Really?

Review Intel Sandy Bridge Quad-Core processors - Notebookcheck.net Reviews

2630QM is faster than 940XM in all benchmarks

 

I don't think I've seen this kind of performance leap between generations before. A lowest end SB outperformed the top of the line Clarksfield? That's crazy

 

lol, where have u been? thats why u see so many threads about the SB's glory. and thats why its such a debacle to choose between an r2 and an r3. but i digress.

 

Guess I'm hiding behind my RGB LED for way too long lol. I still doubt about real performance comparison tho. It's so hard to swallow this

 

And from what i heard, the 2920XM isn't that much faster than the 2630QM.

 

yea but all gens have low and high end of the premium line and its a great way to compare the lowend of the highend compared to the high end of the last gen. its like comparing the Q9800? to 720qm or whatever

side note chick in avatar is smokin ^^

 

^^ Alienware members get so used to her already lol but I guess she's new for so many that not come to AW subforum. She's Leah Dizon (can't count how many times I said this)

What is the line in between?

I've used every generation of Intel since Pentium III time I think. Don't know if I miss anything but here they are Pen III, Pen IV, Duo Core, Core 2 Duo, Core 2 Quad, Quad Cores and i7.

 

Since when is the 2630QM a low tier moddle?

Ok wow. Just looked up the 2630... how is this surprising?

Even if the 2630 weren't SB it's a quad core with multithreading with similar clock rates.

Sorry but what's the big deal? This is basically exactly what anyone would expect.

 

I'm going to follow your digression.

The Alienware M11x R3 hasn't been announced. We have no idea what specs / hardware it will have if/when it is released. The assumption that it will have Sandy Bridge is just that... an assumption.

It makes sense that it would have Sandy Bridge, but I would advise people against making purchasing decisions on what is currently still vaporware.

 

I'm pretty sure he was talking about the M17x R3 (which has practically been released), seeing as the M11x R3 wouldn't be expected to take a quad core...

 

i was, maybe i should have specified, its the m17x r3. from the techies i talked to, its a 40% increase than the 940XM.

 

Because it has smaller cache and lower clock rates and it's priced around 1/3 price of an Intel extreme.

This's not what I expected.

 

definitely not 40% increase. go to Notebook / Laptop Reviews and News - Notebookcheck.net and read their tests

 

Not with a 2630QM, no. But a 2720QM with good cooling will beat a stock 940XM by 40-50% or more in multi-threaded tasks.

 

Yes, but who buys a(n overpriced) 940XM to run it stock?

 

true but i am sure a lowend quad SB could be overclocked alot

 

You can't overclock Sandy Bridge CPU's through the traditional method of overclocking (increasing FSB). You can only get worthwhile overclocks Sandy Bridge CPU's by buying multiplier-unlocked Sandy Bridge CPU's. You can't just buy any old Intel Sandy Bridge CPU and expect an overclock from it (beyond the +400Mhz allowed by Intel).

 

can't be overclocked at all?

EDIT: oh i didn't know that...why did they block it?

 

Well, I don't know if "block" is the right word. It is more a side-effect of an architectural change.

It is a result of the base clock (the "master" clock for the system, upon which all other clock speeds are derived) being moved from the motherboard to the CPU. Intel did this with Sandy Bridge to integrate the different components that go into a computing platform. Overall, this reduces the cost of building a computer around an Intel CPU.

Because of this change, Intel has full control of the base clock (previously, 3rd party motherboard manufacturers had control of the base clock). And since they have control of the base clock, they also have control of all of the clock dividers that control PCI bus speeds, PCI-E bus speeds, memory clock speeds, etc. It is these other clocks (PCI, PCI-E, etc) that prevent you from really overclocking a Sandy Bridge system by changing the base clock.

Intel's response to this is to release relatively inexpensive K-series parts that are multiplier unlocked. You leave the base clock at 100Mhz, but overclock a CPU by changing the CPU clock multiplier. That is why you can't just buy any old Sandy Bridge CPu and expect to overclock the bejesus out of it... you need a K-series part to overclock the bejesus out of it.

Non-K-series parts do allow you to overclock, but only by +400MHz. A K-series part allows you overclock as high as your CPU will support... typically the 5.0Ghz+ range.

 

gotcha but 400 is still a good boost but 5GHz? wow.

 

Because the extreme is incredibly overpriced. The cache might be slightly smaller... but only the L3 cache which is insignificant in most cases.

As I said, this is exactly what you should expect in a processor jump.

As someone above mentioned, it's not like moving from P3 to P4, which was absolutely huge.

 

Im guessing the temps on these CPU's are within respective limits?

 

Actually, there's a hard limit of 5.7GHz.

However, this all only applies to desktops, unfortunately. With laptops, only the Extreme Editions are unlocked, and the rest can't be overclocked at all (not even 400MHz).

 

also going to a p3 to a p4 was vary large depending on which p3 and p4 you compare ^^ P4 was like the longest lasting naming for a processor ^^

 

True. And second gen p3 is definitely different from first gen p3. But I'm talking about the first gen p3 to the first gen p4 (which came out before second gen p3)

 

This is a wonderful deal for the performance then... You can get a new SB laptop with the i7-2630 (while they were still at newegg) for less than the i7-940xm. (Itself, no other parts.)
Newegg.com - Intel Core i7-940XM Extreme Edition Clarksfield 2.13GHz 8MB L3 Cache PGA988 55W Quad-Core Mobile Processor Model BY80607002526AE - Processors - Mobile

 

It's very interesting that this thread is becoming a lot more informative. Anyway, the SB move of Intel is what hardware developers should ever do rather than keep releasing "new model" that is only slightly better in order to rip us consumers off

 

they have other reasons than the last laugh, though.. the actual systems are very intelligent self-overclockers that balance themselfes out as needed, make sure they fit termal constraints, power constraints, and balance out the best usage for it. to provide all that functionality, they have to combine much more close together and integrate it in defined ways, so it works that fast and well.

for that 0.1% of users that maybe thought about overclocking, it's not worth caring about, at all. the new systems are selftweakers that optimize themselfes to the environment they're in. that is very important for all the other customers.

and for those that want high performances, get an unlocked cpu and cool it well. it'll overclock itself as good as it can.

 

how am i wrong? going from a PIII to a PIV can be a big or small jump depending on what ur comparing. Comparing a PIII 800 mhz card to a PIV 3.0 GHz E edition is huge. PIV had a long lasting naming system for all sorts of processors. the only ones i can remember off my head was the prescott and preslor but there were several others in that category too. There is nothing wrong in my statement.

@Dave also there is no good reason to block the overclocking. The 99.9% of population that doesn't care about overclocking wont touch it anyways. Only real reason to block it is to make the people who want to overclock pay extra for an unlocked one. Second the only people that would want to block it would be manufatories of a computer. Reason being if it breaks form overclocking you can make them foot the bill, which gives them the reason to block it. Hence most laptops have limited bios which annoys the crap out of me

 

There actually is a good reason... it makes the cost of building a system cheaper.

Integrating things like the GPU and base clock onto the CPU package means that a motherboard manufacturer doesn't need to buy those separate components and build them on to a motherboard. It makes sense from Intel's perspective. The only real advantage that AMD has is price / performance, so anything Intel can do to drive down price is a win for them.

And the effects of this are relatively minor, when you look at Intel's customer base. The overwhelming majority of CPU's that they sell get sold to large system OEMs like Dell, HP, Lenovo, etc. A small portion get sold as loose parts to enthusiast system builders, and an even smaller portion get sold to system builders that overclock. And for the overclockers that are affected by this, Intel released very affordable K-series parts that allow massive overclocking.

Don't get me wrong - I don't like the fact that I have to buy a specific overclocking part / model # from Intel, because of a change they made. All I'm saying is that I understand why they did it, and am also willing to put up with that inconvenience for the performance benefits of Sandy Bridge.

 

I thought the K series and regular ones don't really have an real difference? Because i thought the 720qm 820qm and 920xm are all the same but they only let the multipliers be adjusted on the 920xm to get all the overclockes/ people who want the most performance possible to get the 920xm because overlocking a 720qm could reach stock 920xm fairly easily. If there is a physical limitation i don't know about i understand but it seems to me they did this to milk the market.

 

(1) The K-series parts and non-K-series parts are identical, except the K-series parts are multiplier-unlocked. You will only be able to get meaningful overclocks out of Sandy Bridge parts by buying a multiplier-unlocked K-series parts.

Core i7-2600K
- 3.4Ghz core clock at stock speeds
- 4 physical CPU cores w/Hyperthreading (8 virtual cores)
- 8MB of L3 (Level 3) cache
- Multiplier unlocked. Max multiplier of 57x (5.7Ghz)
- $317 (vs. $296 for Core i7-2600)

Core i5-2500K
- 3.3Ghz core clock at stock speeds.
- 4 physical CPU cores. No Hyperthreading.
- 6MB of L3 (Level 3) cache
- Multiplier unlocked. Max multiplier of 57x (5.7Ghz)
- $216 (vs $205 for Core i5-2500)



(2) The limitation of only allowing K-series parts to overclock is because the base clock was moved from the motherboard into the CPU package itself.

The fact is that taking the base clock off of the motherboard and integrating it into the CPU package lowers the overall cost of building a computer. It is also a fact that it gives total control to Intel in regards to who can overclock, and how that overclock can happen.Whether you interpret the motive behind those facts as "milk[ing] the market" is up to you.

But personally, I think it isn't to "milk the market". First of all, the requirement of K-series part only affects overclockers (a ridiculously tiny speck out of the people who will eventually buy Intel Sandy Bridge parts). Cost savings affects everyone, and helps to remove the only advantage that AMD has vs. Intel. Second, if they wanted to "milk the market", they'd charge a lot more than $11 / $23 to buy a K-series part.

 

i see what ur saying but it still seems like milking to me..if there is no pyshical difference but whatever

 

i agree i think its stupid...and shady