Q9000 VS Core i7 720QM

Hi guys

I'm sure I'm not the first person who has asked this but I still need more convincing. Which one do you think is better? I do mostly video editing, encoding, photo editing stuff. How much would the performance gain from 720?

The question extends to M15x or M17x, any opinions and suggestions are also welcome.

Thanks in advance.

 

For video editing, encoding, photo editing the i7's are the best. HERE is one of many reviews to prove it.

 

I read the article on Bit-Tech, and I agree that for things like photo editing the i7 is the better choice but when it came to crysis, both the i7 and the q9000 got pretty much the same fps (and we can say that the 280m and mobility 4850 are very close in terms of performance) does that mean at least when it comes to gaming it does not matter what the cpu is because the limiting factor is the gpu?

If that is the case it doesn't matter if your laptop is equipped with a q9000 or a 720qm if you want to game with it.

Is this a fair assumption?

 

In my opinion it's a fair assumption, and if they're equal in games and the i7 720QM wins in everything else the i7 wins. I myself went with the i7 720QM as I believe it's superior overall.

 

Thats not quite a fare comparison, a 2.8ghz quadcore was faster than a 2ghz quadcore how shocking.

 

It's not a 2.8 GHz quad core lol. That's only in dual core mode. In quad core mode the actual clock frequency is less than the Q9000.

 

link to z info?

 

http://www.google.com use it.

 

I did and nothing seems to show what you are claiming and you seem to be avoiding providing the proof. Not even reviews on this very website of i7 cpu's say the cpu runs at lower clocks while all 4 cores are being used. Does it overclock when only two cores are needed yes....I have yet to read anything that says the cpu only runs at 1.6ghz when all 4 cores are being used.

 

http://www.intel.com/products/processor/corei7/mobile/specifications.htm, read the white paper or technical manual. Though I doubt you'll bother seeing how you were too lazy to even do a proper search on google.

 

lol 10char

 

For a possibly easier way of looking at things, try this page http://en.wikipedia.org/wiki/List_of_Intel_Core_i7_microprocessors

The frequency is the base speed. The turbo is how many bins (jumps of 133 MHz) that the processor can self ramp up on 4 cores/3 cores/2 cores/1 core, assuming that heat and power tolerances are not exceeded. Thus the i7-720QM can, assuming perfect cooling and infinite power, run at 1.73 (1600 MHz + 133 MHz) GHz on 4 and 3 cores, 2.4 (1600 MHz + (133 * 6) MHz) GHz on 2 cores, and 2.8 (1600 MHz + (133 * 9) MHz) GHz on one core. Note that this does not include hyperthreading, which means each core could run 2 threads (virtually) simultaneously, so in theory, you could run the 4 threads of a Q9000 (which lacks hyperthreading, so only 1 thread per core at a time) on 2 cores of an i7-720QM. Apparently, non-server Windows 7 prefers to use physical cores over logical cores, however, so unless you deliberately set it so, you'll probably end up with just the 4 physical cores at 1.6/1.73 GHz.

This is, of course, ignoring third party overclocking, just using Intel's own Turbo-Boost self-overclocking.

 

I found this out a few days from The_Moo. It really disappointed me about the i7's. But, I guess that just means the i7's aren't for me.

 

Why? They don't perform equally per cycle. The i7's are quite a bit more efficient than the previous generation quads. In just about everything, the i7 will win even with a lower clock speed.

 

Yeah, from what i understand, i7 is easily efficient enough at the lower clock speeds, and dont forget the hyperthread extention.

I was on the m15x the other day with lower config + 1080/260m

And its definately no slouch when comparing it to my machines hardware, the i7 and q9 perform toe to toe with the i7 even feeling slightly faster.

 

The number I hear quoted most often is about 20% more efficient for the i7s vs Core 2, so that 1.6 GHZ should effectively be about 1.92 GHz. At that point, the from a pure clock speed perspective, it's probably pretty close to a dead heat, as in, you wouldn't notice the difference, and if you go beyond 4 threads, the i7 pulls ahead thanks to hyperthreading. Really, the advantages of i7 over C2 lie mostly in its adaptability. Need to save power? Shut down cores and operate at low speeds on one or two cores only. Need to hyper-power that single thread to completion? Overclock one core to 2.8 GHz and finish that task ASAP and then throttle back down to save power. Need to do super-multithreaded processing? Run 8 threads at a time on 4 cores and finish the task (almost) twice as fast as the Q9000.

Although, admittedly, the truth of the matter is that most of us will not be doing much of anything that would push the limits of a Q9000 or an i7-720QM, even in gaming. We'll usually hit the GPU limit first long before we stress out the CPU.

 

I just prefer raw power. I'd much rather over clock a CPU to get a higher GHz rating rather than shutting down cores to increase the GHz rating of fewer cores.

But, I do see the efficiency factor come into play and I can't argue that i7's are great at that. Don't get me wrong I don't hate the i7's I'm just not all that impressed right now. Once they start making better ones I'm sure I'll be more open to the idea.

 

If you really want raw power in a notebook format, go with the desktop i7 (and probably soon to be available i9) in the Clevo D900F. That's starting to wander a bit far afield, though.

More on-topic to this discussion, supposedly there are third party software utilities to manually overclock the i7 in the M15x. I don't know what they are, though, and you might need an i7-920XM with its unlocked multiplier...

 

I'm sure the i7's will get better and better as the whole idea behind them is brilliant. Just wanna see where it leads.

 

I'm still learning about this whole i7 CPU so I need to do more research and learn what threading is, but any idea if the i7 will be available for the M17X? I know its been covered just didn't find it. Also definately will be interesting to see how it develops further.

 

I don't think the i7's will be available for the m17x. The m17x was finished before the the i7's were released so they currently don't support them. I don't see why they wouldn't come out for the next m17 model though.

 

You're probably thinking about this thread (Caution! Very long!)
http://forum.notebookreview.com/showthread.php?t=420148

I still figure the main limit for an i7 in a M17x sized machine is the heat issue, especially with the 2 cards in SLI. From the numbers people list for the M15x and the 2 i7 Sagers, the i7s idle in the 50-60 degrees Celsius range. Couple this with the fact that the 4870s ended up as GDDR3 instead of GDDR5 due to heat issues, and I'm not sure there's enough overhead thermally for an i7 in a M17x right now.

 

It might be do to heat...That would explain why I haven't seen any laptops that have the i7's with any SLI or Crossfire configurations. So I think you're right it is likely a heat issue. I suppose we'll have to wait for future installments before we see them with SLI or Crossfire.

 

I strongly recommend to have a look here.
This is one of the most extensive reviews out there. It's long and boring till you get into benchmarks, but when you get there take a closer look at Clevo w870cu v.s. Eurocom M98nu (same hardware as M17X ). Clevo has the top i7 920XM+GTX280M v.s. M98NU QX9300(@2.5GHz) + GTX280M. They also run benches with M98NU SLI.
As you can see from those i7 920XM outperforms QX9300 on most of them by a narrow margin. And since QX9300 can be easily OC'ed to 3.06-3.2GHz and gain a HUGE boost in performance the margin will almost entirely disappear (i7 can be OC'ed too of course but alas with no real boost in performance).
And to verify the strength of my assumption those of you with QX9300 can run those benches @3GHz and post the results. I'm sure we all gonna be surprised

 

Wow. Nice find dude. That's an informative read.

Rep+

 

Thanks, been looking for it for weeks, as I wasn't convinced by Intel's marketing moves regarding i7s.

 

It is due to a different motherboard requirement to accommodate both rather than heat IIRC - You are going to be seeing i7 with SLI early next year.

 

Marketing aside, my Core i7 820QM appears equivalent to my QX6850 extreme quad core desktop processor OC'd to 3.2Ghz, in most games.

How you ask?

Well most games are single threaded still, and so the 820QM is running at 3.06Ghz. Then add in its clock-for-clock advantage, and integrated memory controller and DDR3 memory... and you've got a winner.

My desktop uses a Geforce 9800GT, which is slightly better than the 260M in this alienware, but games like L4D2 and Dragon Age Origins show improved framerates when run on the same monitor at 1920x1200 resolution.

I am impressed by the sheer power this laptop posseses, and how cool and silent it is at idle (due to the idling of the cores).

Just my thoughts. Perhaps in time we'll be given more control over the i7 OCing.

 

What is your FPS when in Denerim (Market area) and Inside the Cultists Ruins (snow area) with 820QM and 1900rez? Really curious about those 2 locations.
Also is your FPS stay steady in GTA4? what about Empire - Total War? I'm asking because seen a lot of negative feedback on those from owners of mobile i7 systems when playing all these on max rez+set.

 

Look at the CPU limited game benchmarks, not the ones where the GPU is the bottleneck:

1. Assassin's Creed:
Clevo: 60.54 fps
Eurocom M98: 47.3 fps

2. Call of Duty: World at War
Clevo: 55.62
Eurocom: 45.39

3. Crysis @ Medium settings:
Clevo: 58.71 fps
Eurocom: 48.78 fps

You can't take a general look at gaming benchmarks and conclude the QX9300 is equivalent to a 920XM or even close because it's not. The majority of the game benchmarks will be GPU limited at a high resolution. When the CPU does come into play, you see noticeable differences like I noted above (Jarred also mentions CPU limited games in the article). In each of the CPU limited situations, you're looking at between 17-21% difference in FPS between the two CPU's. If you OC'd the QX9300 10-15% it still wouldn't catch up to a stock clocked 920XM. That's still not even looking at multimedia performance which the new i7's excel at. Finally, reliable overclocking tools haven't been released for i7 CPU's but if/when they are, that gap will increase even more.

 

On the other hand, tests like these always fail to test the T9900/X9100 processors. As has been said many games are single thread/dual thread orientated. These processors at 3.06Ghz, would theoretically offer much more comparable performance in these situations, than than the 2.53Ghz QX9300 providing you weren't running intensive background tasks or programmes like AV software etc, even more so if you overclocked the two chips.

Its annoying they leave these processors out of the test, as although they offer us an insight into C2Q vs I7, they dont even consider how a higher clocked dual core would compare, or even include them in scenarios where they might have an edge, nor does it mean they cover the rather large segment of market still using dual cores.

Its a real shame as they could answer many questions! If these games are single thread orientated CPU limited, then theoretically a dual core at 3.06Ghz would be much closer in performance to the I7 which would be turbo-boosted to 2.8Ghz or similar. Throw in slightly higher IPC, and then you end up in the interesting scenario that the dual core, higher clocked chips may fare about equal; although obviously both the I7 and C2Q would be better for 4 core enabled multimedia/encoding applications.

Just a shame as if they'd included the X9100 or T9800 then they could have answered a lot of questions.

 

Do you know the performance difference in games when a C2D/Q is OC'ed .5Ghz? I see a 15-30 FPS jump with P8700 virtually in ALL games I play (about 10).
Now the examples you bring are nice but don't forget that those benefit most from a highly clocked CPU. So of course 3.2 GhzX 2 Threads is gonna be better than 2.5GhzX2. Now if you update your game to fully utilize 4 cores ( Like Empire-Total War patch) you will see that your i7 will improve by 30-50% (it would be 60% for desktop i7s) while your C2Q jumps to 100% boost in FPS.
Not to mention that QX9300 is meant to be OC'ed to above 3GHZ. Mandrake got almost 5K in 3Dmark with M17X @3.2GHz and I can only wonder what will be the result with M17 (older one) after OC'ing to 3.7GHz.
Now even at 3.2 stable run those video encoding tests and see the difference.
I'm willing to do any test and compare the OC'ed 920XM to QX9300(when my chip arrives before Christmas), for despite the beautiful definitions about new architecture/features I feel a huge disappointment with Intel's last update. Maybe it's just my own subjective impression, but I think that the i-core (mobile) is just a cheap reallocation of CPU potential. It's not a leap forward but rather a slight optimization.

 

I totally agree with you here, that's why I propose to make an extensive comparison of high-end C2D,C2Q and i7 CPUs both stock and OC'ed. It would be nice to see how T9900/X9100 perform @3.6GHz versus 920XM.
But since the first chip I'm getting is QX9300, that is gonna be a start. I will run all those tests OC'ed with a single GTX280M and we will compare them to Clevo W870CU results. Those who own T9900/X9100 can run the same tests and we shall see the full picture.

 

Aye, I mean stock 3.06Ghz they're already fast (not noticed much difference between it and my old [email protected]), and hell if the XTU software and mandrake got a Q9300 to 3.7, I wonder what the X9100 could do..4Ghz in a laptop would be insane...and probably dance all over the mobile I7 in dual/single threaded apps.

I mean this X9100 I've got could easily do 3.33Ghz with just a multiplier change, if I had access to RAM dividers, timings and voltages etc, then I cant see why it couldnt go past that easily, the M980NU I've got now has much better cooling than my old M17 (10 degrees or so cooler load, and about 4-5 degrees cooler at idle) so its not like the potential isn't there if only Clevo would release that unlocked bios they promised.

 

You get a 15-30 FPS jump using which game, setting, established baseline fps etc? That's a very generic and general statement you made. I could say, "well I play quake 1 and i normally average 220 fps but with my OC'd cpu I hit 270 fps" which is entirely meaningless. Better yet, show me a credible review where any Core 2 was OC'd by 500 mhz to yield a drastic improvement in gaming performance. I have two Core 2 based systems at home, one laptop and one desktop. The desktop doesn't show large gains when OC'd, even in CPU limited games. But as I said, it depends on what game, settings etc. you are referring to.

You're assuming all Q9300's OC the same which they do not. How many people are hitting Mandrake's CPU OC on their M17/M17X? Are these extreme OC's even stable?

In fact, this entire discussion of a therotically OC'd CPU vs a stock one is apples to oranges. You can do a proper comparison once i7 OC tools are available. Right now you're basing your argument on a stock CPU vs an OC'd one which has no guaranteed OC baseline.

The stated gains are about 20% and if you look at CPU limited games, that is roughly what is seen. The i7 series for notebooks are new, they do not have any OC tools yet so you can't make a fair comparison of the two using one at stock and the other overclocked.

 

1. You are right, I need to bring more details. all tested with P8700 CPU.
Empire - Total War (1900rez Max set) - Lowest 7FPS(unplayable) @2.53GHz, Lowest 25 FPS @3.06GHz.
Americas Army 3 (1900rez Max set.) - 17-45FPS @2.53GHz, 28-70FPS @3.06GHz.
Dragon Age - Origins (1900rez Max set) - 20-60FPS @2.53GHz, 30-90FPS @3.06GHz.
Crysis - (1024rez, all High, no AA) -25-45FPS @2.53GHz, 30-60FPS @3.06GHz.
These are the once that were recorded with stock drivers.

2. Read M17X owners SIGs. The lowest OC for QX9300 is 2.8GHz, while 80% hit 2.9-3GHz STABLE.

3. 920XM systems do not offer a BIOS OC anymore (Clevo) simply because of some stability and performance issues. People were hitting 4GHz and not getting any performance difference. And not without a reason.

4. Now about baselines.One system has a built-in OC capabilities that make you enjoy your extreme CPU in full measure. Another system doesn't have those but has a different CPU with built-in turbo boost feature (According to Intel it was meant to provide a dynamic OC-ing and remove the need to manually control the CPU frequency.
Would it be completely unfair to compare 2 different systems at their best? Even if we cut to CPU netto, would it be unfair to compare the two the way they exist today? Or would we wait till MAYBE some day the tools will be there to OC 920XM more efficiently and compare then?

 

I am tempted to get the qx9300, but all the tests I have seen (max detail/1920*1200/AA) show the same scores as my rig at stock speeds (because we are pushing the video cards hard). The big exception is GTA IV.

So I guess its all about balance.

Dragon Age: 1920*1200, Graphics Detail Very High, AA 4x, Texture Detail High, Frame Buffer Effects On



NFS: Shift: 1920*1200, 4x MSAA, All Other Settings at Maximum Quality



Resident Evil 5: DX10 Mode, 1920*1200, AA 4x, Motion Blur On, Shadow Detail High, Texture Detail High, Overall Quality High

 

I have need of babyhemi's help, I would get through to you, how do you can?

 

Did anybody post my thread on this here? I dont think so so take a look its 100% related to the OP's question.

http://forum.notebookreview.com/showthread.php?t=457919

In short the Q9000 wins for now because it can be overclocked. Even with the advances the i7 has made over the core2quad series cpu's it cant make up for several ghz of processing power between a stock 1.6ghz i7 and a overlclocked 2.6+ghz C2Q

Every Asus W90 unit did 2.7ghz on the Q9000 pretty much and I think the M17X was a bit lower like 2.5 to 2.6ghz right?

Nothing else matters here, turbo boost doesnt mean jack, hyperthreading doesnt help enough, the C2Q wins.

My best time for Wprime is ~15.5 seconds with the 720QM my best time with the Q9000 is under 14 seconds.

1.5 seconds is pretty large in that benchmark. If we were talking about encoding a 2 hour movie with x264 that may be the difference in waiting an extra 20 minutes or so for it to get done.

 

Excellent debate, this is why I come back to this site.