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

Page 2 - 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.
Something just does not seem right in your review. While I dont argue that Ivy Bridge is certainly the new king of mobile and there is no other competition it seems like your testing of the AMD APU is inaccurate.

I noticed in most of the benchmarks that the AMD APU always exhibited really low power consumption usually around 20 watts which is a good 15 watts below its thermal envelope. Plus your game benchmarks are giving the APU very low frame-rates compared to what I have seen in multiple other reviews. There is absolutely no way Intel IGP 3000 if roughly equal to the A8 APU in graphics performance like your charts indicate.

It appears you tested the AMD with the wrong power profile compared to the Intel solutions.
 
[citation][nom]cknobman[/nom]Something just does not seem right in your review. While I dont argue that Ivy Bridge is certainly the new king of mobile and there is no other competition it seems like your testing of the AMD APU is inaccurate.I noticed in most of the benchmarks that the AMD APU always exhibited really low power consumption usually around 20 watts which is a good 15 watts below its thermal envelope. Plus your game benchmarks are giving the APU very low frame-rates compared to what I have seen in multiple other reviews. There is absolutely no way Intel IGP 3000 if roughly equal to the A8 APU in graphics performance like your charts indicate.It appears you tested the AMD with the wrong power profile compared to the Intel solutions.[/citation]

I think that the A8 had power savers while the Intels had high performance. At least, that's what this page of this review seems to imply:

http://www.tomshardware.com/reviews/core-i7-3720qm-ivy-bridge-mobile-ultrabook,3185-8.html

Dropping Intel's performance by 50% and their power usage by 40% to 50% would bring them more in line with Llano's power usage and would leave Llano as much more competitive. Unfortunately, Llano does not seem to do as well as the Intel chips do with a high performance profile, so they seem more comparable when they use their power saver profile. Maybe some tweaking could alleviate some of the problems. Either way, Llano is obviously more competitive than let on by this review.

They should have either made this much more clear in each page, or if they were feeling particularly generous, maybe even do it over with the power profiles and power usage during each benchmark kept in mind at each step in the review. As it is now, it does not seem too useful to gauge and compare Llano, Sandy, and Ivy Bridge because of this. Also, with only i7s included from Sandy and Ivy, we don't even know how well Llano stacks up against Intel's competition that's actually in it's own budget range. This review just has a very unfinished feel to it. Come on Tom's, you can do better than this!
 
The real question with IB processors is whether Intel will go back to the proven solution for heat dissapation and replace the heat paste it is currently using with the solder. The IB is a great mobile solution, but it may be relegated to being a niche product if Intel decides to stay with its current design over the fluxless solder (which is 20X more efficient at heat dissipation). Unless rectified the SB will predominate with gamers over the IB platform, and Intel's new chip will be a moderate failure. The IB can be super, but only if Intel corrects its design mistake.
 
[citation][nom]BenchmarkGenius[/nom]People keep talking about the GPU performance but that is only 5% of the story. I know everyone loves the GPUs but they do little actual work outside of the gaming community (which doesn't use integrated graphics). The real story there was the IvyBridge CPU performing sometimes more than 5x better than APUs and up to 2x better than Sandy. The Adobe test was staggering. Did anyone notice Ivy was 2x better in WinRAR? Holy crap! The quicksync performance is awesome. I am definitely going to get a system with Ivy although I want the 3820 (call me greedy, but an extra 2 MB of cache and another CPU bin is the what the doctor ordered). I will wait for Haswell on the desktop side, but will definitely be upgrading to Ivy on the mobile side.[/citation]

I recommend that you wait and see how Trinity comes around. By then, it's VCE might be up and running and then we will see if it can stack up against Intel and Nvidia's hardware encoding. Heck, for all we know, it could be even faster than HD 4000 Quick-Sync (maybe, maybe not, but we don't know much about it, let alone how it will perform, so who's to say that it won't be king, even if only for a little while?). Besides, we already knew that the i7s would beat low end APUs in CPU performance. That's just common techy sense.

If you didn't already think that the i7s were several times faster than the APUs in CPU performance even before reading this, I'm kinda surprised.

Also, you can bet that a lot of people will be gaming on these IGPs. Their gaming performance is more than just 5% of the picture here. With Trinity and Haswell supposed to both bring great leaps in IGP performance yet again, IGP gaming is probably going to actually become a fairly common thing, at least compared to its past.
 
the bad thing about owning a first gen Core i7 series mobile processor was you didn't have an intergrated IGP to save power when you didn't need to use something heavy on graphics. My Radeon 5730 Mobile kills my battery life and since i'm stuck with a 6 Cell battery because Asus never made a 9 cell battery compatible with my laptop the best i can get out of my laptop is an 1 hour and 55 minutes surfing the net anything gaming or movies and it just sucks the battery life right out of it.
 
Intel igp is still the low end can't even match last years amd 6550D igp
it does not couple with a gpu
Trinity will have no competition
As far as 4core cpu is concerned amd & intel cpu are all more than capable though bragging rights go to Intl
and the winner for best all in one solution AMD
 
Great Review Andrew!

Could you provide Idle power numbers for the systems used? I'm curious to see if there is an improvement between the Ivy and Sandy systems when you remove the display and and drive as variables. Thanks.
 
[citation][nom]triny[/nom]Intel igp is still the low end can't even match last years amd 6550D igpit does not couple with a gpu Trinity will have no competitionAs far as 4core cpu is concerned amd & intel cpu are all more than capable though bragging rights go to Intland the winner for best all in one solution AMD[/citation]

HD 4000 doesn't need to beat the 6550D in gaming performance, it just needed to beat the 6620G and that it does, by a considerable margin too.
 
What in blazes are you talking about Triny? Show me the 6550D benchmark please. The coupling with the GPU thing didn't work so well in the first gen AMD CPUs although that is a nice feature (that only works with AMD cards of course... can't couple with nVidia).
-Trinity will have no competition? Oh noes, a fanboi.
-4core CPUs are all more than capable? Well, I use VisualStudio 10 on mine. If one platform compiles 4x faster than another, guess what, one is considerably more capable than the other. If one is doing photoshop at 5 to 6x the performance of another, then one platform is completely more capable than another. Somehow doint real world tasks gets minimized by you as "more than capble", but you put this enormous premium on doing a few 3D content things faster on AMD as somehow making it a better all around solution?

You are either 12 or no one is paying you to use your computer. If someone is actually giving you money to use your computer, then you want the tasks you do to be done faster. Now cheaper I can see. Saying somehow that a small bump in 3D performance somehow makes one platform the best all around solution though is just plain stupid.
 

These are the high-end units that will be in more robust laptops. I'm pretty sure the smaller ultrabooks will be running LV i3s and maybe i5s
 
Something is definitely fishy in this test. How is the 6620G no threat to the HD 3000 when tomsharware's own review of the A8-3500M clearly shows that the Radeon HD 6620G soundly thrashes SB's HD 3000 in pretty much every mobile scenario:
http://www.tomshardware.com/reviews/a8-3500m-llano-apu,2959.html
So, what exactly has changed here? Every site has the HD 3000 being trounced by the HD 6620G in the mobile space. Something is rotten in Denmark. 🙁
 
Lol.

First Intel pays off Anandtech to split off the horrendous overclock performance and stock thermals into a separate article.

Now Intel pays off Tom's Hardware to completely fudge tests comparing their $400 mobile processor to a mobile Llano in graphics tests.
 
lol now I understand why when I use APU graphics on my laptop, nothing works as it should. 6620 on APU my ass! Thanks god for dedicated 6650m. Naming suggests only marginal difference, but in reality it's at least twice as fast, when i switch to 6650m. Oh, and no stuttering. only thing I can use apu-6620 graphics is CrossFireX, it helps, although not much
 
I agree that the AMD were most likely set to default in which means "max battery life".

Overall I think that skews the results by much. Intel 3000 graphics are not as good as the a good portion of AMD graphics.
 
Benchmarking with MAX Battery profile on AMD? This for real on TH?! Waste of time article thx. Benchmark both on MAX Bat and then MAX Performance. Who cares if AMD doesn't have balanced?! Use the LCD's!
 
We know that upcoming Trinity-based APUs will include the company's VCE capability, but because that feature isn't even enabled on the Radeon HD 7000-series add-in cards, we don't know how it'll match up.
Click to expand...
Is this true? Why would you add VCE to desktop gpu die design, if it's never ever enabled. For FirePro series? Trinity will use cpu and gpu on the same die, so reusing those "standalone" chips is out of the question.
 
[citation][nom]SuperVeloce[/nom]lol now I understand why when I use APU graphics on my laptop, nothing works as it should. 6620 on APU my ass! Thanks god for dedicated 6650m. Naming suggests only marginal difference, but in reality it's at least twice as fast, when i switch to 6650m. Oh, and no stuttering. only thing I can use apu-6620 graphics is CrossFireX, it helps, although not much[/citation]

Hmm..I'd be curious to see your evidence on either of those 😉
 
[citation][nom]Communism[/nom]Lol.First Intel pays off Anandtech to split off the horrendous overclock performance and stock thermals into a separate article.Now Intel pays off Tom's Hardware to completely fudge tests comparing their $400 mobile processor to a mobile Llano in graphics tests.[/citation]

I'd be curious to see your evidence for that 😉
 
Hi Toms editor,

Thanks for the review. However, is it possible to do a separate comparison to an overclocked A8? Llano APUs are well known for having significant overclocking headroom, and it's generally very easy to elevate the stock 1.6 Ghz clock to somewhere between 2.6 to 3.0 Ghz. Doing this will give us overclockers a better sense of where the Llano chip falls in comparison to Intel's chips, but also let us known when is a good time to upgrade.

Thanks!
 
[citation][nom]blazorthon[/nom]Actually, that clearly shows that the 6620G of the A8s is only marginally better than the HD 3000 and that with the 6630M, the A8s are then closer to twice as fast (some of the time) as the HD3000-using equipped Sandy systems. Keep in mind that the mobile versions of Intel's IGPs are similar to the desktop versions, but the mobile Llano IGPs are much slower than the desktop versions, so on the mobile side, they clash much more, instead of Llano wiping the floor with Intel's IGPs. Trinity will almost certainly let AMD take the lead in mobile graphics IGPs again. Until then, AMD always has the ability to do CF with the IGP and still use similar amounts or even less power than Intel while beating Intel for graphics performance, although Llano clearly can't touch Sandy and Ivy in CPU performance.[/citation]


Are you sure about that? Looking at the charts on the link you provided, it seems that the A8-3500M with 6620G tends to have anywhere from 25-75% higher FPS than the HD 3000
 
[citation][nom]carbonfountain[/nom]Are you sure about that? Looking at the charts on the link you provided, it seems that the A8-3500M with 6620G tends to have anywhere from 25-75% higher FPS than the HD 3000[/citation]

I already corrected that.
 
4 core APU to 3.0ghz? I hear 2,3-2,5 is max. for everyday use. But I guess it's more of a cooling design problem, if anything. So it could be different between two laptop manufacturers and even between cpus of the same model. I wouldn't let it above 80° celsius for long though.
 
[citation][nom]SuperVeloce[/nom]Is this true? Why would you add VCE to desktop gpu die design, if it's never ever enabled. For FirePro series? Trinity will use cpu and gpu on the same die, so reusing those "standalone" chips is out of the question.[/citation]

VCE is going more or less everywhere because it is a part of the Radeon 7000 series (like Blu-Ray playback with Radeon 6000 cards, even the 6770 and 6750, which are re-brands of the older 5770 and 5750, were upgraded to include features of the Radeon 6000 series, such as the native Blu-Ray playback capability). Also, Quick-Sync is pretty much everywhere at this point, so why not let AMD have VCE everywhere that they can too?
 
Yea, I agree with greghome. These results make no sense at all. I thought multiple tests had shown that the HD3000 was considerably slower than the A8, and the HD4000 only brought the performance to about equal levels to high end Llano.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom