 Mhz front side bus as compared to the 667 on the mini. Memory speed may play a minor role in that as well.

Here's the information from Dell's support site:

1 C1431 PROCESSOR, 80532, 2.8G, 512K, 533, SOCKET N, DECISION ONE

 

[ricgnzlzcr]

I wish I understood more the difference between the processors. I had a 3.2 Ghz Pentium 4 Extreme and I'm sure a core solo 1.5 would be faster than it. I wonder why that is for the most part

 

[balamw]

Also don't forget hyperthreading. The 2.8 GHz P4 is most often seen by the OS as two 1.4 GHz cores. If the app isn't optimized to use both virtual cores, you won't see the beenfit. So unless you turn off hyperthreading, a single 1.5 GHz Core with a shorter pipeline will smoke the P4.

B

 

[Makosuke]

I wish I understood more the difference between the processors. I had a 3.2 Ghz Pentium 4 Extreme and I'm sure a core solo 1.5 would be faster than it. I wonder why that is for the most part

Mmm, in that case it would be close, but the edge would probably go to the P4. Now, put it up against a 2.16GHz dualcore and it would eat it for lunch.

The sad fact is that in pushing the Megahertz Myth Intel hamstrung itself for years with a creaky processor design that was designed for clockspeed over actual speed. They finally realized the corner they had painted themselves into with marketing sometime around when their mobile architecture (on which the Core chips are based, if memory serves--I haven't followed Intel stuff very closely) started making everything else they made look like an overheated dinosaur.

Heck, their server-only Itanium chips were speedy at something like 1/3 the clock of a P4.

Bottom line is, after a long time of milking the P4 for MORE than it was worth, Intel finally has a decent foundation on which to build. The Core chips really do perform extrordinariy well, and at low power. Heck, a 2.16GHz Core Duo has been shown in real-world tests to be competitive with 3.75GHz P4EE chips. When your new "mobile" processor is FASTER than your top of the line desktop processor, you know you've been doing something wrong, and are doing something right now.

 

[Eidorian]

The sad fact is that in pushing the Megahertz Myth Intel hamstrung itself for years with a creaky processor design that was designed for clockspeed over actual speed. They finally realized the corner they had painted themselves into with marketing sometime around when their mobile architecture (on which the Core chips are based, if memory serves--I haven't followed Intel stuff very closely) started making everything else they made look like an overheated dinosaur.

Well, pumping more megahertz into a CPU does work. It's just that with a Prescott core you hit 200 watt CPU's and crazy cooling for that. Intel hit the wall at 4 GHz and gave up at 3.8 GHz.

 

[ricgnzlzcr]

Mmm, in that case it would be close, but the edge would probably go to the P4. Now, put it up against a 2.16GHz dualcore and it would eat it for lunch.

The sad fact is that in pushing the Megahertz Myth Intel hamstrung itself for years with a creaky processor design that was designed for clockspeed over actual speed. They finally realized the corner they had painted themselves into with marketing sometime around when their mobile architecture (on which the Core chips are based, if memory serves--I haven't followed Intel stuff very closely) started making everything else they made look like an overheated dinosaur.

Heck, their server-only Itanium chips were speedy at something like 1/3 the clock of a P4.

Bottom line is, after a long time of milking the P4 for MORE than it was worth, Intel finally has a decent foundation on which to build. The Core chips really do perform extrordinariy well, and at low power. Heck, a 2.16GHz Core Duo has been shown in real-world tests to be competitive with 3.75GHz P4EE chips. When your new "mobile" processor is FASTER than your top of the line desktop processor, you know you've been doing something wrong, and are doing something right now.

Wow, I guess these new merom chips are going to do wonders for macbooks and macbook pro's. I can't wait to see what's going to be out in 2008 when I can hopefully stop resist upgrading this G4 powerbook.

 

[andrewheard]

The 2.8 GHz P4 is most often seen by the OS as two 1.4 GHz cores.

Where on earth did you hear that?

 

[balamw]

Where on earth did you hear that?

I'm just making it up.

http://en.wikipedia.org/wiki/Hyperthreading

Actually, I've benchmarked plenty of P4 systems with HT turned on at (clock speed)/2 for various floating point heavy single threaded apps. If a P4 system has HT turned on, it can easily be outpaced by a Pentium M of half the clock speed.

EDIT: I'm attaching a screenshot from my own 2.8 GHz P4 system with HT and youcan clearly see two "virtual cores" one which is in heavy use, and the other not.

B

 

[Killyp]

I can't wait for the Mac Pro so we can be completely sure we're wiping the floor with any Dell or Alienware machine in all aspects, ie, MHz, Number of processors etc....

Then 'they'll' have no way of retaliating...

 

[Capt Underpants]

I can't wait for the Mac Pro so we can be completely sure we're wiping the floor with any Dell or Alienware machine in all aspects, ie, MHz, Number of processors etc....

Then 'they'll' have no way of retaliating...

They'll get the same processors Apple gets...

 

[andrewheard]

I'm just making it up.

I was really hopeing that you were. That just seems so dumb. I just don't understand how it could be possible for games to run well with HT enabled if thats true. Any insight?

 

[balamw]

I was really hopeing that you were. That just seems so dumb. I just don't understand how it could be possible for games to run well with HT enabled if thats true. Any insight?

Same way they run well on Core Duo processors and the like. Multi-threading baby. Do more than one thing at the same time and HT of multiple cores will give you a benefit.

B

 

[DannySmurf]

Same way they run well on Core Duo processors and the like. Multi-threading baby. Do more than one thing at the same time and HT of multiple cores will give you a benefit.

Yes, but while HT does give some boost if the applications you run are written to be multithreaded, HT will not give you anywhere near the boost that two actual cores (or two actual processors) will. A HT processor can still only execute one instruction at a time (whereas a dual-core system can execute two simultaneously).

And of course, the software has to be written to take advantage of that. Software that isn't doesn't get any speed boost from HT OR multiple cores.