How much faster is intel i7 than say AMD Athlon 2 X3 445 ?
 

Is intel i7 1.5 times faster or 2 times faster than a AMD Athlon 2 X3 445?

I have this HP desktop computer that I have for about 1 year and it is too slow for video editing.I use the comnputer every day for video editing.

I have AMD Athlon 2 X3 445 , 4GB RAM and 1TB hard-drive . I'm thinking of putting it on ebay and getting a new computer.Well price is not a problem and will spend 1,500 on a faster new computer .

My question is how much faster will new computer be ? And is the AMD Athlon 2 X3 445 any good than say intel i7 ?Or did I got a low budget computer?


If the intel i7 is only 1.5 times faster or 2 times faster than a AMD Athlon 2 X3 445 than is it worth it to get a new computer?

Here's a benchmark:
http://www.cpubenchmark.net/cpu_look...hlon+II+X3+445

An Intel Core i7-3960X @ 3.30GHz processor will be roughly 5 times faster than your current processor.

So video editing should also be much faster, 5 times faster? I don't know. There may also be some video editing benchmarks available.

How fast your system is in video editing doesn't only rely on the CPU, you have also to consider the environment. You need a large amount of fast RAM, a fast storage subsystem, ... . Also the software you use will make a huge difference.

There is also a significant difference between 32-bit and 64-bit especially with things like video encoding.

Um, both are 64 bit!

for a benchmark, you could also look here
http://www.anandtech.com/bench/Product/523?vs=202

(you may have to change the exact processors compared, particularly as i7 isn't a processor but a range). And note, that while one processor wins more than the other, that doesn't mean that the processor that wins the majority of the tests is ahead in all, and even loses some tests by a significant margin.

But, beware: with benchmarks, what you get is very dependant on exactly what and how you do the test, and you probably won't find anyone testing exactly your workload and exactly your set of conditions. And, obviously, if, for example, the bottleneck is getting data off disk, even a hyperfast processor will make hardly any difference.


Well, there is a question that only you can answer - the one concerning 'what is worth it to you?'.

Incomplete linking

i7-3960X 'Extreme Edition', 14019 passmarks-
http://www.cpubenchmark.net/cpu.php?...0X+%40+3.30GHz

X3 455, 2969 passmarks-
http://www.cpubenchmark.net/cpu.php?...hlon+II+X3+455

That doesnt mean that the i7-3960X will be 5 times faster, anywhere. Passmark is a dinky little artifical benchmark, it doesnt mean a huge amount in the real world. The i7-3960X isnt a valid choice for nec207 anyway with the possible budget posted....you're looking at $1K US+ for the CPU alone. Whenever you see intel wack 'Extreme Edition' on anything, it increases the price a huge amount. Then you have to add another $200 (absolute minimum) for a motherboard. Then there is the 'new' platform issues that always pop up, etc..

Even if nec207 had money to brun, a i7-3960X isnt a great choice for now anyway IMO.

Hp desktop? Even if it came with a 'top end' CPU (eg Phenom II X6, or one of the intel i7s) HP systems are almost always 'budget'. HP (and the other 'big name' companies) dont really change that much between the low end computer and the top end, so you'll still have a cheap motherbaord, probably on older chipset, slower RAM, slower HDD, etc..

Intel i7 performance varies a lot between the oldest, slowest models (i7-920) abnd the newest, fastest versions (i7 'Extreme Edition' i7-3960X).

+1. What software are you using?

You could possibly drop a Phenom II X4/X6 into your HP system nec207. They have a lot more power than the Athlon IIs for video editing work. Not just from the number of cores (X3 has 3 cores, X4/X6 4/6 cores) but also the CPU cache (athlon II has 512k per core, Phenom II has 512K per core + 6MB L3 cache).

BTW, you generally get what you pay for. $175-200 on a AMD CPU will be similar in performance to $200-225 on an intel CPU. Its only if you are using 1 program in particular, and check some benchamrks for it (if they exist) that you can stop being general, and start being more exact.

Both will run on 32bit or 64bit OSes though. For video editing, a 32bit OS is a fair bit slower than 64bit.

Okay I checked http://www.cpubenchmark.net/high_end_cpus.html

The AMD Athlon II X3 445 comes at 2,707 and the Intel Core i7-2600K come at 9,108 so that is like 4 times faster looking at it.

The best of best if one has lots money is Intel Core i7-3960X 3.30GHz that comes in at 14,038 there is no faster CPU.


I could get a Intel Core i5-2400S 2.50GHz or Intel Xeon 2.67GHz that come in at 5,000 that would be 2 times faster than what I have.

The Xeon 2.67GHz that come in at 5,000 and the top of the line is the Intel Core i7-3960X 3.30GHz that come in at 14,038 almost 15,000 that is 3 times faster than Xeon 2.67GHz that come in at 5,000 .

I know but why would you deliberately run 32 bit on 64 bit? Unless it was absolutely necessary.

Some distros like to 'recommended' 32bit (and after they shut the 64bit part of that distros forum as 'its not needed anymore'). Theres lot of people, even on forums like this that like to suggest 32bit 'just in case', or because they ran into a problem in 2005-2008.

Lots of reasons why someone could end up on 32bit with a 64it capable computer.

Like I said above, you cant just check passmark for the number of passmarks for 2 CPUs and compare performance based on the scores. It_does_not_work.

So passmark (right now) lists these CPUs/scores (using slightly older CPUs due the page I am going to compare to)-

Intel Core i7 980X = 10234 passmarks
Intel Core i7-2600K = 9109
Intel Core i5-2500K = 6742
Intel Core i7-870 = 6172
Intel Core i5-2400S = 5045

AMD Phenom II X6 1100T = 6204 passmarks
AMD Phenom II X6 1055T = 5179
AMD Phenom II X4 965 = 4197
AMD Athlon II X3 455 = 2974 (opps, should have used X3 445, thats 2707)

Then have a look at this set of results for Adobe After Effects CS5-

http://www.tomshardware.com/charts/d...-CS5,2427.html

All 'scores' are time in seconds to apply after effects-

i7 2600K = 69
i5 2500K= 70
i7 980X = 73
i7 870 = 79
i5 2400S = 91

X6 1100T = 82
X6 1055T = 95
X4 965 = 100
X3 445 = 157

Dont forget that the numbers from the 'Adobe After Effects CS5' set of results _only_ apply to that program, and only to "Rendering 3 Streams into 1 (210 Frames)". Other secotions of teh program can have very different results.

The X3 445 is a bit under half the speed of the i7-2600K. The i7-2600K is nowhere near 4 times faster than a X3 445. The i7 980X which passmarks better than the i7-2600K is slower with AAECS5. The X6 1055T gets more passmarks than the i5-2400S, but the intel is slightly faster at AAECS5. BTW, AMD tradionally does worse with AAECS than intel....Do not assume that these figures will apply to whatever program you are using.

If that doesnt show you that you cant take 'passmarks' and try to convert them into real world performance, I dont know what will.

The X6 1100T would get you into the same sort of performance levels as the newer intel i7s, and you might only have to change CPU/heatsink. Much easier, and cheaper, than getting a whole new machine.

BTW, yes, there is no faster single CPU than the i7-3960X. For the cost of a 6 core i7-3960X, you could get up to 32cores (yes, 32, 3-2, 10x3+2) cores from an 2 x opteron CPUs (dual CPU board). Even though the AMD cores are slower, with 6 cores vs 32, the opteron will be faster for multi-core capable programs.

Also keep in mind that those numbers are only true for the exact system that was used for the benchmarks. Use a different motherboard/chipset, different RAM (speed and size), ... and you will get different scores.
If you really want to know how much faster a specific i7 is for your workload then there is only one way to find out: you have to make your own benchmarks, with your applications and the hardware you have/intend to buy.

What do you mean the Passmark is artifical ? How does the passmark work? What is the point of the Passmark than?

You know people would tell me in past that evey 18 months CPU would double is speed. By looking at all those CPU listed at that web site it cannot be no more than 3 years. So a CPU made in 2005 and one made in 2010 would be no more than 3 times faster .

So by looking at the Passmark that it would be way way way way off moores law.


Ya that not worth it . Unless I can get some thing 3 or 4 times faster it is not worth it .

Actually, Moore's Law doesn't say anything about speed (and he said two years). It is about the count of inexpensive integrated transistors. The timespan was reduced to 18 months later by an Intel executive and he changed the doubled transistor count with a doubled processing speed.
But twice the amount of transistors doesn't double the performance, have a look at the example here.

Its 'artifical' (or 'synthetic') because its not measuring a real world task. How exactly passmark generates its scores is is not something I know (and I cant tell from the passmark site).

Its still a useful benchmark. Nowhere near as useful as real world benchmarks on tasks that ther user is actually going to do IMO, but still, it has its uses.

It not passmarks fault that you are trying to apply passmark scores in a way that they were never intended to be used.

TobiSGDs link to the wikipedia 'moores law' page is well worth a look. People misquote moores law all the time.

Even if you DO believe 'CPUs double in speed every 18 months', that doesnt mean that any CPU from september 2010 will be twice as fast as any CPU from march 2009.

Not worth getting a whole new system. Have you even considered the Phenom II X6 idea I suggested?

The PhenomII X6 is a great processor & the newest 1 out is cheaper than when I bought mine.
http://www.frys.com/product/6484202?...H:MAIN_RSLT_PG
http://www.frys.com/product/6219140?...H:MAIN_RSLT_PG
Look at the price of the intel processors here
http://www.frys.com/catreq/-13350