Well...I'm disappointed. A core i7 story.

So i bought a core I7 2720qm ES on ebay a few months ago and was kind of disappointed by the performances barely above a core i7 2630qm.

A few days ago, someone almost gave me a core i7 2720qm OEM. So i decided to try it in my Vostro 3750. Guess what ? Exactly the same performances than the ES version : 4,90-5 in cinebench 11.5 64 bits and 4950-5000 pts as cpu score in 3dmark06.

Sandy Bridge seems to be a bit tricky. I guess it depends more on how your bios will manage the cpu than the cpu itself.

Temps are low for both, it doesn't throttle at all. 75°C max. A bit lower for the ES actually. Exactly the same framerate for HD 3000 on BBC2. Can't see a real difference between both. Except than i can force the OEM to run forever at 2,6ghz with throttlestop while ES stays at 2,5ghz no matter what. But the ES has 8mb cache instead of 6mb. So...it's pretty much the same.

Well, will sell the OEM at a good price i guess.

Anyone else disappointed by his core i7 perfs compared to benchies found on the web ?

 

It really comes down to what you do with your computer, you won't notice much a difference in most situations except CPU-intensive programs like video encoding/3d rendering, even you are encoding a video or rendering a complex 3d scene, the difference between a 2630qm and 2720qm won't be staggering as the latter is only about 12-13% faster.

For most day to day uses, a faster CPU won't make a difference, a higher clocked quad like 2720 or 2820 may benefit a few games a little, but usually they won't provide a tangible performance gain over the lower quad.

 

there are no big differences between those cpu's, as all the numbers (including the actual name of the cpu) tells you. what do you expect? if MAX difference is 10-13% or something, then that's the MAX difference. not the average, or typical, or anything.

 

Oh i wasn't expecting much. It's more than enough for my use anyway. My point is just that i'm pretty far from the core i7 2720qm numbers found here :

Mobile Processors - Benchmarklist - Notebookcheck.net Tech

Not a big deal but still.

And i saw many people telling that ES cpus are underperforming...Well, it's not the case for me. It's more bios related in my case. OEM or not.

 

Uhm... just how much did you expect from a 10% speed difference?
Because, the 2720qm IS only 200MhZ faster per core than 2630qm.

Benchmarks on notebookcheck should be taken with a grain of salt... use them as a 'rough estimate' that give you a general idea but by no means an accurate measuring tool (and even then you have to remain skeptical).

Even in CPU intensive programs such as 3dsMax, the difference will be barely visible unless we are talking about renders that would usually take say an hour to render on 2630qm... and 2720qm would realistically take 54mins instead.

10% is minimal... it saves you minimal amounts of time.
20% is better and produces a better effect, but still not 'drastic'.
30% is much better and produces a more noticeable effect.
40/50/60% is of course a lot better.

Of course... we are talking about CPU intensive tasks, and predominantly for things where timing is essential - but even so, 10-20% is not justifiable... and even in games that are more cpu bound, you wouldn't see that much difference between a SB i7 quad of 2.0Ghz and 2.4Ghz.

Perhaps with 2.6Ghz and above...

But seriously, Intel is pricing their CPU's with increasing stupidity.
Anything above 2630qm and 2760qm has minor instructions that people won't find useful and performance wise, we are talking about 10-15% differential as you jump from one cpu to the next, whereas the prices jump by 100% or double - which is a pointless/unjustified price gauging (even their lower clocked quads are way overpriced for what they are worth) .

 

Hi Deks Owner of the most powerfull Acer 5930G ever upgraded.

I know high end cpus are pretty useless considering the extra price, that's why i never bought them and will never do. I took a core i7 2720qm ES cause back then it was very cheap on Ebay. Cheaper than a used core i7 2630qm.

I'm just saying that somehow it's not normal that people with a core i7 2630qm get better scores than others with core i7 2720qm considering the extra price (For OEMs cpus i mean). I mean when you buy a desktop core i5 2500K, it performs like a core i5 2500K. Sandy bridge mobile performance seems a bit like a a lottery. Guess mobile segment is more adaptable and it's up to manufacturers to decide how the bios will push cpu turbo according to heat concerns...Again i don't really care since the perfs are already amazing. Just an observation.


Worst case is manufacturers that sell expensive quad i7 notebooks that throttles like hell and performs like garbage. Big scam.

 

Exactly.
They hype up this technology like it's something revolutionary... and it's at least 6 decades behind of what we could be using.
Anyway...

Benchmarks as I said can be misleading... so be sure to check for signs of Throttling and if your OS if fully up to date.
But having said that, the numbers themselves will be really close to each other... in some cases they might not even reflect 10% differential properly.
you need to do some real-life testing using cpu intensive programs and compare results from cpu to cpu.


As for my Acer... lol... I don't think I'm done upgrading it yet.
I'm waiting on some copper shims to arrive so I can put them on the cpu/gpu/northbridge in order to further drop the temps, and some time after that I will likely put in an SSD into the laptop's primary bay while putting my hitachi into the CADDY to act as a second storage drive.
As for the gpu... I'm thinking about upgrading that too, but finding a good candidate for an affordable price-tag is not easy.
:-D

 

And the final thing is that the cpu is a ES model. Its not the final product. ES chips are known to be everywhere in performance.