Memory Latency and Bandwidth

We've never been able to look at some of the low level characteristics of Intel's Core architecture, and although we didn't have enough time to do a thorough run of low level benchmarks we were able to run ScienceMark 2.0 in order to get an idea of how the Core 2 Extreme stacked up against the FX-62 in terms of memory latency and bandwidth.

We had seen Conroe performance results that showed the new architecture being able to offer fairly competitive memory access latencies to AMD's architecture, without the need of an on-die memory controller. Our ScienceMark 2.0 results confirm just that:

ScienceMark 2.0 - Memory Latency (256-byte stride)

While AMD still offers lower memory latency, the Core 2 Extreme X6800 is very close in comparison - especially considering that it has no on-die memory controller. With lower clock speeds than its Pentium D siblings and a faster FSB, memory access latency is reduced tremendously with Conroe. On a larger scale, through a very effective cache subsystem as well as memory disambiguation, Conroe can offer significantly improved memory performance compared to its predecessors, including the Athlon 64 X2/FX.

ScienceMark 2.0 - Memory Bandwidth

ScienceMark's memory bandwidth results offer a very telling story, showing us the bandwidth limitations of Intel's FSB architecture. While the FX-62's peak theoretical bandwidth is not achieved in real world, you can see how AMD's Direct Connect architecture offers higher limits for chip-to-chip communication.

Index Business Application Performance
POST A COMMENT

133 Comments

View All Comments

  • PCSJEFF - Wednesday, June 28, 2006 - link

    If you wanna test the CPU in games, why don't you use also Grand Prix 4 and Everquest 2: those two games have a 3D engines that use a lot more the CPU than the video card. Reply
  • drewintheav - Wednesday, June 07, 2006 - link

    I will not be surprised if conroe ends up costing twice as much in the retail channel as what is being stated at the current time. It might take a while for the prices to stabilize and then fall back down towards the prices that are being quoted at the present time... Reply
  • mine - Friday, June 09, 2006 - link

    OH MYGOD I am so tired of these Conroe vs. FX , Intel vs. AMD benches -
    You find them everywhere in every incarnation on every website in every printed medium..
    and i am tired of this Intel-AMD fanboy crap...
    Anand had a chance of testing the first mainstream motherboard for Conroe that will be sold in 100 k quantities in the next months. This would have opened a ideal chance of testing the Intel claims of the new 965 chips much better performance.

    and why so silent about the used motherboard ....

    gone.... anand took the "SUN " way

    sorry for my..english it is not my native language



    Reply
  • macugenwatcher - Wednesday, June 07, 2006 - link

    do think on single application the conroe extreme is somewhat better than fx62 but remember Anand's comparsion was with ghz of 2.93 for conroe and 2.8 for athalon fx....

    Also his benchmark of fx62 must be on poorly optimized system. Here's the proof:

    sysmark 2004 overall office

    anand core extreme--- 266
    anand fx62---210
    pc stats fx62--- 261

    sysmar 2004 document creation

    anand core extreme--- 366
    anand athlon fx62--- 280
    pc stats fx62--- 297

    sysmark 2004 data analysis

    anand core extreme--- 280
    anand athlon fx62--- 185
    pc stat fx62--- 214

    business winstone 2004

    anand core extreme--- 32.8
    anand athlon fx62--- 27.9
    pc stat fx62--- 36.4

    Check out the Sysmark2004 overall and Business winstone 2004!!!

    What the heck happened to Anand's FX system?

    Where the heck is the multi-application tests?

    Why couldn't Anand run the full gamut of tests if he had these systems in his hotel room? He said he had to hurry up before people figured out that he had a Coroe chip in his room.... SAD SAD SAD.....

    Definitely, Anand's FX system was crippled by poor memory... This is very very sad statement on Anand's objectivity....

    Athlon FX going to 65nm soon with gigzhertz up to 3.2 to 3.4, will be the performance leader in 4q-2006. Intel will have no answer. INTC is now saying "competitive processors" not "AMD beating processor"

    http://www.pcstats.com/articleview.cfm?articleID=1...">http://www.pcstats.com/articleview.cfm?articleID=1...
    Reply
  • Gary Key - Thursday, June 08, 2006 - link

    quote:

    Definitely, Anand's FX system was crippled by poor memory... This is very very sad statement on Anand's objectivity....


    Both systems were setup identically to ensure an unbiased test platform with both boards being configured at default settings, the way a system will be delivered to the customer. I can ensure you that lower latency settings also improve Conroe's performance.

    quote:

    Athlon FX going to 65nm soon with gigzhertz up to 3.2 to 3.4, will be the performance leader in 4q-2006. Intel will have no answer.

    We have already know that Intel has production ready Conroes at much higher stock speeds than what will be released in July. However the transition to 65nm allows AMD to do signficantly more changes than just clock speeds, like improving the memory controller which is where they will realize significant improvements beyond the current DDR2 memory setups. Of course, we will be talking about DDR3 and 45nm production for Intel next year so just be happy the CPU landscape is competitive again. I must have missed the Conroe scores in your PCstats article. ;->
    Reply
  • Gary Key - Thursday, June 08, 2006 - link

    Where's the edit button? "We already know...." Reply
  • CobraT1 - Wednesday, June 07, 2006 - link

    do think on single application the conroe extreme is somewhat better than fx62 but remember Anand's comparsion was with ghz of 2.93 for conroe and 2.8 for athalon fx....

    Your opinion which is not widely shared. CPU speeds were noted and commented on. Everyone who read the article was fully aware.

    Also his benchmark of fx62 must be on poorly optimized system. Here's the proof:

    Check out the Sysmark2004 overall and Business winstone 2004!!!


    This only proves PCstats used a different configuration. Previewing their article I could not delineate if they were testing at 2.8 or 3.08, the 2 speeds for the FX-62 they listed in their system configuration. (I could have missed it, I quickly previewed it.) They also used a different system configuration then what Anand had at his disposal.

    What the heck happened to Anand's FX system?

    It appears nothing. These tests echo results that have been known for sometime.

    Where the heck is the multi-application tests?

    Why couldn't Anand run the full gamut of tests if he had these systems in his hotel room? He said he had to hurry up before people figured out that he had a Coroe chip in his room.... SAD SAD SAD.....


    I do believe Anand answered your question on the very first page and has been referenced numerous time on this very page.
    "Keep in mind that we had a very limited amount of time with the hardware as to not alert anyone that it was missing and being used for things it shouldn't be (not yet at least), so we weren't able to run our full suite of tests. We apologize in advance and promise we'll have more when Conroe launches, but for now enjoy."

    Completely up front as Anand usually is.

    Definitely, Anand's FX system was crippled by poor memory... This is very very sad statement on Anand's objectivity....

    In Anand's comparison, identical memory was used in like systems. PCstats comparison cannot make the same statement. PCstats numbers are hardly comparable and certainly not grounds for accusations.

    Athlon FX going to 65nm soon with gigzhertz up to 3.2 to 3.4, will be the performance leader in 4q-2006. Intel will have no answer. INTC is now saying "competitive processors" not "AMD beating processor"

    As I have no idea where you came up with these statements, I see no need to comment on what appear to be your beliefs. Beliefs require blind faith, something I have none of.

    If you just have to believe Anand has been corrupted, stay away from the other sites that are and will be reporting like comparisons with similar data and conclusions.

    Anand's comparison was a great early indicator and has for the most part confirmed what has been reported on various sites for the last few months.

    Bottom line, nothing shocking has been revealed.
    Reply
  • cscpianoman - Wednesday, June 07, 2006 - link

    I was very impressed with the Conroe numbers. I have an Uncle who works for Intel and I really liked his demeanor about the chip. He commented they have a great chip in line for consumers, but he didn't seem over the top. He knew AMD would respond and it sounded like Intel's attitude has changed for the better. Instead of flaunting gigahertz numbers they are getting down to business. The Conroe intro a few months back was controlled by Intel probably to generate discussions like these and increase market awareness. Now that we know they were legit, it has cemented more purchasers in the future. I suspect there will be a lot of AMD users on the forums asking which intel chipset to get to match the Core 2. Reply
  • zsdersw - Wednesday, June 07, 2006 - link

    You think clock speeds for Conroe won't change by the end of 2006? What are you smoking? Reply
  • macugenwatcher - Wednesday, June 07, 2006 - link

    Sure it can increase. But what I am upset about is these SEMI-OBJECTIVE tests on a pre-production Conroe model versus production AMD chips....

    Why are there such discrepancy in sysmark and winstone for the AMD chip?

    If Anand professes to be a true journalist/researcher/benchmarker, he should be more objective in his testing.

    I am sure Conroe will be very good. But in different applications, especially in multi-application setting, AMD chip could be better.
    Reply

Log in

Don't have an account? Sign up now