Information

Dear forum members,
We are delighted to inform you that our forums are back online. Every single topic and post are now on their places and everything works just as before, only better. Welcome back!



Discussion

Discussion on Article:
AMD Phenom II X4 965 Black Edition: The Peak of Deneb Evolution

Started by: sstass | Date 08/13/09 02:25:10 AM
Comments: 11 | Last Comment:  12/27/09 04:20:32 PM

Expand all threads | Collapse all threads

[1-7]

1. 
I used to enjoy this website quite a bit. Recently I've seen some ridiculous pro-Intel/nVidia stuff, and really dislike the bias. This was the last drop. I'm not coming back here again. And for those interested in real Phenom II X4 965 BE performance, read here: http://anandtech.com/cpuchipsets/showdoc.aspx?i=3619
0 0 [Posted by: sstass  | Date: 08/13/09 02:25:10 AM]
Reply
- collapse thread

 
I don't really know what your problem is with this review. X-bit's results are practically identical to Anandtech's in terms of standings, even despite the different suite of tests. Anandtech's simply look more favourable toward the Phenom II because they place it among the top performers in a set of results also including a lot of CPUs with much worse performance. But practically everyone has an idea of the performance of at least two of the CPUs in the X-bit test, so Anandtech's extensive comparison with old or low-end CPUs such as Athlon X2 4600+ or Pentium E2140, while interesting, was largely unnecessary. X-bit can only report the results they collect: it is up to the reader to make a sensible interpretation of the graphical presentation of the numbers. In this case, it seems that you have simply been misled by the presence of a long tail in Anandtech's results that is not present in X-bit's.
0 1 [Posted by: MTX  | Date: 08/13/09 02:59:00 AM]
Reply
 
This article is not biased in anyway. It uses a standard set of applications for benchmarking. Nothing indicates that Intel was given some unusual advantage in the benchmarks. Just because we don't like the current trend in processor benchmarks does not mean that they are biased.
0 0 [Posted by: philosofool  | Date: 08/13/09 08:18:33 AM]
Reply

2. 
Your i7 power measurments are flawed. i7 consumes power for the whole Uncore part(L3, Memory controller, QPI links) via the 5V / 3.3V rail through the chipset, not only via the 12V CPU power rail. Thus your statement here is not correct, too:
And it is not just twice as much as a Core 2 Quad processor with comparable performance will consume, but is way higher than the actual power consumption of the Core i7 with 8 virtual cores and not 4. In other words, Phenom II X4 965 has very frustrating power consumption.


Read maybe this:

Interestingly, the entire Uncore is uncoupled from the CPU’s main power supply coming in through the auxiliary 12V rail. Most of the power for the uncore, that is the power for the analog and digital portions of the memory controller, the QPI link and the shared L3 cache are directly derived from the QPI bus voltage, which was formerly the AGTL voltage and is specked at 1.20V typical. As the power consumption is heavily dependent on load, Intel’s data sheet only lists the max values for the analog and digital components each as 5A and 23A, respectively, for a combined ITT max of 28A. At 1.20V, this results in 33.6W power for the logic portion of the uncore.
(...)
As a result, using conventional methods, that is, either measuring 12V supply current or else measuring at the MOSFETs directly will only generate data pertaining to the core section of Nehalem but will not take into account the uncore, which, at least according to Intel’s specifications can weigh in with as much as 44.58 Watts.


http://www.lostcircuits.c...;limit=1&limitstart=2
So you can add ~44W to your i7 power consumption measurement - in the worst case.

Please correct your article accordingly.

regards

Bingle
0 0 [Posted by: Bingle  | Date: 08/13/09 02:51:57 AM]
Reply
- collapse thread

 
Well spotted. The difference between the two sets of power consumption results did look rather too large to be accounted for by differences in chipset consumption alone.
0 0 [Posted by: MTX  | Date: 08/13/09 02:58:26 AM]
Reply
 
Thank you, Bingle. This is very useful info. I will investigate Core i7 uncore power consumption and will publish the results in my next articles. I think, I will add wattage measurements with mainboard power consumption via 24-pin power connector for all contemporary platforms.
0 0 [Posted by: Gavric  | Date: 08/13/09 07:26:31 AM]
Reply

3. 
Test is different on AnandTech ,but AnandTech use different hardware,but this new AMD CPU show advance in video compression ,now take second place behind CoreI7 965 in x264 compression and wmv encoding,game etc.. My opinion that Intel Nehalem price begin to shake,matter of time is when Intel cut price.
0 2 [Posted by: Blackcode  | Date: 08/13/09 10:39:41 AM]
Reply

4. 
I've seen some ordinary reviews in my days....

This one takes the cake.
0 0 [Posted by: wasteoftime  | Date: 08/15/09 07:42:08 AM]
Reply

5. 
Come on xbitlabs, until when you're going to do worthless gaming reviews in your articles. Is just a prehistoric way of "demonstrate" the cpu usage in games, nobody cares about medium or low details.


Put 1680x1050 / 1920x1200 high/vhigh with no AA or 4xAA.


Then people talk about bias, using this pathetic "medium" gaming gives reason to them.
0 2 [Posted by: Nintendork  | Date: 08/15/09 11:00:57 PM]
Reply

6. 
0 2 [Posted by: Blackcode  | Date: 08/16/09 12:02:08 AM]
Reply

7. 
A good test
can you compare the test use win 7 as a new operations system ( will become favourite for computer user) to know are they different from win vista ? thank you
0 0 [Posted by: spwwhite  | Date: 12/27/09 04:20:32 PM]
Reply

[1-7]

Back to the Article

Add your Comment