Core i7 720QM vs Core i7 820QM - Benchmarks

Overview

The purpose of this thread is to provide actual data with which to help a prospective-buyer make the decision between a Core i7 720QM
and Core i7 820QM. One some systems, this upgrade can range between $200-$400, and as such, it is my hopes that this post will help
consumers make an informed decision.

Test Systems

The two systems used in this test were very similar systems. The first, being a Clevo W860CU, and the second being a Clevo W870CU.
Both systems had idential RAM and GPU's. The HD were slightly different, the W870CU having a 320GB 7200RPM drive, while the W860CU had
a 500GB 5400RPM drive. Obviously the CPU's were different, the W860CU having the Core i7 720QM, while the W870CU had the Core i7
820QM.

General specifications for the laptops can be found below.

Clevo W870CU - Specifications
Clevo W860CU - Specifications


Testing Methodology
All benchmark utilities were downloaded and installed from their respective official websites, and were run with all defaults taken.
Where relevant, I've tried to mention the settings used, but if you have any questions as to the settings used, please download the
benchmark and install it to view the default settings before asking.


Benchmarks

The following benchmarks were run, links to the results for each test.

1. WPrime

2. Fritz Chess Benchmark

3. 3DMark Vantage

4. 3DMark 06

5. BarsWF (CUDA)

6. SuperPi 2M




WPrime
*lower is better*

Core i7 720QM = 32M - 15.911 sec - 1024M 494.534 sec
Core i7 820QM = 32M - 12.903 sec - 1024M 419.56 sec

Back to Benchmark index

Fritz Chess Benchmark

Core i7 720QM = 12.12 / 5818 kN/s
Core i7 820QM = 15.23 / 7308 kN/s

Back to Benchmark index

3DMark Vantage
1280x1024 - Preset: Performance

Core i7 720QM = P6572 - CPU - 26095 - GPU - 5260
Core i7 820QM = P6582 - CPU - 28663 - GPU - 5237

Back to Benchmark index


3DMark '06
1280x1024

Core i7 720QM = 11,901
Core i7 820QM = 12,409

Back to Benchmark index


Bars_WF

Core i7 720QM = 551.92 MH/s - CPU ~ 16 MH/s per core/thread
Core i7 820QM = 565.28 MH/s - CPU ~ 19 MH/s per core/thread

Back to Benchmark index


SuperPi
( **Thanks to trvelbug for providing the720QM score.)

*lower is better*
Core i7 720QM (2M) = 35 seconds
Core i7 820QM (2M) = 30 seconds

Back to Benchmark index



W870CU Specifications

Display

* 17.3" (16:9) FHD LED (1920x1080) with Super Clear Glare Type Screen

Processor & Chipset

* Intel® Core™ Mobile i7, Intel® Core™2 Quad Mobile Processors
* Chipset: Intel® PM55 Express Chipset

Storage & Drives

* 2 Detachable 2.5" 9.5mm (H) SATA Hard Disk Drives
* Raid 0/1 Supported
* 1 12.7mm (H) Optical Drive Bay, ATA Interface,
Interchangeable with DVD±R/RW Combo drive, Blu-ray/DVD±R/RW Combo Drive

Memory

* Supports Dual Channel DDRIII SDRAM
* Two 204Pin SODIMM sockets
* Expandable up to 8GB DDRIII 1333MHz

Video Controller

* nVIDIA® GeForce™ GTX 280M with 1024MB DDR3 Video Memory
* PCI-Express™ 16X
* Microsoft® DirectX® 10 Compatible
* HDCP supported
* Dual-View capable Supports two different applications open at one time; one on the External Monitor, and one on the Laptop

Screen. Multi-tasking has never been more convenient.

Audio & Multimedia Features

* Built-in High Definition Sound System
* 3D stereo enhanced Sound system
* S/PDIF Digital output 7.1CH
* 1 Built-in Microphone
* 4 Built-in Speakers + 1 Subwoofer
* Sound Blaster compatible

Network / Communication

* Built-in 56K MDC modem with V.90 & V.92 compliant
* Built-in Gigabit Ethernet LAN
* Bluetooth™ V2.1 + EDR module
* Intel® Wi-Fi Link 5300AGN 802.11a/b/g/n wireless LAN

Keyboard / Pointing Device

* Full Sized Keyboard with Numeric Keypad
* Windows Hot keys
* Integrated with Hot Keys for LCD Brightness, Suspend, Panel/CRT Display
* Touch Sensor Hot Keys for E-Mail, Web Browser, and Mute
* Integrated Touchpad with Scrolling function


Input / Output Ports

* 1 DVI output Port
* 1 HDMI output Port
* 4 USB 2.0 Ports
* 1 Headphone Jack
* 1 Microphone Jack
* 1 Line-In Jack
* 1 S/PDIF output Jack
* 1 RJ-45 LAN (10/100/1000Mbps)
* 1 RJ-11 Modem
* 1 IEEE-1394a Fire Wire
* 1 eSATA Port

Slots

* 1 Express Card 34 / 54 Slot
* 7-in-1 Card Reader (MMC/RSMMC/MS/MS Pro/MS Duo/SD/Mini-SD)
* 1 MiniCard Slot for WLAN

Power System

* 1 Removable11.1V smart Li-Polymer battery pack, 3800mAH
* Full Range 120W AC-in 100~240V, 50~60Hz, DC output: 20V, 6.0A AC Adapter

Cooling

* Arctic Silver 5 thermal compound standard
* Copper CPU Heatsinks

Size & Weight

* 8.35 LBs with Battery Pack
* 16.25" (w) x 11" (d) x 1.69" ~ 2.25" (h)

Additional Features

* Kensington® Lock
* Built-in 2.0M Digital Video Camera
Back to Benchmark index



W860CU Specifications


Display

* 15.6" (16:9) HD+ LED (1600x900) with Super Clear Glare Type Screen
* 15.6" (16:9) FHD LED (1920x1080) with Super Clear Glare Type Screen

Processor & Chipset

* Intel® Core™ Mobile i7, Intel® Core™2 Quad Mobile Processors
* Chipset: Intel® PM55 Express Chipset

Storage & Drives

* 1 Detachable 2.5" 9.5mm (H) SATA Hard Disk Drives
* 1 12.7mm (H) Optical Drive Bay, ATA Interface,
Interchangeable with DVD±R/RW Combo drive, Blu-ray/DVD±R/RW Combo Drive

Memory

* Supports Dual Channel DDRIII SDRAM
* Two 204Pin SODIMM sockets
* Expandable up to 8GB DDRIII 1333MHz

Video Controller

* nVIDIA® GeForce™ GTX 280M with 1024MB DDR3 Video Memory
* PCI-Express™ 16X
* Microsoft® DirectX® 10 Compatible
* HDCP supported
* Dual-View capable Supports two different applications open at one time; one on the External Monitor, and one on the Laptop

Screen. Multi-tasking has never been more convenient.

Audio & Multimedia Features

* Built-in High Definition Sound System
* 3D stereo enhanced Sound system
* S/PDIF Digital output 7.1CH
* 1 Built-in Microphone
* 2 Built-in Speakers
* Sound Blaster compatible

Network / Communication

* Built-in 56K MDC modem with V.90 & V.92 compliant
* Built-in Gigabit Ethernet LAN
* Bluetooth™ V2.1 + EDR module
* Intel® Wi-Fi Link 5300AGN 802.11a/b/g/n wireless LAN

Keyboard / Pointing Device

* Full Sized Keyboard with Numeric Keypad
* Windows Hot keys
* Integrated with Hot Keys for LCD Brightness, Suspend, Panel/CRT Display
* Touch Sensor Hot Keys for E-Mail, Web Browser, and Mute
* Integrated Touchpad with Scrolling function


Input / Output Ports

* 1 DVI output Port
* 1 HDMI output Port
* 4 USB 2.0 Ports
* 1 Headphone Jack
* 1 Microphone Jack
* 1 Line-In Jack
* 1 S/PDIF output Jack
* 1 RJ-45 LAN (10/100/1000Mbps)
* 1 RJ-11 Modem
* 1 IEEE-1394a Fire Wire
* 1 eSATA Port

Slots

* 1 Express Card 34 / 54 Slot
* 7-in-1 Card Reader (MMC/RSMMC/MS/MS Pro/MS Duo/SD/Mini-SD)
* 1 MiniCard Slot for WLAN

Power System

* 1 Removable11.1V smart Li-Polymer battery pack, 3800mAH
* Full Range 120W AC-in 100~240V, 50~60Hz, DC output: 20V, 6.0A AC Adapter

Cooling

* Arctic Silver 5 thermal compound standard
* Copper CPU Heatsinks

Size & Weight

* 7.38 LBs with Battery Pack
* 14.75" (w) x 10" (d) x 1.65" ~ 2.0" (h)

Additional Features

* Kensington® Lock
* Built-in 2.0M Digital Video Camera


Back to Benchmark index

 

Thanks for this!

Will be using it to base my purchase decisions next year for sure.

 

You're very welcome! It seems to be a common question people have, if they should upgrade to the 820QM or stick with the 720QM. Obviously everyone is different but at least we have some numbers to work with now!

 

Correct me if I'm wrong, but essentially, computationally-intensive tasks will see a big boost, while games will see a much smaller performance increase from upgrading to the 820QM?

Seems like the aggressive turbo modes on Clarksfield didn't manage to fix the "good but not downright awesome for gaming due to Nehalem's limited L2 cache" problem first observed in last year's Bloomfield CPU's.

 

That's about right, unless of course your game is multi-threaded and is heavily-CPU dependent. The only thing like that that comes to mind would be GTA IV.

 

Almost forgot...

+rep for an awesome comparison

 

Thanks man!

 

I was expecting the Core i7's to perform better. In terms of per core performance, they are not that great compared to the Core 2 duo's.

Core i7 720QM = 12.12 / 5818 kN/s (taken from your test)
Core i7 820QM = 15.23 / 7308 kN/s (taken from your test)
C2Duo T9900 = 9.15 / 4393 kN/s (my laptop)


+ Rep for a nice thread.


--

 

Strange how the 720 has a higher gpu score than the 820 , could be a fluke.

3DMark Vantage
1280x1024 - Preset: Performance

Core i7 720QM = P6572 - CPU - 26095 - GPU - 5260
Core i7 820QM = P6582 - CPU - 28663 - GPU - 5237

 

I noticed that as well.. I haven't run enough sample tests to determine how much the CPU/GPU scores fluctuate by, even on the same system, but seeing as how they're the exact same card, I think you're right.. just a fluke.

 

It's almost certainly a fluke - the difference is <1% so it lies well within any reasonable margin of error.

 

Supreme Commander is another.

Excellent work in this thread.

I'm with you on this one.

 

It seems the difference between them is hardly worth the $200~ price premium it costs.

While the extra cache would be nice, the 720 is hardly settling, considering the performance gains you get from it.

The only question is how will the 720 compare with the fastest arrandale.

 

Depends on what you're doing! But that's what the results are there for, to help people make that call. I think most people would probably agree with you.

It also depends on the upgrade cost. I know for Clevo/Sager it's only a $200 upgrade, but for other manufacturers it can be as high as $400.

 

If you build a barebones from RJtech and buy the CPU seperately it's $370~ on ebay for the 820, which i'm still considering. But by the time I buy my new laptop the Arrandales will have come out...

Recent issues have screwed over my intended laptop purchase

 

Excellent comparison, well-addresses the many threads out there full of prospective buyers forking over $400 for the 820QM. Repped.

 

Thanks MS, I appreciate it.

 

i think youll see a fairly big gain in realworld applications that do a lot of number crunching like video editing and transcoding.

 

You should remove any GPU benchmarks from that list, OP. Like in 3DMark Vantage, uncheck the Use PPU option, and in bars_wf, use the SSE2 test, not the CUDA test.

 

Finally got the 2M 720QM SuperPi score (thanks trvelbug!!) added.

Thanks jenesuispasbavard.. I might consider that, if I can rope trvelbug or someone else who has regular access to a W86C0 w/ 720QM into adding some more tests, or re-running BarsWF SSE2.

 

Good job!
Now if someone with 920XM can run these benches the picture will be complete

 

So we can prove how much of a waste the XM is? , considering they locked it out on Clevo's I have a hard time believing people are going to drop the extra money on something for the overclocking ability that's blocked by the manufacturer

 

I plan on making an extensive comparison between mobile i7 and C2Q CPUs. But will probably focus on QX9300 versus 920XM.

 

Well, keep in mind that you're comparing the lowest end i7 to just about the highest end C2D. Compared to a standard Q9000 as shown in the previous Fritz thread ( http://forum.notebookreview.com/showthread.php?t=421863), the 720QM beats a non-overclocked Q9000 by a little (200-800, need more numbers), and that's at a 1.6 GHz clockspeed vs a 2 GHz clockspeed.

Hm. I wonder. Supposedly, Fritz loads the CPU to 100%... I wonder how the scores change if you disable hyperthreading?

 

I will be interested in this comparison, see how much Nehalem helps the mobile chips. I can't live without it in my desktop.

 

That's actually quite good. You can see the top two results that the i7 quad is about 33% faster per clock compared to the Core 2 quad(Q9650 3.33GHz vs i7 975 3.33GHz).

The lower clock is holding back the mobile Core i7's. 3.06GHz dual vs a 1.6-1.73GHz quad and the latter is still 32% faster. It's a well threaded program, but clock speed difference is significant.

Hyperthreading adds 25%: http://forums.techpowerup.com/showthread.php?t=89777

 

Well, that's kind of my point, Hyperthreading theoretically makes it 8 (less efficient) cores vs 2, so to get the true 4 cores vs 2, you should turn it off, and at that point the numbers change. If hyperthreading does add 25%, that would seem to largely wipe out the apparent advantage of the i7.

 

No, you shouldn't turn it off, because the older Core 2 doesn't have it. By disabling Hyperthreading, it becomes yet another inaccurate benchmark. Disabiling features are fun for computer architecture theorizing, but not relevant to actual usage.

There are programs that benefit from the Hyperthreading part in Nehalem. And there are others that only care about the "other" side of Nehalem, which is the improved memory performance.

In reality, it'll be somewhere between both. The way Hyperthreading works its more of "4 efficient cores" than "8 less efficient cores".

 

From my tests, HT on vs HT off makes a negligible difference for most apps. Even running the Fritz benchmark you'll end up with very similar scores.

The only reason I ran with HT (and all benchmarks were with HT on) was because it was set as a default in the bios. I have yet to see a manufacturer turn it off by default, but I haven't researched very carefully.

 

That is just wrong. Not you, IntelUser (and thanks for sharing that link!) but whoever reported that on techpowerup was wrong (to say HT added 25%). I've run the test on multiple systems with and without HT and there's no way HT adds 25%. I believe that person did see a 25% increase, no reason to lie about it, but lots of factors (such as background processes) could contribute to that.

One thing to keep in mind is that the Fritz benchmark itself can vary almost +- 3% when run on the same system, even on back to back to back runs.

 

Ok, so if HT on or off makes no appreciable difference, then probably Fritz does truly load cores to 100%, such that the "spare" time that hyperthreading usually makes use of simply isn't there to be used. That makes sense, I think.

 

Not sure why the different results, I have Deep Fritz 11, ran the Benchmark (Lists for 8 Processors) using Windows 7 64 Bit Home Premium and got the following;

13.17 / 6323 kN/s

Full Spec in Sig below


As a comparison my Desktop runs a Q6600 overclocked to 2.95Ghz, 8GB PC 800 DC DDR2 RAM, Windows 7 Professional 64 Bit scores;

16.17 / 7763 kN/s

 

You can't say that until you have a good sample of numbers. If the 720 scores, let's say, 1000, 1000, 1000, 1000, 1000, and the 820 does 1002, 1002, 1002, 1002, consistently, then 1% is NOT within a "margin of error". Run the test 20 times and let me have the data and I'll tell you whether it's statistically significant or not.

 

Well, you can, it just depends on the confidence interval used. I understand what you're saying though, and we don't have enough sample data to provide entirely accurate results. However, the point of the post was to provide info to people deciding on the 720 vs 820, not an exercise in providing 100% valid statistical data.

Again, I agree with your point, but we're not publishing this to Scientific American or anything!

 

http://www.tomshardware.com/reviews/Intel-Core-i7-Nehalem,2057-12.html

Fritz 11: 23.8%

 

Interesting, thanks for sharing that. Obviously they used a bloomfield chip, but I will have to go back and look at that. I'm very surprised at those results. Either way, since both the 720QM/820QM have HT, it doesn't affect the results of this test, but it is something to look into for my Fritz benchmark post.

 

It's weird that the difference is so large. It's about 25% and the clock speed difference is only between 8-10% depending on turbo boost. Are you sure they are both done at the same resolution?

 

Absolutely. I didn't pay for 3dmark'06, and thus was unable to adjust any resolution related settings.

Also, note that the overall 3DMark score increase was just under 5% with the exact same GPU's.

 

I disagree. I give a rat's tail about transcoding and whatnot. What it comes down to for me when looking at two different components is "How will this affect my gaming experience." So in that case, I want the full run, including gpu.

In fact, I was wondering if having the slower HDD was dropping the 720m's 3dMark06 score. Also general architexture, since I've seen exact copy builds show increased performance on the larger chassis.

 

I do not think hard drive has any impact on 3dmark scores (and if it does, it's very small) - it only has a large impact on PCmark.

The size of a notebook should not have any impact at all on the 3dmark scores... unless somehow the smaller notebook is unable to handle the heat output of the GPU + CPU and downclocks components.

 

+ 1 rep (both for BS and MS). Great post to show us the difference. Dell/alienware can totaly forget im gona spend 430 euros more for the 820QM. Yet, I do feel that getting one should make your laptop last a longer time in case more future games will be as heavy as GTA4. For me it will depend wether it will matter on Assassins creed 2, if not, il go for the cheaper model, or whatever Dell/Asus are gonna offer me for 1.5-2000 euros ^^

 

I didn't think it would be major. Hard drive speed would mostly affect game loading times.

As far as size goes, I would assume that airflow/heat would be the main difference, though there might be differences in motherboard as well, I've no idea.

 

I wonder if upgrading to the 820qm would be 250€ well spent :S

From what I've been reading, 3D apps take full advantage of the 4 cores, and I'm hoping to be using this laptop in the next 3-4 years with 3DS Max and Zbrush.

Does the 820qm really pushes itself to 2Ghz at the 4 cores when needed?

 

I can't for the life of me understand why some people are so involved in raw processing power. How many really use their computer primarily for number crunching? A Pentium IV can probably beat all the participants of this thread at chess at the same time, and if you're into protein folding you might want to use the computers at your university.

Even if for some strange reason the CPU is the one major bottleneck in your daily computer use, I can't understand how one can be so enthusiastic given the certain knowledge that a much more powerful piece of silicon will be out in six months or so?

Anyway, I just really wish someone would write a little about things that to me are at least equally important: battery drain (in laptops) and heat generation. Here's a question I'd like answered:

Would my laptop be more silent with an i7 720 than with an i7 820??

 

Because we are enthusiasts, as someone tactfully reminded me as well, and I've yet to see any significant CPU change that didn't make me like using my system as it's more responsive. It's only fast enough when I don't think about how long ANYTHING takes to run...

Hence the term "Enthusiast"... the same reason I didn't opt to buy an Inspiron 15. Also, as for power, yes, every 6 months... but the i7 is in particularly a tech shift.... and since things DO change so much... you spend what you can to get the best you can to future-proof.... or you just like the latest thing (some people do)...

Battery drain is pretty well discussed as a basic benchmark... the GPU's I think are the worst offender here.... as for heat... people do talk but the gaming laptops generally deal with it pretty well now unless there is a problem with ventilation.

 

BrandonSi,

Thanks so much for the info, just bought a new m15x and was debating this exact issue. For my uses,the 720 is ample enough, thanks again +1

 

Quality thread from you there mate

+1

 

There was some disscussion on HT on and off so here are some links if someone is still interested...

http://ixbtlabs.com/articles3/cpu/ci7-turbo-ht-p1.html
http://ixbtlabs.com/articles3/cpu/archspeed-2009-3-p1.html
http://www.pcgameshardware.com/aid,694495/Intel-Core-i5-and-Core-i7-Lynnfield-CPUs-reviewed/Reviews/

 

No, Supreme commander is not. The 3rd and 4th thread of the game is for Audio and UI, In fact a faster dual core is better with SC anyday. My computer is barely at 40% CPU usage when running SC ( Q6600 @ 3.0ghz.)

Dirt 2 and Unreal TM 3 are pretty good quad core games.