Product: Core 2 Extreme Quad Core QX6700
Author: Sean Kalinich
Category: Enthusiast
Reviewed on: 12 October, 2006
Product cost: $999.00
Manufacturer: Intel
Spelling and Grammatical editor:
|
Introduction:
It was not that long ago that Intel was sitting around the lab, scratching their heads, wondering what they could possibly do to reclaim the lead they had lost to AMD. Then someone came up with the right answer. That answer was Conroe, better known as the Core 2 Duo. Much to the chagrin and dismay of many an AMD fanboy, the Core 2 Duo not only caught up to AMD?s flagship CPUs, it simply strolled by like AMD was sitting still.
Then in the midst of those same AMD fans preparations for a Hemlock and Tonic, came Intel?s insult to injury. This arrived in the form of the Quad Core Kentsfield to be better known as the Core 2 Extreme quad-core (QX6700). We have had one of these little marvels in our labs for the past few days and have been simply amazed at its ability to thumb its nose at the law of diminishing returns. Now this is not to say that AMD has simply given up. There is the 4×4 that is due out sometime soon and AMD?s own quad core Barcelona. But none of these are here now. This leaves AMD in a position they have not been in for a while? behind and playing catch up. So let?s rub a little salt in the still fresh Conroe Vs AM2 wound and show you our experience with Kentsfield.
|
Click to Enlarge
|
{mospagebreak title=What?s New?}
If a tree falls on you in the woods and no one is around to hear it will you still be dead?
The first thing that people will notice is different with Kentsfield is the fact that there are four CPU cores. The second thing they will probably notice is the rise in temperature around their work area while the system is on. Besides heat and an extra pair of CPU cores the Kentsfield is basically the same in features as a Core 2 Duo. But Kentsfield does have a few other new features and goodies thrown in.
We will cover some of the more interesting ones below.
Unlocked ?
The Core 2 Extreme Quad Core will be unlocked; this means that the usual Bus locks that are in place have been taken away so the Enthusiast is free to ?tune? (that is pronounced ?overclock the sh*t out of?) the CPU for the best performance.
Digital Thermal Sensors ?
The Core 2 Extreme quad core CPUs come with a digital thermal sensor this allows for more accurate monitoring of the internal core temperatures. Combined with the Quiet System technology in the 975X chipset this allows for better thermal and noise control.
Intel Wide Dynamic Execution ?
Although this is not new to Kentsfield it is worthy of mention. With Wide dynamic Execution, the Core 2 Extreme quad core CPU is capable of executing 16 full instructions simultaneously.
Advanced Smart Cache ?
Another item that was introduced with Conroe that shows up in Kentsfield.
This system of shared cache allows dynamic allocation of cache to the CPU cores based on workload. So if a core needs more cache to aid in execution of code, Intel?s Advanced Smart Cache can provide it.
A quick word on the 8MB of cache sported by Kentsfield, the cache is split into 4MB of cache per pair of cores. Each core pair is able to share the 4MB available to them but cannot share the cache assigned to the other pair.
Power:
The Core 2 Extreme Quad Core is a power hungry CPU. Coming in at a whopping 130Watts TDP. This CPU is not for the faint of heart or the weak of PSU.
The power consumption is just about equaled by the performance output unlike many of Intel?s older power draining CPUs.
Heat:
With four physical cores and 8MB of cache you can bet the QX6700 puts out some heat.
The ambient temperature around my testing station gets rather warm while running this CPU. During testing I measured temperatures at a few different fan speeds at idle and under load. The revision of the D975XBX2 that Intel sent out to test the QX6700 does not support fan speed control using the PECI (Platform Environment Control Interface)
. This meant that the FAN speed was manually set to full speed in the system BIOS. At this speed the fan was so loud I could barely concentrate. So I began testing to see what speed I could run the fan at while maintaining good temperatures and my sanity.
The settings in the BIOS allow for adjustment from 100% to 60% in 10% increments
I have included the temperatures at idle and under load for each step.
|
Fan Speed
|
Idle Temp
|
Load Temp
|
|
100%
|
44c
|
73c
|
|
90%
|
45c
|
75c
|
|
80%
|
45c
|
77c
|
|
70%
|
45c
|
78c
|
|
60%
|
47c
|
81c
|
In the end I ran the fan at 70% for my testing.
{mospagebreak title=Methods}
Do you know that under ?redundant? in the dictionary it says ?see redundant??
Methodology:
For this evaluation I wanted to test the CPU with as many real world tests as I could.
There will be very few synthetic benchmarks and no time demos in this evaluation. They do not serve to show off real performance.
In addition to the software used for testing, I also wanted to simulate a real world environment, for this I installed the following commonly used applications:
Microsoft Office 2003
SpeedFan
Diskeeper 10
Avast Home Edition
Acrobat Reader
FireFox
A HighSpeed PC TechStation was used for all testing.
{mospagebreak title=Test Systems}
Intel Core 2 Extreme QX6700
Intel Desktop Board D975XBX2
2GB Kingston HyperX DDR2-800 CL4 (SLI Ready)
2x
74GB
Western Digital 10,000RPM Raptor SATA HDDs (RAID0)
Leadtek PX7900GTX TDH Extreme 512MB
SilverStone Zeus ST75ZF 750 Watt PSU
Intel Reference Cooler
Generic 16x DVD-ROM
Generic 1.44MB Floppy Drive
Microsoft Windows XP Professional/ Windows XP x64 Edition
Core 2 Extreme X6800 (2.93GHz)/ Core 2 Duo E6700 (2.66GHz)
Intel Desktop Mainboard DX975XBX
2GB Kingston HyperX DDR2-800 CL4 (SLI Ready)
Leadtek PX7900GTX TDH Extreme 512MB
2x
74GB
Western Digital 10,000RPM Raptor SATA HDDs (RAID0)
SilverStone Zeus ST75ZF 750 Watt PSU
SilverStone NT06 CPU cooler
Generic 16x DVD-ROM
Generic 1.44MB Floppy
Microsoft Windows XP Professional/ Microsoft Windows XP x64 Edition
AMD
AM2 FX-62 (2.8GHz)/ X2 5000+ (2.6GHz)
MSI K9A Platinum
Corsair XMS2 Twin2X2048 6400C4 DDR2-800 CL4 (SLI Ready)
Leadtek PX7900GTX TDH Extreme 512MB
2x
74GB
Western Digital 10,000RPM Raptor SATA HDDs (RAID0)
SilverStone Zeus ST75ZF 750 Watt PSU
Stock AM2 Cooler
Generic 16x DVD-ROM
Generic 1.44MB Floppy
Microsoft Windows XP Professional/ Microsoft Windows XP x64 Edition
{mospagebreak title=Performance Tests Gaming}
Gaming:
These four game titles have heavy physics and particle effects throughout the game. They are in many cases capable of bringing even the fastest system to a crawl. All games were run at 1024×768 with no additional graphical options on. The only exception to this was HalfLife 2 Episode One. HDR was set to Full, the HDR used in HalfLife is a more CPU intensive integer based HDR.
Need For Speed Most Wanted:
Need for Speed Most Wanted is just plain fun, I can always load up this game and get a kick out of driving at breakneck speed, running into EVERYTHING, and never trashing my car. On the system side the physics calculations and the flock-like AI (if you have played the game you know the Police vehicles move in unison like a flock of birds) can have a significant impact on system performance. For my testing I ran Challenge number 9 out of the challenge circuit. This little tour in the country side has you running between tollbooths to beat certain times. All the while the Police are chasing you and trying to get you to stop. There are a couple of 90 degree turns that really show off the effects of physics in the game.
|
XP Pro
|
XP x64
|
|
|
|
Not much of a gain here, in fact there is a slight performance hit for the QX6700.
F.E.A.R.
F.E.A.R. is another game I can really get into; there are physics, particle effects, and some devilish AI that combine with Monolith?s penchant for supernatural themes to make for an immersive gaming experience as well as a game that can throw the breaks on your speedy system. For testing I measured frame rates during the Interval 02 level from just after the cut scene of the little girl burning up the Special Ops soldiers through to where the same little girl hurls you out of the window.
This section of the game allows for two very nice little firefights, with multiple enemies in both open and channeled terrain.
|
XP Pro
|
XP x64
|
|
|
|
Again we see the same thing the QX6700 is lagging slightly behind the E6700
Call of Duty 2:
I have said it before and it bears repeating Call of Duty 2 is a bar fight of a game.
Where F.E.A.R. gets you with clever AI, Call of Duty 2 gets you with an over whelming number of enemies, add to that the minor issue that your allies tend to jump right in front of you while you are shooting (and if you kill them by accident you are branded a traitor and have to restart) combines to make Call of Duty 2 a very intense and chaotic game. Call of Duty also has some good particle and minor physics effects that can affect game performance in a system with a slow CPU.
For testing I used the Demolition level and played from the beginning of the level until the destruction of the German building at the end.
|
XP Pro
|
XP x64
|
|
|
|
Here is something interesting; the game with the most basic AI seems to benefit the most from the extra cores. Of course Call of Duty 2 does have a multi-core patch, but then again so does F.E.A.R. and we did not see this improvement there. Either way the QX6700 beats out even the X6800 in Call of Duty 2.
HalfLife 2 Episode One:
HalfLife 2 Episode One was an obvious choice for game testing. The Havoc physics engine combined with the AI to be found in all of the HalfLife titles was something that I could not pass up. Not to mention I simply enjoy playing the game.
For my testing I chose a section of the LowLife level, where Gordon and Alyx move though the ruins of an underground parking garage full of antlions. The numerous enemies and the need to move several destroyed cars to cover up the antlions access holes allow for the physics engine to come into play. I continued with this level until I reached the elevator and was able to board it (after the fight in the dark with the various zombies).
|
XP Pro
|
XP x64
|
|
|
|
HalfLife 2 is another place where the extra cores seem to hinder gaming performance instead of helping.
Gaming Observations:
Are four cores better than two at gaming? Considering that most gaming developers have yet to make the leap to multi-threaded games, I was not surprised to see the QX6700 come in with gaming performance numbers similar to the E6700, the only notable exception to this was in Call of Duty 2 where the QX6700 showed close to a 10 FPS difference over even the X6800. Remember that the numbers shown above are only indicative of the CPUs impact on gaming performance. No graphical enhancements were enabled with the exception of the Integer based HDR in HalfLife 2 Episode One. As you enable more options to improve the video quality your gaming experience becomes more dependent on your GPU.
The tests above were only meant to show the ability of the CPU to handle the back end of gaming performance, the AI, Physics, and positional audio calculations involved in fleshing out the games.
{mospagebreak title=Performance Tests Applications}
If a tree falls on Britney Spears in the woods and no one is around will anyone care?
Cinebench 9.5:
Cinebench is a test that we have only recently implemented. Based on the Cinema 4D engine it is an excellent method for showing the benefits of Multi-Core Processors in image rendering.
In Cinebench our scores were close to double to scores of the E6700, this is almost a linear performance scaling and shows what the QX6700 can do with a truly multi-threaded application.
LightWave 3D:
The second test in our rendering applications, Newtek?s Lightwave 3D, has been the choice of many professionals for years now. Starting with version 8.5 Newtek included 64 bit support. In our article on the benefits of 64 bit over 32 bit found here we saw how a multiple core SMP system was able to take advantage of the 64 bit memory addressing and run rings around a similar 32 bit system. With the Core 2 CPUs we have multiple cores and 4MB of cache to buffer rendering information, add into this the improved performance of the EM64T instructions. With the Core 2 Extreme Quad Core we literally double this with two Core 2 CPUs stuffed under the IHS for a total of four Cores and 8MB of cache.
My render sample is available from the content CD. It is the moonbase scene. I set the camera resolution to 3200×1600, for my AA I used a PLD ? 5 Pass run with Gaussian (sharp) for my motion blur. Thread was set to 2 with segment memory at 128MB.
|
|
X6800
|
E6700
|
X2 5000+
|
FX-62
|
QX6700
|
|
XP Pro
|
41:25
|
45:39
|
1:12:58
|
1:05:34
|
20:55
|
|
XP x64
|
41:45
|
45:48
|
1:12:29
|
1:05:02
|
21:10
|
Time in Hours, Minutes, and Seconds. Lower is Better
Ok all I can say is Wow, the times that we saw were just plain impressive. Again we see the QX6700 show close to linear scaling in terms of performance.
Photoshop CS 2:
Photoshop CS2 is an excellent test for measuring the Core 2 Extreme quad-core CPU?s ability to handle image manipulation. Photoshop was designed to be multi threaded and scales nicely with multiple CPUs. For our testing we ran Diverheaven.net?s Photoshop test. The times for each filter run during the test are shown below.
|
XP Pro
|
XP x64
|
|
|
|
Now here we have something of a surprise. Although in some filters the QX6700 was faster than the E6700 it was not faster than the X6800. And only barely beat out the E6700 in total time. I would have expected the QX6700 to show similar scaling in Photoshop as we saw in LightWave and Cinebench. This just did not materialize.
SuperPi:
What can be said about SuperPi that hasn?t already been said? I could say it is one of the more boring tests to run but that really doesn?t add much now does it? The reason we included SuperPi is that for many people it is a measure of performance and stability. The speed that your CPU can calculate the number Pi out to 32 million places is just as important to some as 3DMark or PCMark are to others. For our testing we ran a single instance of both the 1m run and the 32m run. After this we ran one run per core with core affinity and without.
With SuperPi I was not expecting any performance improvement at all. I did end up seeing a very small improvement in the single run but nothing to write home about. In our multiple run testing there was actually a performance hit by running one instance on each core. This could be due to the amount of information flooding the memory bus. Whatever the cause it is something to take note of.
I have included the numbers in a screen shot below as the sheer amount of information is more than Most will want to go through.
All 32M run times are listed as MM.SS.SSS
All 1M runs times are listed as SS.SSS
Easy DVD to DivX SVCD Converter :
Encoding DVD movies to a DivX AVI is another test that can show the scaling available with Kentsfield and running multithreaded applications. DivX 6.4.2 detects and uses all four CPU cores.
|
|
X6800
|
E6700
|
X2 5000+
|
FX-62
|
QX6700
|
|
XP Pro
|
36:47
|
50:18
|
53:42
|
50:53
|
29:50
|
|
XP x64
|
30:31
|
46:07
|
39:43
|
37:35
|
25:46
|
For DivX encoding we see a healthy improvement in performance with the QX6700 beating out the X6800 by roughly five minutes even though it is clocked slower. Here again we see the power of multi-threaded applications.
{mospagebreak title=Numbers!}
Numbers Game:
Again for all of you that just have to know how much better your 3DMark and PCMark scores will be I have included those here just for you.
|
3DMark06
|
PCMark05
|
|
|
|
{mospagebreak title=Value}
Price:
You knew there had to be a catch. Well if anything price might be it. The QX6700 runs at $999.00 this is the same price point as the X6800 and keeps in line with the Extreme series pricing. Although this is a steep price it is still significantly less than buying 2 Dual Core Xeons or Dual Core Opterons.
{mospagebreak title=Conclusions}
Conclusion:
So are four cores better than two? The answer is both a resounding yes and a disappointing no.
For gaming, four cores is not going to show much of an improvement over two or even one until game developers begin coding games to take advantage of them. Now you may say that Crysis and Alan Wake are due out soon and that they can take advantage of multiple cores. To that I say, ?They are not out yet? the performance numbers that we have seen have been on demo systems that can be tweaked however the person running the demo wants them to be. Or you may say that games like Quake 4 and F.E.A.R. take advantage of multiple cores.
In truth neither of them, nor any other game currently on the market was coded to use multiple cores or CPUs. Even Call of Duty 2 where we saw a performance gain had the multi-core support added as a patch. For the here-and-now and for the near future, Multi-Core CPUs will have very limited impact on your gaming performance especially at high resolutions. In our testing at 1280×1024 and above it was clear that the GPU was the determining factor and the CPU power was removed from the equation. This is why we chose to test at such a low resolution. Once you bump up the resolution and start adding the eye candy (AA, AF, Fog, etc.) you become GPU reliant. Although we were not able to test SLI or Crossfire with this evaluation both of those Multi-GPU solutions can be limited by your CPU.
The Core 2 Extreme Quad Core is not going to improve your current gaming performance if you are already running a Core 2 CPU. If you are running a Pentium D or even an AM2 CPU then the jump to the Core 2 Extreme Quad Core will probably give you a healthy boost. But for both of those you will probably be upgrading your platform as well.
For applications, the answer is again yes and no. If the application is capable of handling true multithreading then yes (as you would probably have guessed) the Core 2 Extreme Quad Core will give you an almost linear scaling in terms of performance. Our LightWave and Cinebench testing showed us this. In other performance tests that were multi-thread aware but did not seem to know what to do with them, we saw a marginal performance increase.
The biggest surprise was Photoshop; this application is supposed to scale nicely as more CPUs are added. The testing we ran did not show this. Due to time constraints we were unable to delve too deeply into the reasons for this and will be looking into it at a later date.
For your Workstation needs the Core 2 Extreme Quad Core is a handy answer that gives you amazing performance at a lower Total Cost of ownership. This puts it right in the Pro-Sumer market that has been largely ignored by Intel. You can expect the Small (and large) Production companies, Photographers, Digital Animators, Editors, and Cinematographers to line up to grab this one.
In the end the simple fact is that Intel has solidified the position as performance leader that they took by force with the release of the Core 2 Duo line. With Kentsfield they are first to market with a desktop Quad Core CPU. AMD?s much vaunted Barcelona will not be out for some time (perhaps Q2 07) this will put Intel in a position to grab back some of the Pro-Sumer workstation market share that AMD has been gobbling up. It looks like Intel has AMD on the ropes and backpedaling. I hope that AMD?s upcoming releases are show-stoppers or they may be in real trouble.
Acknowledgments:
I would like to thank Dan Snyder at Intel for providing us with the QX6700 for this evaluation.