Opteron 165 - Upgrade to X2

CPU by SidneyWong @ 2006-01-17

Last year I had the pleasure in getting my first taste of Upgrade to A64-Venice 3000+ . The high overclock headroom and ease of thermal output plus the low price made the Venice 3000+ a terrific buy. So much buzz on the Opteron Dual Core S939 since November last year; I was hoping I would get the same result my Venice delivers.

  • prev
  • next

What happens?

What Happens:

Like a child, I rushed towards my computer room. Within minutes, I ousted the Venice and popped in the Opteron using the existing Phoenix 70 cooler which has served me well for over two years. Since you all have read all about this dual core, I am not going through the history, specs, FPS, SuperPI and anything alike. What I am going to do and write is simply my experience as a regular Joe in using this processor. At the end, you are likely to tell me what I've done wrong.

I was delighted to see Windows booted up using the identical setting; except I lowered the vcore. Bear in mind, the Venice 300HTT x 9; and memory setting have been running without any trouble since last June 2005.

  • DFI NForce 4 Ultra D
  • BIOS version N4D704-3
  • 300HTT x 9
  • GSkill DDR4400 2x512mb - 2.5-3-3-7, divider 1:1
  • Seasonic S-12 500
  • Hitachi 165 G HDD SATA
  • NV 6200 softed modded to 6600
  • BenQ DVD R/W
  • X-Dreamer II case - 2x80 intake, 2x80 exhaust, 1x80 topblower
  • Chill Vent II
  • Windows XP SP2 + latest updates

    A few minutes later I was greeted by this screen.

    Madshrimps (c)

    When I tried running SuperPI in 2 threads, it failed. I raised vcore to 1.45.

    Madshrimps (c)

    This was the high point, from hereon things would become a bit disappointing. I was unable to get Prime to run stable using two threads no matter what voltage I set. I could get it to run SuperPI + 3DMark benching; Prime95 + SuperPI except 2x Prime.

    2x Prime95 was stable for 15 hours at 2.31 Ghz, default vcore with CPU temp ~44°C and identical PWMIC temp; telling me that the Pheonix 70 is still capable.

    Green Line = CPU temp Red Line = PWMIC Grey Line = Chipset
    Madshrimps (c)
    Click to enlarge

    I blamed the Phoenix cooler; so I switched to the AMD Stock 4-heatpipe cooler and hoped for the best, the fact that the temp only reached upper 40 °C before Prime95 failed meant it couldn’t be temp related. The interesting part is, Prime95 failed at anything over 260HTT regardless of voltage increase. The PWMIC temp worries me when I crank the vcore up slightly, signaling dual cores do run hot. The 4-pipe AMD Stock cooler performs the same as the Phoenix 70 which I still believe one of the best coolers in its size range.

    Insofar, I have been setting the memory timing, divider, voltages other than CPU the same as the Venice 3000+ which was running smoothly and Prime-stable: 300HTT, 1:1, 4X, etc.

    Madshrimps (c)

    Next: more about temps, multi-threads other than Prime95 ->
    • prev
    • next