Two different articles comparing Pentium M & Turion

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
Hello all,

I don't frequent over to this neck of the woods often, but I used to when I tinkered with my desktop ;) Now I'm into laptops and I'm a senior editor at the website in my sig. We recently released a pretty in-depth article on the Pentium M vs Turion 64, analyzing the architectures and running benchmarks. And coincidentally, THG released a very similar article just after we did. Obviously we are nowhere near in size to THG, but frankly I wasn't impressed by their article. We've had a couple readers comment in the forums that the THG article was better and I just don't see it.

I was hoping to get some impartial feedback (since I know I'm not :p) from some intelligent people, because frankly I think the people who commented on our forums are the sort of people who take THG as their end-all/be-all existence.

Whenever you all have some free time, I would appreciate you taking a look at both articles completely; it is a doozie ;) But look at not only the tests & results, but the testing methods, analysis, etc etc. I just wasn't impressed with THG and I need to know if its just me :(

THG Part I
THG Part II

LaptopLogic Article

LaptopLogic discussion thread with THG comments

Also note that some users have questioned the battery results of our Acer 8104 tests: consequently we have another 8104 in-house and are re-testing it with the Winstone Batterymark and our new Bapco MobileMark (or whatever its called). We will be presenting updated results with the old & new batterytest platforms soon.

Thanks in advance! :)
 

Pabster

Lifer
Apr 15, 2001
16,986
1
0
Originally posted by: fbrdphreak
Also note that some users have questioned the battery results of our Acer 8104 tests: consequently we have another 8104 in-house and are re-testing it with the Winstone Batterymark and our new Bapco MobileMark (or whatever its called). We will be presenting updated results with the old & new batterytest platforms soon.Thanks in advance! :)

I'll be interested in those battery test results. Your review was the first one where a Turion-powered lappy lasted longer than a Dothan-powered one assuming equals.

Overall I thought Laptop Logic's review was good. Tom is a constant source of debate around these parts :D
 
Nov 11, 2004
10,855
0
0
How about you give me the 8104? :D

Anyhow, personally I would like to see 100% full load battery tests as that's where it counts for me.
But I prefer your site there to Tom's.. Tom's is.. questionable in their methods and writers..
 

n7

Elite Member
Jan 4, 2004
21,281
4
81
Very nice review.

BTW, everyone knows THG is full of BS, which would explain their results, & also explain why you have idiots jumping onto to your forum claiming their testing was better.
Also, they were comparing an 1860MHz CPU to an 1800 MHz one...looks like a few other things weren't fair either...
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
Thanks Kensai :) I will pass on your idea of full load battery tests (gaming I assume?). I was actually contemplating that earlier, I agree that it would probably be valuable to a lot of people.
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
Originally posted by: n7
Very nice review.

BTW, everyone knows THG is full of BS, which would explain their results.

Also, they were comparing an 1860MHz CPU to an 1800 MHz one...who know what else wasn't fair...
I noticed that too. 2.0GHz vs 2.0GHz is the only thing that made sense to us when comparing two different architectures that way :confused:
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
On the topic of full-load battery testing: what suggestions do you all have? Loop a game benchmark? 3DM05? Real-world gaming with a stop watch until the battery dies?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,099
16,014
136
Originally posted by: fbrdphreak
On the topic of full-load battery testing: what suggestions do you all have? Loop a game benchmark? 3DM05? Real-world gaming with a stop watch until the battery dies?

I would try 2, F@H is cpu only, and a game would be cpu+gpu.
 

Furen

Golden Member
Oct 21, 2004
1,567
0
0
The only things that THG did better than you guys were to use DDR333 as well (since most, if not all, Turions come with that to begin with) and to include the Sempron (35w) as well. Their power consumption table was very nice but I'm guessing they just calculated that from the Total power/time and didnt measure it directly. Lastly, they used DDR333 on their system, which makes it "match" the Turion system a bit better.

They did some very stupid things like say that Doom 3 ran faster on the Dothan because of its 2mb cache (I mean, it could run slightly faster because of it but 2x faster tells me that something is wrong with the AMD system). They also didnt match the speeds of the video subsystems (the Centrino's video ram is running about 15% faster, I believe) and the Intel CPU is clocked 60MHz faster than the Turion. They didnt explain how their testing scenarios for battery life worked ("Office Applications" doesnt explain if its just continual CPU load using office applications or if it's like your "life" scenario that mimics real-world use) but (overall) I think their battery life tests show the same trend as yours with the exception of gaming: basically at Full load, which I'm guessing "Office Applications" is, the Centrino system has better power consumption while at lower load, playing a DVD wont load the system too much, the Turions end up being better. Gaming is a bit of an exemption but this could be due to the lower ram frequency, for all I know.

EDIT: Now I just have to join the LaptopLogic forums to flame... err... discuss my point of view (in a rational and tolerant manner) with these guys.
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
Originally posted by: Furen
The only things that THG did better than you guys were to use DDR333 as well (since most, if not all, Turions come with that to begin with) and to include the Sempron (35w) as well. Their power consumption table was very nice but I'm guessing they just calculated that from the Total power/time and didnt measure it directly. Lastly, they used DDR333 on their system, which makes it "match" the Turion system a bit better.

They did some very stupid things like say that Doom 3 ran faster on the Dothan because of its 2mb cache (I mean, it could run slightly faster because of it but 2x faster tells me that something is wrong with the AMD system). They also didnt match the speeds of the video subsystems (the Centrino's video ram is running about 15% faster, I believe) and the Intel CPU is clocked 60MHz faster than the Turion. They didnt explain how their testing scenarios for battery life worked ("Office Applications" doesnt explain if its just continual CPU load using office applications or if it's like your "life" scenario that mimics real-world use) but (overall) I think their battery life tests show the same trend as yours with the exception of gaming: basically at Full load, which I'm guessing "Office Applications" is, the Centrino system has better power consumption while at lower load, playing a DVD wont load the system too much, the Turions end up being better. Gaming is a bit of an exemption but this could be due to the lower ram frequency, for all I know.

EDIT: Now I just have to join the LaptopLogic forums to flame... err... discuss my point of view (in a rational and tolerant manner) with these guys.
Thanks a lot for the comments!

There is value to using DDR333 in comparison (I dunno how they found a Sonoma laptop that uses DDR333, those are few & far between these days LOL), but we used DDR400 in order to truly evaluate the platform & architecture. If we were comparing the NOTEBOOKS themselves, then definitely we would use the stock memory. But if Sonoma is designed to use dual channel DDR2 533 and Turion is designed to use lower latency DDR400, then so be it :)

I wish we could've included Sempron as well. I think we're going to do a "budget shootout" or something and do a similar article comparing Sempron & Celeron M.

I think they did OC the slower GPU to match them up, but they didn't match drivers. Drivers between different models, let alone different manufacturers, vary a LOT in notebooks. Hell, in our Ferrari 4005 & 8104 (same company, Acer), the Ferrari had a newer driver (and higher clocks). We used the Ferrari driver & clocks for both notebooks.
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
Originally posted by: Markfw900
Originally posted by: fbrdphreak
On the topic of full-load battery testing: what suggestions do you all have? Loop a game benchmark? 3DM05? Real-world gaming with a stop watch until the battery dies?

I would try 2, F@H is cpu only, and a game would be cpu+gpu.
What about something like looping the Prime95 torture test for CPU?
 

Lonyo

Lifer
Aug 10, 2002
21,938
6
81
Yay for THG including the Sempron based machines.
No offense, but THG is better in the fact that it has more processors included, which for me is VERY useful( trying to find the best deal for a laptop for the gf, and I want AMD, but didn't know what Semprons would be like, now I do know).
Haven't finished readong THG yet tho, just about to start #2 article.
 

Lonyo

Lifer
Aug 10, 2002
21,938
6
81
/me thinks the Doom 3 results are BS.
UT2004 is fairly CPU limited, especially at 1024x768, ytet all of the processors were fairly close, yet in Doom 3 the Centrino PC gets double the fps of the other systems?
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
I too wish we were able to include Sempron. I guess I'm more looking at these two articles are evaluations of the platforms themselves, Turion & Pentium M. We plan on doing a budget comparison of Celeron M & Sempron systems, but this is more of an analysis of the architectures & platforms than a "product review."
 

Furen

Golden Member
Oct 21, 2004
1,567
0
0
Originally posted by: fbrdphreak
I too wish we were able to include Sempron. I guess I'm more looking at these two articles are evaluations of the platforms themselves, Turion & Pentium M. We plan on doing a budget comparison of Celeron M & Sempron systems, but this is more of an analysis of the architectures & platforms than a "product review."

Yeah. I, personally, like your review better because it seems better structured and meant to put everyone's "the system specs are not the same" claims to rest and finally figure out which is better (which depends on what your priorities are, fancy that). THG's review is also one of the most boring reads I've done in a while. The sempron does come out great in THG's review, but the price difference between it the other systems is not that great.
 

linkgoron

Platinum Member
Mar 9, 2005
2,598
1,238
136
I don't get it, in THG's review the Centrino LT had a higher clocked x700, or am I wrong?
 

fbrdphreak

Lifer
Apr 17, 2004
17,555
1
0
Originally posted by: linkgoron
I don't get it, in THG's review the Centrino LT had a higher clocked x700, or am I wrong?
The clocks were different but I believed they OCed the slower chip.
 

Hacp

Lifer
Jun 8, 2005
13,923
2
81
As for testing, I think doing a DVD battery life test with WI fi on, downloading something is probably best for showing battery life. Also, you should use DDR333 for AMD turions as that is the memory that most people use for those laptops, and DDR2 533 memory for the Pentium M laptops. Everything else should be the same or very similar.

Personally, I don't care about performance benchmarks. As long as it doesn't feel slow, I"m ok. The benchmarks I'm really wondering about is battery life.
 

Pabster

Lifer
Apr 15, 2001
16,986
1
0
Originally posted by: Hacp
Personally, I don't care about performance benchmarks. As long as it doesn't feel slow, I"m ok. The benchmarks I'm really wondering about is battery life.

:thumbsup:

I suspect a high percentage of notebook users would gladly sacrifice some performance for longer battery life.

 

Concillian

Diamond Member
May 26, 2004
3,751
8
81
I think the following statements really put the two reviews into perspective:

Laptoplogic:
However, with both benchmarks, the FPS is already so high that a fps higher will not make a difference in your gaming experience.
A reasonable statment many review sites will never bring up.


THG:
Obviously, the only thing that matters is the largest possible L2 cache.
The graph under that statement shows the Sempron (256k cache) performing nearly identically to the Turions with 1MB cache, clearly indicating the importance of the cache size. :confused:
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Excellent article...
I would have enjoyed a 64bit software comparison of Turion to itself in 32bit as well..
Also, as you point out that the monetary competitor is actually a 1.6 GHz PM, it would have been nice to see a comparison there as well...

As for the THG "article"...it's THG! Seriously, most people who read and understand the hardware sites don't take THG seriously...