Dual CPU Benchmark scores suck real bad
Home | Reviews and Features | Special Reports | Forums |

Results 1 to 9 of 9

Thread: Dual CPU Benchmark scores suck real bad

  1. #1
    Join Date
    Aug 1999
    Location
    North Carolina
    Posts
    1,750

    Dual CPU Benchmark scores suck real bad

    I'm not satisfied with the numbers I'm seeing on CPUmark99 and SANDRA. I can't say for sure I'm getting an accurate reading from these two benchmark apps, either. I'll start off by telling you an embarrasssing truth about my dual 624's. The CPUmark99 score is 45.5 when running Win2000 and using both processors. Pretty dismal, huh? But when I boot into Win98 the score goes up to 47 with just a single Celery. That's hard to believe. And SANDRA scores have a similar characteristic, as well. Hell, the system FEELS slower when running Win2000. I have to say something in defense of the dual CPU low scores, and that is the score of 45.5 is ALWAYS reported by CPUmark99. Even when I run several applications WHILE I'm running the benchmark program, the score is at least 45.5. It's almost as if it keeps half its power in reserve, and doesn't like to use it. Nothing slows it down, but it still doesn't seem fast. I have to set my BIOS to the MPS Version Control 1.1 setting, because Win2000 acts funny when I set it to 1.4. Win2000 supports 1.4, but I can't get it to work. Is that the reason I'm seeing low scores? because version 1.1 is less efficient at CPU utilization? Please tell me that's all it is. Hey BP6 owners! What are YOUR scores?
    P4 2.8@4.0 dual R134A phase change chillers with 600watt thermoelectric array. Capable of -70F
    Nixie tube numeric display control panel and pneumatic motherboard tray.

  2. #2
    Join Date
    Nov 1999
    Location
    San Leandro CA
    Posts
    681
    Hey TNproud2b you get my email.

    ------------------
    You did what to who's dog for how many jellybeans?
    You did what to whos' dog for how many jellybeans?

  3. #3
    Join Date
    Aug 1999
    Location
    North Carolina
    Posts
    1,750
    Yes, in fact I just looked. Thanks a lot for the offer!
    P4 2.8@4.0 dual R134A phase change chillers with 600watt thermoelectric array. Capable of -70F
    Nixie tube numeric display control panel and pneumatic motherboard tray.

  4. #4
    Join Date
    Sep 1999
    Posts
    1,249
    I found it to feel slower than Windows 98 also. Even though 2000 utilizes 2 processors. It did have the feeling of raw horsepower. However, Windows 98 felt snappy and Windows 2000 just, well, just, well, just is there.... I am wondering if I should have overclocked an Athalon 500@700 instead of dual 366@550. I agree what kind of performance are you guys who have your systems up and running properly getting with normal programs? I haven't had enough time to tweak yet so maybe that's it.

  5. #5
    Join Date
    Jan 1999
    Posts
    203
    I hate to be the bearer of bad news, but dual CPU's on a home desktop system will give you no performance increase over a single CPU. The Intel GTL bus can only access one CPU at a time, not both simultaneously. This is likely why your benchmarks are so low. The Athlon EV6 bus can access both CPU's simultaneously and will provide a much larger difference in performance. If you want some proof, go to Thresh's Firing Squad and read the article they posted about two months ago, comparing single and dual CPU systems. In all but one or two benchmarks, the single CPU system was as fast or faster than the dual. In the tests where dual was faster, the difference was so small it would never be noticed. If you want a dual CPU system that will rock, get yourself a single Athlon system now, and upgrade to dual CPU's when the motherboards are available in 2000. A single Athlon will be much faster than any dual Celeron or PIII system anyway.

  6. #6
    Join Date
    Aug 1999
    Location
    North Carolina
    Posts
    1,750
    I agree that an Athlon system would be faster, but the chips don't cost $70 either. It ain't real impressive to make a fast cpu go faster. I'll throw in something I should have mentioned in my post, using dual processors results in SETI processing times that are exactly twice as fast as a single Celeron. Explain that. While it may not access these 2 chips simultaneously, there is indeed a performance gain.
    P4 2.8@4.0 dual R134A phase change chillers with 600watt thermoelectric array. Capable of -70F
    Nixie tube numeric display control panel and pneumatic motherboard tray.

  7. #7
    Join Date
    Sep 1999
    Posts
    1,249
    I usually upgrade maybe once every 18 months or so. This machine is really really fast any how. I'm sure this would be some peoples dream machine. Think of all the 166s 350s and so forth that are out there. You could go broke upgrading every month when they come out with a new processor or motherboard. My Abit BP6 is a great motherboard. By the way I heard the newer Celerons have the SMP disabled again. How long will it take for someone to figure it out once more. Let's try Linux or something. Anyone know what the best version is to get.

  8. #8
    Join Date
    Oct 1999
    Location
    Anchorage, Alaska
    Posts
    19
    Hi TNproud2b,

    I just ran SANDRA 99 Standard on my dual celery in Win2K and it warns you in the beginning that this version is only for Win95/98. If you go on and perform the tests anyway, the tiny notes at the botoom of the page say that while my system has two processors, the tests are run as a single processor mobo.

    That may account for the SANDRA scores but I don't have CPUmark99 so I can't give any figures for those benchmarks.

    Don't know if this helps....

  9. #9
    Join Date
    Aug 1999
    Location
    North Carolina
    Posts
    1,750
    Thanks Doug, you're absolutely right. There is a warning at the bottom of the window that clearly says "results for single CPU". Thank God my dual system isn't as slow as it said.
    P4 2.8@4.0 dual R134A phase change chillers with 600watt thermoelectric array. Capable of -70F
    Nixie tube numeric display control panel and pneumatic motherboard tray.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

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