Measure of Frequency. Sooo… technically it is the QUANTITY of cycles of alternating current signal makes in ONE Second.
Speed is not a factor when it comes to processors. It just seems that way because the higher the frequency the MORE they can do.
That’s why an 800 MEGAHERTZ cpu that is playing a video will NOT play that video any faster on a 3.06 GIGAHERTZ cpu that is playing that same video.
THAT is the biggest misconception of the whole gigaherts myth.
okay?
Right. But that’s the misconception. They all get to the speed. Your 600 hp vehicle is also geared incorrectly for speed, it’s geared for power.
Let me try again.
Say you have an 800mhz pc. Open calculator. Do the operation 100 X 100.
Do the same with a 3.06 ghz pc and you’ll find they both do the calculation at the same speed.
When you do calculations in the billions, that’s where MORE is needed to see and effect.
So it is NOT the cpu that will make the computer SEEM faster, it’s amount of data you are subjecting it to.
If you installed Windows 98 in each of the two computers, you’ll find they both perform quite similarly. If you put Winblows Xp in BOTH, you’ll find the PC with the cpu that can DO MORE calculations will SEEM faster.
Yes, but I’m not talking about calculator. I’m talking about converting 120 minutes of movie from NTSC to PAL, or changing a video from .avi to .mpg so I can burn a VCD. But I guess I’ll have to accept that the software in question can’t make use of the higher performance of the computer.
Of those things mentioned, booting up XP, starting CivIII, and watching video are all hard drive limited (unless the video is crazily-compressed). A faster CPU will not speed them up much at all because the hard drive is slow.
A lot of Photoshop and Illustrator work and video encoding will be memory speed limited. A faster CPU will not help much (although a bigger cache, or more advanced branch prediction, or a bunch of other processor “frills” will probably help).
It is true that there are harder physical and electrical problems with getting faster and faster processors. But for the most part, advancement in this direction is stalling because it doesn’t help any. Why invest R&D in faster chips just to bring your cache-miss penalty up a few hundred cycles?
While you are correct that you won’t notice the speed difference when applied to a very small problem, it makes a big difference on problems that are not trivial.
MHZ and GHZ measure the cycles per second (the rate the electricity moves back and forth between the low and high voltage).
Each cycle you have the ability to do “something” (although some chips operate on the rising edge and the trailing edge, in which case they can do “something” 2 times for each cycle).
If you have 2 CPU’s identical in all respects, except CPU A is 1ghz and CPU B is 2ghz, then CPU B has the ABILITY to be TWICE as fast as CPU A (excluding memory latency issues, etc.)
I couldn’t have said it better.
^ to add to this,
the proof is, that Mac OSX Tiger ran circles around Windows XP Pro when run on the same High powered machine.
Yes you could. The problem isn’t most likely a processor-speed issue in the first place.
From what I understand, multimedia processing isn’t necessarily CPU-bound, but I/O-bound.
Kinda like saying that the time it takes the 600 hp car to take something from point A to point B is limited by the time it takes to load and unload that thing, instead of the time spent driving.
If it takes you an hour to load or unload the vehicle, and you’re driving 30 miles, it doesn’t really matter whether you have a car that goes 60 mph or 70 mph because you’re still spending 2 hours loading and unloading, regardless of whether it takes 30 or 35 minutes to get where you’re going. Sure, the 70 mph car is faster, but the real bottleneck is the loading and unloading, not the speed of the car.
I agree that the marketing literature is misleading at best- it took me several years and several analogies to impress upon my father that a new CPU is not the answer to his problems; memory and system bus speed are much more important these days.
It’s not a processor/OS optimization issue at all, in most cases.
Sigh, I’m guessing this comes from the perennial Photoshop face-off tests, Mac v. Intel. Cite please?
On reviewing the thread: leandroc76, please disregard my question.
As others have said, booting and loading tasks are probably disk-limited, so you shouldn’t see much improvement.
As far as “watching video” is concerned, I’m assuming that’s a mistype on your part, since surely you wouldn’t expect a real-time process like watching video to run faster on one machine compared to another (providing both met the minimum requirements for video)?
With AVI-MPEG transfers, the CPU is going to be thrashing the cache and the main RAM a fair amount. This is one area where you want to run at the highest RAM speeds possible, and the difference between systems with different FSB speeds can be greater than the difference between systems with different CPU clock speeds. Are you running the same FSB speed in each of your CPU comparisons?
This statement runs counter to my (rather extensive) experience. Of course, some applications are disk-limited, some are RAM-limited, but it’s not hard to write applications that are CPU-limited. The characteristics of the above applications are often OS-independent. To suggest that Win98 applications perform similarly on CPUs of different clock speeds, but that WinXP magically sorts out the sheep from the goats, runs contrary to my experience. Perhaps I could ask you for clarification, or a cite?
I’m intrigued. What machine was this that managed to run Mac OSX and WinXP Pro, both in native mode? Naturally, if one of them wasn’t running in native mode then the comparison is meaningless, and if it wasn’t the same machine, but two different machines, then the quoted statement of yours is demonstrably false. Could you please clarify, preferably with a cite?
Well, Apple is switching to Intel processors, and several Pentium 4 based development machines have trickled into the hands of those who develop software for Apple. These machines can dual boot Windows XP natively. I presume that is where he is getting his info one, though I haven’t seen any articles comparing the two OS’s on the same machine, just some brief articles that say “OSX runs blazing fast”.
Well, 30 fps is better than 25, innit? 
No, but loading the file, to see it. But I guess that’s just a memory thing.
Still, none of this changes the fact that the second to last change was between machines that essentially doubled performances WRT Ram, CPU and such, and this was such a jump as well. And the second to last jump made a comparatively larger leap in perceived performance than this one did. Hence my OP.
You are very correct. The point I was making to the OP was that the CPU is NOT the factor of performance. The Gigahertz is just a number of quantity and has nothing to do with OVERALL Speed.
BTW: There was an article about Tiger on an Intel PC and Win XP on another Intel PC with the same specs. Now I can’t find it. :mad: Randomletters basically put it correctly.
leandroc76, it seems to me that objecting to the use of Gigahertz as a speed because it’s a “quantity” is like saying “My speedometer doesn’t tell me the speed of my car, just the quantity of miles it would travel in an hour.” I’d say the number of Gigahertz is a legitimate measure of the speed of the processor. The fact that the processor speed isn’t the determining factor in the overall speed is a separate issue.
The problem is that clockspeed is only valid in predicting speed within the same processor design - IE, a 3ghz Pentium 4 will always be faster than an otherwise identical Pentium 4 running at 2ghz. Throw different processor designs into mix, and clockspeed becomes useless when it comes to measuring performance - a 2ghz Pentium M or Athlon 64 will perform better than a 3ghz Pentium 4 in most applications, for example.
A better car analogy would be a tachometer - obviously, a 4 cylinder car that’s engine can run at 5000 RPM would let you go faster than a 4 cylinder car that has an engine limited to 4000 RPM. Throw in V-6 and V-8 cars though, and knowing how many RPM the engine can turn becomes less useful, when making comparisons.