How much better is P4 vs. AMD for Video Editing

[heiber]

I am thinking about upgrading from an AMD XP 1700 and have read that a P4 is better than AMD for video editing.

Are there any benchmarks or information that show this and how much the improvement? Taking into account the cost difference, I would like to know the impact. Also, what about a Celeron vs. the P4 vs. AMD.

Thanks.

[gshelley61]

You can find others who recommend AMD over P4. I have an XP2400+, but either is fine. More RAM is what is helpful to many. Large HD's for storage are useful, too.

[Piccoro]

What specific applications do you use?

Link
Link
Link
Link

[guns1inger]

Probably the biggest advantage that intel has over AMD for video work is hyperthreading on the P4 3.xGhz chips. It doesn't make a great deal of difference to the actual editing process, but for applications that can use it, you can see substantial (60 - 70%) improvements in rendering times.

I would not touch a Celeron for anything that requires grunt.

[DereX888]

I am thinking about upgrading from an AMD XP 1700 and have read that a P4 is better than AMD for video editing.

The only, but significant, advantage P4 has over AMDs is its SSE-2 instructions set, which speeds up i.e. encoding. *However* your software have to make use of it to take advantage of it (meaning: it have to be specifically written to use SSE-2).
Some hardware (i.e. ATI All-In-Wonder Radeon cards) also work better with machines with SSE-2 equipped CPUs.

Also with video stuff AMD's labeling/marketing convention doesnt work (as in XP 2000 is "equal to" Pentium4 2000MHz). XP 2000 with its 1.6GHz is exactly what it is - comparable to Pentium 4 1.6GHz, not to P4 2GHz as AMD would want you to believe. That marketing slogan works only with office/business stuff.

Are there any benchmarks or information that show this and how much the improvement? Taking into account the cost difference, I would like to know the impact. Also, what about a Celeron vs. the P4 vs. AMD.

Ask anyone who does video on PCs.
In this example:
Making backup copy of DVD to SVCD on 3 machines similarily equipped with same amounts of RAM, same software, same configurations etc etc -but with different CPUs - P4 will do it about 20-25% faster than same speed AMD. Celeron will be behind AMD.
But what you may save on CPU (if you buy AMD) you should spend on adding more RAM to your machine, because it is something what you never have enough if you do any video editing on it.

Just my $0.02

[flaninacupboard]

The only, but significant, advantage P4 has over AMDs is its SSE-2 instructions set, which speeds up i.e. encoding. *However* your software have to make use of it to take advantage of it (meaning: it have to be specifically written to use SSE-2).
Some hardware (i.e. ATI All-In-Wonder Radeon cards) also work better with machines with SSE-2 equipped CPUs.

Wrong. AMD chips support SSE-2 now.

Also with video stuff AMD's labeling/marketing convention doesnt work (as in XP 2000 is "equal to" Pentium4 2000MHz). XP 2000 with its 1.6GHz is exactly what it is - comparable to Pentium 4 1.6GHz, not to P4 2GHz as AMD would want you to believe. That marketing slogan works only with office/business stuff.

Wrong, an AMD chip does more work per clock than an intel chip. the numbering system is sometimes optimistic (for encoding) sometimes spot on, (general office apps) and sometimes sceptical (games). all over it's fairly well balanced and makes sense to compare them in that way.


If you were to buy a P4 3.0 chip here it would set you back £150.16. an athlon XP 3000+ would be £106.90 and an XP3200+ £139.37. Even an A64 3000+ is £144.04.
AMD clearly represents the better value.
Heading into silly territory with a P4 3.4 at £287.79 the A64 3400+ is £285.90 - the price difference is no longer relevant and the clock to model disparity is high. still not unjustified though http://www.tomshardware.com/cpu/20040601/socket_939-17.html#directx_8
shows a xp3400 as same speed as P4 3.4.

It's a complex issue, but AMD does seem to be the best for all round performance and best on price. yes, if you want the fastest encoder possible, buy the best P4 chip. if you want a sensible balance of price and performance, buy an AMD.

[DereX888]

flaninacupboard wrote:

Wrong. AMD chips support SSE-2 now.

True, finally latest XPs got it too.
But with introduction of SSE-3 the 'chase' will start all over again, assuming Intel will licence it to AMD again (it is Intel's copyright, isnt it?)
And yes, I know there are no apps written to use SSE-3 yet, but so it was when SSE-2 was introduced - for short time.

If you were to buy a P4 3.0 chip here it would set you back £150.16. an athlon XP 3000+ would be £106.90 and an XP3200+ £139.37. Even an A64 3000+ is £144.04.
AMD clearly represents the better value.

Wrong. It represents LOWER PRICE, thats all.

The value should be only counted in a tie with specific duties the machine with its CPU was built for and has to perform.
Otherwise you'd have to say that best value are actually Celerons 2.0 since they are cheapest, and i.e. at speed of 2.4GHz they don't perform any less than AMD XPs in the office world, as well as in the home users world - for which simple tasks most use their computers for, web/emails plus occasional disc burning; these ppl consist of lion's share of PC users in the real world.

Whenever I build new machine for someone i always try to determine what is the purpose for this box, because most of ppl initially want "P4 with HT". However in the end I usually build for most of them AMD-based machine - and not because AMD is any better, but just because the tasks their machine will have to perform mostly dont require more than 1+GHz speed, and thats why I suggest AMD chips instead of Intel's - to save their money, or actually 'shift' them i.e. into better graphics card.

There is no such thing as "fairly well balanced comparison" between the chips from AMD and Intel. Its an AMD's marketing telltale.
Except for the short period of time (when first socket 423 Willamette P4s came out) excepot for that short period of time AMD never had any edge over Intel in performance of 'heavy duty' work horse machines. Yet again - them AMD chips are already more than average Joe may need (ATM).


IMO in general:

You want best machine for heavy work like 'playing' with video editing/stuff - get P4 and lots of RAM

You want good home machine, for gaming etc - get AMD XP and spend what you saved on better GPU

You want cheap web/mail home box - get Celeron, Duron etc, theyre still more than average person need today. After all they made i.e. all "Titanic" digital effects on a farm of just Pentium 133 or 166 Mhz CPUs, didn't they?

[Treebeard]

Also with video stuff AMD's labeling/marketing convention doesnt work (as in XP 2000 is "equal to" Pentium4 2000MHz). XP 2000 with its 1.6GHz is exactly what it is - comparable to Pentium 4 1.6GHz, not to P4 2GHz as AMD would want you to believe. That marketing slogan works only with office/business stuff.

Wrong, an AMD chip does more work per clock than an intel chip. the numbering system is sometimes optimistic (for encoding) sometimes spot on, (general office apps) and sometimes sceptical (games). all over it's fairly well balanced and makes sense to compare them in that way.

Actually the labeling of Athlon XP was not based on comparison to the Pentium chip, it was comparing it w/ the previous AMD chip (athlon).
An athlon 1.4ghz had same clock speed as XP 1600 but the 1600 ran like what would have been considered an athlon 1.6ghz.

[DereX888]

Actually the labeling of Athlon XP was not based on comparison to the Pentium chip, it was comparing it w/ the previous AMD chip (athlon).

Yes, it is very true.
However the notion out there is that is based on AMD XP comparison to P4s, not to the older Athlon's. Most of people dont know this fact and AMD's marketing doesn't even try to explain it - I remember AMD's own banners where it was very suggestive and misleading (something like "why spend a fortune for 2GHz CPU if you can have it for less").

Anyone remember AMD's tests where they compared performance of first XPs to P4s on a machines both using DDR266 (PC-2100) instead of more adequate (back then) comparison of AMD XP with PC-2100 DDR RAM vs P4 with PC-800 Rambus RAM? How reliable are such tests when one is using 'crippled' setup of competitor product... I don't like Intel for its almost monopolistic grip on the market, but at least I have never seen such twisted 'comparisons' on their part.

Its all marketing strategy, nothing else.