Core i7-3720QM: Ivy Bridge Makes Its Mark On Mobility

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
[citation][nom]frozentundra123456[/nom]Yea, I agree with greghome. These results make no sense at all. I thought multiple tests had shown that the HD3000 was considerably slow er than the A8, and the HD4000 only brought the performance to about equal levels to high end Llano.[/citation]

Actually the review stated somewhere in there that they were testing the machines all on battery. As a result, the A8's performance was considerably slower compared to while plugged in.
 
[citation][nom]carbonfountain[/nom]Actually the review stated somewhere in there that they were testing the machines all on battery. As a result, the A8's performance was considerably slower compared to while plugged in.[/citation]

The A8 showed so badly because it had a power savings profile active and the Intel systems all had high performance profiles active.
 
So, Mr. Ku, are you going to comment on the apparent disparity between your results with the A8-3500 and every other benchmark of it, including previous Tom's Hardware benchmarks? Is it possible you mistakenly (I presume not purposely) ran it with a battery savings profile active on the Llano notebook, which would obviously skew the results heavily?

I don't mean this with any sort of hostile intent. It's okay with me if you made an honest mistake, as long as the mistake is admitted to and corrected. And, if the power profile wasn't the issue, then an explanation for the unusual benchmark results is still needed.
 
[citation][nom]blazorthon[/nom]The A8 showed so badly because it had a power savings profile active and the Intel systems all had high performance profiles active.[/citation]

Actually, I think the Intel systems were on the 'balanced' power profile. The reviewer stated that the tests were performed at the default which infers power saving for the AMD and balanced for the Intel.

Although I'm sure that the AMD would still be beaten performance-wise, I'd like to see what the APU could do with a performance power profile. My guess is they'd be closer (especially in graphics) performance-wise and have a slimmer lead in power consumption.

The rubber will really meet the road when the i3's and i5's come out for the mobile IB segment. That would be a fairer test against the A8 in regards to price.
 
[citation][nom]Kyuuketsuki[/nom]Is it possible you mistakenly (I presume not purposely) ran it with a battery savings profile active on the Llano notebook, which would obviously skew the results heavily?[/citation]

He explicitly stated that tests were run under default battery profiles.
 
Seems like a lot of folks are only reading part of the story here.

1) Our original Llano review unit was tested plugged into the wall.
2) These systems were run at their default power profiles. We can't help that AMD lacks the intermediate setting that Intel ships at.
3) Rather than simply presenting the data for Balanced vs. Max Battery Life, we took the extra step to re-run tests with all profile permutations to show what you'll get for performance with Llano cranked all the way up.
 
[citation][nom]cangelini[/nom]Seems like a lot of folks are only reading part of the story here.1) Our original Llano review unit was tested plugged into the wall. 2) These systems were run at their default power profiles. We can't help that AMD lacks the intermediate setting that Intel ships at.3) Rather than simply presenting the data for Balanced vs. Max Battery Life, we took the extra step to re-run tests with all profile permutations to show what you'll get for performance with Llano cranked all the way up.[/citation]

Most people read it. But the majority of the tests were not done as such. It seemed intentionally misleading.
In a drag race the Camaro beat the Corvette. 5th page in, happens to mention that the Corvette came with a trailer and was therefore tested with the trailer attached.
 
Quick note for those of you in this thread that this pertains to:

Registering multiple usernames through fake email accounts for the purposes of steering discussion on the Tom's Hardware Community is dishonest, disingenuous, and very much against the Rules of Conduct. Users that do so will be permanently banned from commenting on Tom's Hardware.
 
[citation][nom]jpishgar[/nom]Quick note for those of you in this thread that this pertains to:Registering multiple usernames through fake email accounts for the purposes of steering discussion on the Tom's Hardware Community is dishonest, disingenuous, and very much against the Rules of Conduct. Users that do so will be permanently banned from commenting on Tom's Hardware.[/citation]
... until they create a new email account, unfortunately. This happens often enough that simply banning a single user name seems to be an inadequate deterrent against excessive trolling. I'm beginning to wonder if there are any other reasonable options?
 


You're right... My bad. The Intel's were at Balanced, not high performance. Regardless, the Llano APUs obviously lost this badly only because of the difference in power profiles.
 
[citation][nom]blazorthon[/nom]You're right... My bad. The Intel's were at Balanced, not high performance. Regardless, the Llano APUs obviously lost this badly only because of the difference in power profiles.[/citation]
Ah right, rereading the article he clearly stated he used the defaults and the default on the AMD system was battery saving. So while perhaps not a "mistake", it's definitely poor testing methodology to compare the CPUs when the AMD is on a battery saving profile that cripples performance while the Intels are on a "balanced" profile that obviously doesn't hobble them much, and contributes to a skewed perception of the performance differences.

Also, the only game in the section on the differences between power profiles is WoW. Even though WoW has always heavily favored anything that's not AMD, the Llano still does better than the HD3000 while from the rest of the article you'd think it can't hold its own even against that.
 
[citation][nom]Blink[/nom]Most people read it. But the majority of the tests were not done as such. It seemed intentionally misleading. In a drag race the Camaro beat the Corvette. 5th page in, happens to mention that the Corvette came with a trailer and was therefore tested with the trailer attached.[/citation]
That argument only works if, when you buy the Vette, Chevy gives you a trailer. In this case, we're testing two retail laptops the way they ship. This is *what you'll experience* unless you manually set the Max Performance profile.
 
[citation][nom]cangelini[/nom]That argument only works if, when you buy the Vette, Chevy gives you a trailer. In this case, we're testing two retail laptops the way they ship. This is *what you'll experience* unless you manually set the Max Performance profile.[/citation]

Given the fact that most of us are enthusiasts (at least to an extent), that's exactly what we would do. However, I can see the logic in this... Anyone who didn't know better would probably leave it at it's default power profile and this is representative of what they would get. However, most people here would not do that, so although I see the logic in bench-marking at default, I also think that it would have made sense to put some more emphasis on the high performance profile, or make a custom profile that is similar to the balanced profile found in the Intel machines (if possible, I know that I can make custom profiles on my machines, both Intel and AMD), hopefully getting an even more proper comparison.
 


Luckily, I have larger, more powerful alternatives than just expunging usernames based on email accounts. Though I think the other active, genuine participants in this thread will realize which accounts and posts are being created by shills for other sites and guerrilla marketing companies. The practice of bringing in shill accounts to direct the conversation and skew reviews/feedback has transcended hotel reviews on TripAdvisor. Caveat lector - let the reader beware.

That said, after reading about those spicy N36Vm's, I have something new to add to my Christmas list.
 
[citation][nom]jpishgar[/nom]Luckily, I have larger, more powerful alternatives than just expunging usernames based on email accounts.[/citation]
Oh good. :)
 
[citation][nom]alphaalphaalpha[/nom]Given the fact that most of us are enthusiasts (at least to an extent), that's exactly what we would do. However, I can see the logic in this... Anyone who didn't know better would probably leave it at it's default power profile and this is representative of what they would get. However, most people here would not do that, so although I see the logic in bench-marking at default, I also think that it would have made sense to put some more emphasis on the high performance profile, or make a custom profile that is similar to the balanced profile found in the Intel machines (if possible, I know that I can make custom profiles on my machines, both Intel and AMD), hopefully getting an even more proper comparison.[/citation]
That's a completely fair opinion for the folks who understand gaining 3x% performance might cost 4x% battery life. Default just happened to be the configuration we found most relevant, but of course wanted to include everything else as well. We'll definitely keep the feedback in mind moving forward, though! A definite "thank you" for weighing in.
Best,
Chris
 
[citation][nom]cangelini[/nom]Seems like a lot of folks are only reading part of the story here.1) Our original Llano review unit was tested plugged into the wall. 2) These systems were run at their default power profiles. We can't help that AMD lacks the intermediate setting that Intel ships at.3) Rather than simply presenting the data for Balanced vs. Max Battery Life, we took the extra step to re-run tests with all profile permutations to show what you'll get for performance with Llano cranked all the way up.[/citation]
... put in image quality tests too...
 
[citation][nom]DjEaZy[/nom]... put in image quality tests too...[/citation]
I saw Charlie's IQ comparison and already asked Don to schedule a comparison of HD Graphics 4000 and the AMD competition =) I think that'll be an interesting one!
 
Trinity has already launched! That TH deemed to release this review on the day that Trinity has launched shows to me either ignorance of what is going on in the industry or they really are Intel fanboys.. Tsk-Tsk
 
[citation][nom]The_Trutherizer[/nom]Trinity has already launched! That TH deemed to release this review on the day that Trinity has launched shows to me either ignorance of what is going on in the industry or they really are Intel fanboys.. Tsk-Tsk[/citation]
Our Trinity launch is upcoming. This story was yesterday--in fact, we've had a lot of Ivy Bridge-based coverage planned since launch that just hasn't gone live yet.
 
[citation][nom]hixbot[/nom]Is the Ivy overclocking (desktop) article still coming?[/citation]
Yes, should be edited for next week. Also, I've been working on an efficiency piece!
 
[citation][nom]cangelini[/nom]Yes, should be edited for next week. Also, I've been working on an efficiency piece![/citation]

I know that this is off-topic, but I have a question about the forums. Why is it that sometimes, when someone posts a comment on a forum thread that I have subscribed by email to (or news article or review), I get an email telling me that someone has commented in this forum (or news article or review), and sometimes I don't?
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom