Sandy Bridge Processors Gaming Performance

Just read through the mini review on notebookcheck and thought I'd share it.

It appears that Sandy Bridge dual cores work as well as quads on a selection of games with min-end GPU (GT540m used). To clarify, the reviewers found little or no significant difference between a core i3-2310m and a core i7-2720qm at 1600x900 & medium details.

This is great to hear for the casual gamer, I'm sure some games may have more significant differences (SC2 & DA:O for example), but it's exciting that thin & light laptops with dual core and mid-graphics should have enough CPU power regardless of CPU (i3, i5, i7).

Review-Intel-Sandy-Bridge-Processors-Gaming-Performance

 

Overall Sandybridge is an awesome CPU. I built a whole new desktop Sandybridge desktop with an Core I7 2600k. They run really cool temp wise and overclocks like crazy.

 

Yea that helps alot because I was thinking about getting a new laptop that have the sandy processor in it.

 

Sounds nice an all, but when can we see quantum processors !

 

Id still feel better having a quad than a dual core lol

 

This has possibly been the most useful thing I've seen to help me deciding on what laptop to buy. Going for one with that graphics card (I think even the same base) and have been trying to decide for ages which processor is right for me.

 

Nothing has changed in terms of what has always been. As long as a game is GPU dependent then whether you are using a Dual or a Quad will not really matter. As soon as you play a CPU dependent title, that is where you will notice the difference. Try playing GTA IV with one of those Dual Cores and then compare it with a Quad. You will never get the same level of performance with the Dual as you will with the Quad.

 

cool cool thanks for sharing!

 

This is business as usual. All that shows is that the GT 540M is bottlenecking the more powerful CPUs.

Run this same testing with the GTX 485M, and you'll see the lower end Core i3 and 5 throttling its max framerates.

 

I'd like to know how the AMD 6770m and 6850m would fare with the new i5-2410m and i7-2630qm.

HP DV7 quad's base CPU is 2630qm with a 6770m GPU, while the Envy 17 has a faster 6850m but the base CPU is only a 2410m. The 2630qm is a $200 option on the Envy, which puts the price a bit higher than I want to pay. I'd rather have the 6850 GPU, but I don't know if I want to give up a blu-ray player and a quad core CPU for it.

 

Sandy Bridge isn't such a great deal. Its an improvement over Nehelam but its not anywhere near the scale of difference when i series replaced core 2 series. I'm pretty happy with my i7-740QM. Also taxmantoo, i wouldn't get the envy 17. Its thin, light and powerful but runs quite hot when i looked at it when buying my G73. Its a safer bet to get something other than envy. That was months ago but not sure about now.

 

For laptops, it's the other way around. The Clarksfield replacements of Core 2 were barely better than the mobile Core 2 Quads (their one and only virtue was Turbo Boost) and this improvement came at the cost of battery life. Sandy Bridge is much better than Clarksfield and uses less battery than even Core 2.

Regarding the original post: tests where the component of interest is being bottlenecked by some other component are pretty worthless unless the sample is close to being exhaustive -- and 4 games is nowhere near that.

 

aside that its never worth to change the laptop each year, 2 years is the minium for a good performance boost

 

^ GF100 (2010) -> GF104 {2011) -> 28nm (2012) bucks that trend.

The base Clarksfield vs. base Sandy Bridge (i7-740QM vs. i7-2630QM) is a very substantial leap.

 

I would say large. Clock by clock on base clock 2630 is abouts 0.2GHz faster than 740qm and has a 0.1GHz faster turbo boost.. which is probably more efficient but i doubt its that much faster.

 

The 2630 can outperform the i7-940XM, take a look here for benchmarks and you will see. Or the summary quote in that thread from the testers.

 

Checkmate.

Sandy Bridge is the truth, which is why the 2630QM is as high as a gamer needs to go. 2720QM is a waste, and 282oQM is just ridiculously priced. I won't even speak of the XM.

 

In addition to the above, you're only looking at part of the turbo boost spec. The max TB stated is for one core. When all four cores are active, the 740QM maxes out at 1.96GHz, while the 2630QM tops out at 2.6GHz (though it usually stabilizes at 2.4GHz). Withe two cores, the 740QM can reach 2.53GHz and the 2630QM at 2.8GHz. Then there's the 5-15% clock-for-clock performance increase.

 

Unless you do CPU intensive games, dual core sandy bridge looks enough.

Review Intel Sandy Bridge Processors Gaming Performance - Notebookcheck.net Reviews

 

Those are the same exact tests, which are linked in the OP.

Also, they've been invalidated, because the GT 540M is bottlenecking the CPUs. You'd need to use a GTX 485M, to get accurate readings.

 

Game performance has never been a good foundation to compare or question cpu performance.

There are so many random factors.

Firstly the most important is that in many if not most instances the testing will be turtlenecked by the gpu's performance making any cpu testing irrelevant.

Testing the same cpu with differing gpu drivers or different gpu models can also change results when the cpu can be exactly the same between systems.

The next big problem is that games are not all using the cpu in the same way, some may only be single threaded, others maybe multi threaded. A single threaded game would show better results with a higher clocked dual core than a slightly slower clocked quad core even if the quad core is a much more capable cpu.

Even in most multi-core games most often its a single core of the cpu under heavy load and the other cores just get some "secondary" threads that are not as intensive. Its better than nothing but not too many games can make full use of all cores on a cpu. More and more games are moving in that direction though as best as possible.

The way I see it a quad core these days can fill any dual core task (even more so with turbo boost) but if you are playing one of the few game titles that are extremely cpu demanding and use all 4+ cores on a quad core, the dual core cpu can not fill that role and the games performance will suffer.

So get a quad core cpu, and get one with the best speed possible for the money.

 

I still can't believe Notebookcheck thought that such a test would be valid...

 

My point was for casual gamers it's good news, the GTX cards are in the enthusiast range, hence not casual gamer teritory. So although there's only a few games, at medium details on 1600x900 it is still sufficient.

When going 1080p and GTX 460m and above, I imagine the difference would start to show in certain games which are thought to be CPU-bound, however that is a different topic.

So to summarise, will it play Crysis? Yes, any mainstream SB CPU will play Crysis easily. To what extent it will play Crysis in high details is a different question for hardcore PC gamers