Results 1 to 27 of 27
  1. #1
    Join Date
    Aug 2004
    Posts
    371

    Stress testing dualcore CPU reliability at Tomshardware

    http://www.tomshardware.com/stresstest/index.html

    Quite interesting to watch those live cams
      0 Not allowed!

  2. #2
    Join Date
    Aug 2004
    Posts
    371
      0 Not allowed!

  3. #3
    Join Date
    Oct 2004
    Posts
    149
    Quite interesting to see the AMD with 0 reboots and the Intel with 3,
      0 Not allowed!

  4. #4
    Join Date
    Aug 2004
    Posts
    371
    Reboots are actually crashes...

    These tests are run automatically, when the system crashes, some guy needed to be reset the PC to start it again, so you will periods where INTEL CPU temp reading is 0
      0 Not allowed!

  5. #5
    Join Date
    Aug 2004
    Posts
    371
    http://www.tomshardware.com/stresstest/index.html

    4:0 now, INTEL crashed 4 times, AMD 0
      0 Not allowed!

  6. #6
    Join Date
    Mar 2003
    Location
    New Jersey
    Posts
    1,277
      0 Not allowed!

  7. #7
    Originally posted by wm2100
    http://www.tomshardware.com/stresstest/index.html

    4:0 now, INTEL crashed 4 times, AMD 0
    Amazing what happens when you read. Those performence graphs, also include around 7 hours of time that intel wasnt doing anything, while they tried to figure out the problem.

    "The actively cooled Northbridge of the GA-8N-SLI Royal becomes unusually hot, so in our opinion the chip set is probably the cause of the crashes"
      0 Not allowed!

  8. #8
    Join Date
    Aug 2004
    Posts
    371
    Originally posted by crnc
    There is a technical forum for a reason. you should use that. Btw, I love Intel. Down with AMD.
    yeah, yeah, yeah, INTEL fans love INTEL crash or not
      0 Not allowed!

  9. #9
    Join Date
    Mar 2004
    Posts
    158
    That intel box seems to be having some real trouble. I'm also not clear on how they are reporting reboots, since it showed 3 the first time I looked yesterday evening.. later in the evening it showed 1 with an uptime of around an hour. Now it shows 4 with an uptime of less than an hour.

    However they are doing it, it looks like the intel box is choking quite a bit. I'll be interested to see what they have to say at the end of the test.

    CRNC: Any particular reason why you have such a partisan view? Is it because you use Intel and want the company you have an association with to succeed (kinda like a favorite football team) or is there a technical reason you like them? Just curious.
      0 Not allowed!

  10. #10
    Join Date
    Aug 2004
    Posts
    371
    Personally, I had used a number of INTEL dedicated servers, they ran into hardware problem after a long period. I am not sure if those were crappy components dedi hosts used, or a CPU problem. but my systems were never under this kind of stress.

    Remember, the TDP figure on INTEL CPU label is the typical power not the max, for example, a 105 watt prescott's max power is 151 watts. I guess the stock heatsinks are designed on typical power, under the stress, i think the INTEL CPU is always near max power even though INTEL CPUs can clock throttle to cool down very fast when temps are too high. AMD's power figure is different, if you read AMD's spec, it's labeled as Theoretical Maximum Power, so an Opteron at 67 watt won't exceed 67 watt under any circumstance..
      0 Not allowed!

  11. #11
    Originally posted by angst7

    CRNC: Any particular reason why you have such a partisan view? Is it because you use Intel and want the company you have an association with to succeed (kinda like a favorite football team) or is there a technical reason you like them? Just curious.

    The Intel choice is a personal view. We will do any server on a custom quote, so it doesn't matter if they are AMD or Intel. If a client wants it, we will build it. (not to advertise, but a month or 2 ago, we where offering about 10 different dual opteron specials)


    BTW angst7, my post above says why they believe the intel machine is having issues.
      0 Not allowed!

  12. #12
    Join Date
    Nov 2004
    Location
    UK
    Posts
    191
    wm2100 did you not get it when we slated your AMD post this morning, benchmarks are all very well but both Intel and AMDs are both good processors. AMDs may have the slight edge over Intels at the moment but 6 months down the line it could be the other way around.

    Jona
      0 Not allowed!

  13. #13
    Join Date
    Mar 2003
    Location
    New Jersey
    Posts
    1,277
    Originally posted by angst7
    CRNC: Any particular reason why you have such a partisan view? Is it because you use Intel and want the company you have an association with to succeed (kinda like a favorite football team) or is there a technical reason you like them? Just curious.
    I forgot that i'm not entitled to have my own opinions. Apologies.

    Seriously, I have no affinity towards either processor; I just find it very strange that these threads keep getting posted in this forum, and not in one that is more appropriate to the subject matter.

    I LOVE INTEL AND AMD AND POWERPC AND RISC.

    Hows that.
      0 Not allowed!

  14. #14
    Join Date
    Nov 2004
    Location
    UK
    Posts
    191
    Originally posted by crnc

    I LOVE INTEL AND AMD AND POWERPC AND RISC.
    Heh, I like it :-)

    I think the point people are making is that on a webhosting forum we have no intrest in one sided benchmarks. Many of us already have AMDs and Intels deployed and as I said whilst I believe AMD have a slight upper hand at the moment that is no reason not to buy intel if you have a good reason to do so and give it 6 months the situation could well be reversed. It is a similar arguement to windows vs mac, we're never going to resolve it but posting pointless statistics is pretty useless......

    Jona
      0 Not allowed!

  15. #15
    Join Date
    Aug 2004
    Posts
    371
    In single core CPUs, INTEL and AMD are still quite close. A 3.6GHZ Xeon single core is close to Opteron 248 in many cases. However, once we move to dual core or dual processor, INTEL can't hold up. It has a 50% to 90% performance disadvantage. Soon, dual core AMD chips will go mainstream and we will all benefit(now they are $500 bucks a chip). INTEL won't have dual core server chips until they go to 65nm in 2006.

    INTEL has hit the speed ceiling with their current architecture, 4GHZ is no longer achievable, the whole P4 /Xeon team and INTEL's VPs in the server group have quitted INTEL. INTEL is working on a whole new design based on Pentium 3, which will take some time. So far, INTEL has not been able to clone AMD64 instruction set on to Pentium 3, that's why INTEL won't have mobile 64 bit chips until 2006. But what we heard is AMD will have a big new fab up running in 2006 at 65nm and producing quad-cores...

    Heat production wise, INTEL is at 151 watts max per core.
    AMD is at 67 watts. So AMD can still increase frequencies. In fact, AMD is going to release the 2.8GHZ FX57 soon.
      0 Not allowed!

  16. #16
    Join Date
    Mar 2004
    Posts
    158
    Originally posted by crnc
    I forgot that i'm not entitled to have my own opinions. Apologies.
    You know I wasn't trying to imply that you're not entitled to your opinion. I was just curious what caused you to adopt the preference you so plainly stated.

    Originally posted by crnc
    I LOVE INTEL AND AMD AND POWERPC AND RISC.

    Hows that.
    Excellent.

    As for the stress test, it seems that they've restarted it. Both processors show the same uptime now, and are set at 0 reboots. Perhaps they corrected the chipset heat problem that concept referred to. It'll be interesting to see the results of this test. I love these kind of torture tests.
      0 Not allowed!

  17. #17
    Join Date
    Aug 2004
    Posts
    371
    Last year, they did some similar tests, the INTEL system crashed quite a few times, and Tom blamed the power supply. Guess this time chipset does not work for INTEL?

    But, let's keep in mind this Pentium D EE dual core is a 170 watt monster, it may burn a hole on the MB. I think the stock INTEL heatsink, albeit a big one, is not designed for such torturous tests.

    Another thing to notice is whenever the INTEL chip gets hot, it drops load by 20% for a second or so to cool itself down, a smart throttling technology in INTEL chips, which allowed them to have high nominal frequencies.
      0 Not allowed!

  18. #18
    Join Date
    Aug 2004
    Posts
    371
    INTEL crashed again! It was 4:0, INTEL crashed 4 times.
    Tom probably changed the MB and give both a fresh new start, unfortunately INTEL crashed again!

    http://www.tomshardware.com/stresstest/intel.html
      0 Not allowed!

  19. #19
    Join Date
    Dec 2004
    Posts
    256
    Again, did you do ANY research?

    http://www.theregister.co.uk/2005/04...tel/page3.html

    A quick google of that motherboard shows severe stability issues. Which would pin the problem on both gigabyte and nvidia... as the article clearly states. Yet again, FUD from wm2100
      0 Not allowed!

  20. #20
    Join Date
    Aug 2004
    Posts
    371
    read it again, the link you posted was about pre-production samples, tom is using validated products
      0 Not allowed!

  21. #21
    Join Date
    Dec 2004
    Posts
    256
    Originally posted by wm2100
    read it again, the link you posted was about pre-production samples, tom is using validated products
    And again, keep googling, the problem persisted in the "final revision" as well.
      0 Not allowed!

  22. #22
    Join Date
    Aug 2004
    Posts
    371
    Well, Tom is a die hard INTEL fan, I am not familiar with any of these intel boards, is Tom aware of the issues you alleged? We would certainly like a fair comparison, not something obscured by MB stability. Maybe you can email Tom and ask them to use a INTEL chipset.

    My bet is INTEL will still have stability problems, because they rushed these dualcore stuff onto market just to show they have something, while AMD has designed them in 1999.
      0 Not allowed!

  23. #23
    Join Date
    Dec 2004
    Posts
    256
    The thing you seem to fail to realize is this isn't all "new technology" that intel "rushed out the door". If you follow intel at all you'd realize they don't release a new technology until it's absolutely necessary. They milk the cow for all it's worth, it's called sound business practices. I guarantee intel didn't just "slap together a dual core" mainly because that would be basically impossible. Not to mention stability is the name of the game with intel. I can basically guarantee you as well that no cooling on that northbridge is probably the main cause of the "reboot issues". My a7n8x *shock* *AMD* board is the exact same way. Before I put a fan on the northbridge it was down about once a week.
      0 Not allowed!

  24. #24
    Join Date
    Aug 2004
    Posts
    371
    But they already add a fan to the northbridge, see
    http://www.tomshardware.com/cpu/20050603/index.html

    So far, there is no conclusive evidence the fault the CPU on the crashes, the INTEL cpu is quite smart in doing clock throttles and other self-protecting acts. However, from this test, we see another advantage of AMD64 architecture: AMD64's northbridge is inside the CPU. INTEL systems may need active cooling of the northbridge, which generates additional heat, so INTEL system power consumption is CPU+ NB + Others, in AMD64 systems, you have CPU + Others.
      0 Not allowed!

  25. #25
    Join Date
    Dec 2004
    Posts
    256
    Which again, boils down to business sense. Intel makes their own northbridge controllers, another source of revenue. Why integrate it when you don't have to and can make more money off of an off-chip solution? It may not be the most effective, most efficient, or the best, but it makes the most money

    Not only that, but you're comparing two different generations of chips basically. When intel starts rolling out the new Dothan's based off their Pentium-M's this will be much more relevant. Until then it's pointless. You're comparing the last generation of Intel to the current generation AMD (even if that's what's selling). Which you COULD attempt to claim is AMD being suprerior, but if you did a little research you'd quickly find the biggets setback to intel is that IBM refuses to license SOI to them, which is basically a ginormous setback. I have a feeling that since IBM is basically in neck deep to Apple for not producing with the Power series that intel will be given rights to something along the lines of SOI, which will catch them up overnight (assuming they don't already have something in the labs, which would shock me).
      0 Not allowed!

  26. #26

    Red face

    seraph1 its a lost cause man... I'm not even going to bother on the topic anymore as its just the same thing over n over...

    Seems like he's here more to dictate to people, rather than discuss and learn... wm2100 won't make $$$ in the real world with his current way of thinking and his approach towards hosting in general....
      0 Not allowed!

  27. #27
    Join Date
    Aug 2004
    Posts
    371
    Originally posted by seraph1
    Which again, boils down to business sense. Intel makes their own northbridge controllers, another source of revenue. Why integrate it when you don't have to and can make more money off of an off-chip solution? It may not be the most effective, most efficient, or the best, but it makes the most money

    Not only that, but you're comparing two different generations of chips basically. When intel starts rolling out the new Dothan's based off their Pentium-M's this will be much more relevant. Until then it's pointless. You're comparing the last generation of Intel to the current generation AMD (even if that's what's selling). Which you COULD attempt to claim is AMD being suprerior, but if you did a little research you'd quickly find the biggets setback to intel is that IBM refuses to license SOI to them, which is basically a ginormous setback. I have a feeling that since IBM is basically in neck deep to Apple for not producing with the Power series that intel will be given rights to something along the lines of SOI, which will catch them up overnight (assuming they don't already have something in the labs, which would shock me).
    Dothan? INTEL still has not figured out how to do 64 bits on their mobile chip. Yonah won't have 64 bits. Wait till 2006 and see if INTEL engineers can figure out how to do 64 bits on Pentium M. Before AMD found out how to do 64 bits on x86, INTEL already had 4 teams work on P3 and concluded that it is impossible to extend it to 64 bits.

    INTEL's problem is that they have no innovation, they are just copycating AMD nowadays, and they are not even good copycats. So far, they have only sucessfully copied AMD64 instruction set to the P4, but p4 is dead and INTEL is start afresh.

    On DCA, INTEL is probably 3 years behind AMD, INTEL will not have something similar to hypertransport until 2007.

    On multi-core, INTEL's next generation dualcore is primitive compared to AMD's and AMD will have quad-core then...
      0 Not allowed!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •