Athlon XP Vs the Pentium M

K

klowngoblin

[H]ard|Gawd
Joined
Oct 28, 2001
Messages
1,087
Encoding a 4.3GB dvd from Xvid videos with nero vision express 3 (battlestar galactica eps). i duplicated the file structure on my desktop and laptop so they both have the same files and i saved the project so everything would be the same.

Athlon XP 2500@3200 512KB Cache + 1024MB PC3200 Ram 160GB 40GB 7200RPM --> 160Raid 0 Drives= 6hrs
Pentium M 1.6Hz 2MB Cache + 1024 PC2700 Ram 40GB 5400RPM --> 80G 5400 RPM = 4hrs

what is going on here? how can my desktop cpu be slower than my laptop?
 
because the pentium m is an amazing chip.
and the athlon xp is kinda outdated. :cool:
 
lesman said:
I agree.
Click to expand...

yeah its outdated but i dont have $400 to updgrade CPU, mobo and HSF :(. i want to get a gigabyte mobo (GA-K8NS Ultra-939 ) + Athlon 64 3200 or 3500
 
then use your laptop as your desktop too. just hook up a full sized keyboard, monitor and mouse to it whenever you bring it to your desk ;)
 
(cf)Eclipse said:
then use your laptop as your desktop too. just hook up a full sized keyboard, monitor and mouse to it whenever you bring it to your desk ;)
Click to expand...

once i upgrade my desktop it will reign superior again, the HDD's in my desktop can just slaughter any notebook drive since i have 4 and i can copy across them all. my laptop only has a ext USB drive that maxes out at 13MB/ s (though its dead on 13MB)
 
It would be nice if they would release a new P-M board with something other than the 855 Chipset. I've heard they were coming, but haven't seen one.
 
klowngoblin said:
Encoding a 4.3GB dvd from Xvid videos with nero vision express 3 (battlestar galactica eps). i duplicated the file structure on my desktop and laptop so they both have the same files and i saved the project so everything would be the same.

Athlon XP 2500@3200 512KB Cache + 1024MB PC3200 Ram 160GB 40GB 7200RPM --> 160Raid 0 Drives= 6hrs
Pentium M 1.6Hz 2MB Cache + 1024 PC2700 Ram 40GB 5400RPM --> 80G 5400 RPM = 4hrs

what is going on here? how can my desktop cpu be slower than my laptop?
Click to expand...
hmmmm only takes my main rig 3 hours to rip a 2hour dvd
 
ALL4AMD said:
hmmmm only takes my main rig 3 hours to rip a 2hour dvd
Click to expand...
Wait, I just thought about what you typed...it takes 3 HOURS TO RIP A DVD?????? It takes me 13-26 mins to do that....and then maybe 15-20 mins to burn it....did I miss something??? :confused: :confused:
 
I can encode on a XP1700+@1.47Ghz faster than 6 hours.

I'm assuming he meant encode the DVD to Xvid also :).
 
Perhaps it's the SSE2 of the Pentium M that makes the difference.
Part cache, part SSE2 probably.
 
I know my P4C 2.8 @ 3.2 w/ HT, 1GB DDR 433, Dual Raptors in Raid 0, NEC 16x DVD - RW

Tear Through DVD's :p
 
kirbyrj said:
I can encode on a XP1700+@1.47Ghz faster than 6 hours.

I'm assuming he meant encode the DVD to Xvid also :).
Click to expand...
oops yeah i just realized what i typed but yeah dvd to divx is 3hours
 
I'd be worried too if that setup was getting beat by a 1.6hz processor. :eek:





;)
 
NOTE: most ripping software is encoded to run better under the intel SSEI2 instruction sets - this should count for some of the speed difference + the cache - but the 6 hours sounds slow - what else is running at the same time you were ripping?
 
Dr. X said:
I'd be worried too if that setup was getting beat by a 1.6hz processor. :eek:





;)
Click to expand...

Why? A 1.6GHz Centrino is like a 2.6GHz+ P4 in terms of performance.
 
yeah, generally a64 and dothan have very similar ipc. so you can almost compare them by the clock speed.
 
How well does the pentium m compare to athlon 64's? It looks like it would perform better at the same speed.
 
delusion_2005 said:
Because the The Pentium chip is just a beast ^^
Click to expand...
um.. well it's just a tweaked p3. it shows how much the netburst architecture sucks :(

and andrew, i answered that in my other post, though somewhat indirectly.
 
Josh_B said:
Why? A 1.6GHz Centrino is like a 2.6GHz+ P4 in terms of performance.
Click to expand...
yeah, and his XP should kick the crap out of a 2.6ghz p4

what motherboard are you running? do you have recent chipset drivers installed?
 
(cf)Eclipse said:
and andrew, i answered that in my other post, though somewhat indirectly.
Click to expand...


Yeah, too bad no one will underclock to 1.6ghz to see how exactly they compare. I'm just scared that intel's making two different product lines.
 
i suppose, but dothan has sse, sse2 and sse3, whereas the a-xp just has sse.

(i was also being silly and was thinking of a64's :eek: )
 
klowngoblin said:
Encoding a 4.3GB dvd from Xvid videos with nero vision express 3 (battlestar galactica eps). i duplicated the file structure on my desktop and laptop so they both have the same files and i saved the project so everything would be the same.

Athlon XP 2500@3200 512KB Cache + 1024MB PC3200 Ram 160GB 40GB 7200RPM --> 160Raid 0 Drives= 6hrs
Pentium M 1.6Hz 2MB Cache + 1024 PC2700 Ram 40GB 5400RPM --> 80G 5400 RPM = 4hrs

what is going on here? how can my desktop cpu be slower than my laptop?
Click to expand...
As people have mentioned earlier, it's entirely down to the SSE2 support in the Intel chip. Xvid does considerably better on Intel chips simply because of this.

The AMD chip will generally be better all round especially in FPU intensive apps such as games. But anywhere where there's an SSE2 or 3 optimisation, it's gunna fall behind. Fortunately, these optimisations tend to only really occur in video editing software etc which you have discovered yourself ;)
 
(cf)Eclipse said:
i suppose, but dothan has sse, sse2 and sse3, whereas the a-xp just has sse.

(i was also being silly and was thinking of a64's :eek: )
Click to expand...

Are you sure that Dothan supports SSE3? I thought that was going to be in the successor.
Besides, SSE2 is the instructionset that gets the big gain in A/V encoding/decoding, SSE3 doesn't really do much there.
So I very much meant SSE2.
 
spine said:
The AMD chip will generally be better all round especially in FPU intensive apps such as games. But anywhere where there's an SSE2 or 3 optimisation, it's gunna fall behind. Fortunately, these optimisations tend to only really occur in video editing software etc which you have discovered yourself ;)
Click to expand...

That may change... There have been compilers that can replace pretty much all FPU code with faster SSE/SSE2 equivalents for a while now. Especially on P4 they do wonders.
I suspect these will be used more commonly in the future, especially on the 64 bit platform, since SSE/SSE2 are standard there. There's no reason to support standard x87 FPU code at all on the 64 bit platform.
 
the problem i have with sse2. i remember reading an article on someone who encoded a movie with sse2 optimizations and without. though it was faster with it on, the quality was lower, even though he had set it the same. something to think about.

and no, i'm not positive that the p-m has sse3, but i thought it did. i suppose that since it's based on the p3, it probably doesn't.
 
lithium726 said:
yeah, and his XP should kick the crap out of a 2.6ghz p4

what motherboard are you running? do you have recent chipset drivers installed?
Click to expand...

Not in terms of encoding, it won't.

Besides, I said 2.6+GHz P4. Note the '+' sign.

I do agree that in general performance, the 3200+ should be slightly faster though.
 
(cf)Eclipse said:
the problem i have with sse2. i remember reading an article on someone who encoded a movie with sse2 optimizations and without. though it was faster with it on, the quality was lower, even though he had set it the same. something to think about.
Click to expand...
Seriously? I'd like to see that...
 
spine said:
Seriously? I'd like to see that...
Click to expand...

Actually, I think he might be right. I remember that when using 3DNow! optimizations, the quality would be noticeably lower than using x87 FPU only. I imagine that SSE2 would be similar, but more precise than 3DNow!, due to the fact that it is 64-bit double-precision. Does that mean that x87 FPU code makes use of extended precision? (80-bit).
 
it was a long time ago. i tried to find the article, but alas. for now just ignore what i said because i can't prove it, however it kinda makes sense. it's an optimization.
optimizations mean cutting corners
cutting corners mean reduced quality (in this case, precision)

but then again, what i read could have been skewed a bunch and i'm just not remembering.
i have a feeling that even if it is lower quality, it isn't noticable unless you make a direct comparison. even then, you can just bump up the bitrate a bit.


edit: yes, iirc, x87 is 80bit
 
(cf)Eclipse said:
it was a long time ago. i tried to find the article, but alas. for now just ignore what i said because i can't prove it, however it kinda makes sense. it's an optimization.
optimizations mean cutting corners
cutting corners mean reduced quality (in this case, precision)

but then again, what i read could have been skewed a bunch and i'm just not remembering.
i have a feeling that even if it is lower quality, it isn't noticable unless you make a direct comparison. even then, you can just bump up the bitrate a bit.


edit: yes, iirc, x87 is 80bit
Click to expand...

SSE = 128 bit.

==>Lazn
 
You must log in or register to reply here.
Back
Top