Is a Celeron M any good...read on....

Status
Not open for further replies.

Superczar

Posts: 9   +0
Me & my friends did a Celeron M vs Pentium M vs AMD Sempron vs Pentium IV comparo sometime back.....

Some of you may think this sounds like a weird comparo with all these processors targeted at different segments altogether...but the idea was to compare a Celeron M and Pentium M's performance with similar desktop proccy's

I hope that this will come in useful for a lot of other people too:
If someone wants me write a more detailed report, I will do that ...As of now, I am not even sure if this might come in useful for anyone so don't want to go thru a huge effort for nothing...

Also, Plz Forgive the typos :
( BTW all prices mentioned are in Indian Rupees ...1 US $~ 44 INR)
------------------------------


Test configs:
1-My Compaq V2010 with Pentium-M 725 (1.6 Gig) with 512 MB ram, 4200 RPM 40 GB HDD and Intel Extreme Graphics (1250 $ ~ 55K) and wide-screen ( 1280X768 ) display

2- Friedn's Compaq V2035 with Celeron -M 1.5 ( Ithink 350) (52K Chennai price + 2.7K for extra RAM), 512 MB RAM (2X256) 4200 RPM 40 GB HDD and Intel Extreme Graphics (Rupees-52K)and wide-screen ( 1280X768 ) display

3- Friend's Toshiba S-331 (don't rem the xact no) with P-IV 3.06, 256 RAM, 40 GB HDD, on board gfx Radeon 7000 IGP (think he paid 61K for this 1 month ago)

4- My desktop with AMD Sempron 2400, FX5200, 80 GB 7200 HDD, 256 MB RAM

I have posted the benchmark results that we ran from what remains inmy memory..(but I do have all the results (exact figures) on an xls somewhere, so if someone is interested, I will take the pain of hunting for it :)))

a) Performance: We ran a lot of benchmarks + also tried gaming with Unreal Tourney and UT 2004 @ 800X600

The benchmark whose results I remember (And would be relevant for many of you as this benchmark taxes purely the CPU ) is Super Pi (calculates the value of Pi to a pre-defined number of digits)


Super Pi at 2 Million digits (all non-essential apps shut down)
1- P-M 725- 2 min 20seconds
2- Celeron M Laptop - 2 min 37 seconds
3 - P-IV laptop - 2 min (but it got appallingly hot while running the test)
4- Sempron Desktop: 2 Min 34 seconds

Now in % terms, that's not really much of a difference b/w the P-M and C-M, is it?

Unreal Tourney (Original): Don't rememeber the FPS count but the gameplay was extremely smooth on all the systems (easily <40fps on all)

UT 2004:

Desktop- Very playable with avg fps over 35 fps
My laptop: Decent Playability with fps in the range of 25-30 but with all effects turned off
Celeron Laptop - Surprisingly, it yielded the same fps nearly all the time (probably the bottleneck was the GFX subsystem ) (again with all effects turned off)
Toshiba Laptop: FPS~30, but game kept crashing ...also the laptop was consistently hot to the touch....

Interpretation: On board GFX on current gen laptops are fine enuff for casual gamers, but for someone who wnats to play the latest and greatest stuff, A GFX card is a must...
More importantly, A laptop will never match the performance of a desktop (though we didn't have a laptop with a good GFX card, but other factors like HDD speed etc will drag the sysstem down even on one with a decent GFXcard...)



B) Battery life: Config 1 gives an average battery life of 3-3.5 hours with average load (surfing + word processing etc), so does config 2, config 3 has a poor battery life ~1-1.5 hr

1,2 have a 4.4 AH battery...not sure abouth the battery capacity of the Toshiba Model

C) HDD speed: Tested using HDtune for burst Data access rate
-All the laptops gave an Avg 24-25 MBps (blame it on the 4200 RPM HDDs)
- The Desktop gave ~40 MBPS :)



D) weight : 1 and 2are 2.5 kgs, config 3 is 3.1 kg




FInal Thoughts:

1- The miniscule performance difference between a Celeron M and Pentium M doesn't justify the big price diffference.....
2- A P-IV may be faster than a M series (Dothan or Banias core ) processorbut the high heat dissipation and battery drainage makes it a bad choice for a laptop...Howver, the superior (??) architecture of the M series ensures that even at a ~1/2 clock speed of a P-IV, the performance difference is miniscule
 
Status
Not open for further replies.
Back