Jump to content

scyphe

Members
  • Content Count

    242
  • Joined

  • Last visited

Everything posted by scyphe

  1. What did you have before? CS: Source (and 99.9% of the games out there) only use 1 core, and @ 2.4Ghz it's ~98% of the speed of a single core Athlon64 (3700+). It's also possible that there's some bug that have to be fixed somehow. There's a lot of research needed before jumping into a new platform (dualcore opterons from whatever single core platform you used before).
  2. If chip is disappointing I'd just sell it and buy a known stepping on e-bay (if economy permits). I'm waiting for my 165 to arrive and it's apparently a CCBBE.
  3. May simply be that you've hit the limit on your chip.
  4. Yeah, great OC for a Sempron 2800+
  5. Oh, sorry, missed that part about the sandy..
  6. check this thread: http://www.xtremesystems.org/forums/showthread.php?t=96900
  7. My guess when it comes to differences like that is that the silicon source (the big glassy looking thing) got used up and they mounted a new one to slice up for new die platter. Theory: perhaps the last bits of those silicon "sausages" are of better grade than the rest of the silicon and the last couple of slices provides the highest quality dies??? Just a very uneducated guess, I don't even know what the stuff is called. Read an article about the process but that was yeeeears ago (the Pentium 1 days)..
  8. AND.. I don't know about ATI but nVidia is improving the dualcore support in their forceware drivers with every new release.
  9. Done.. have something else going instead..
  10. Gah, I hope I don't win that bid then.. lol according to that thread the CCBWE 0607UPMW's suck donkey...
  11. Heh, okay, I was just going on the ebay auction description (US Only) I'll see what happens with the stuff I've put a bid on. May get back to you if you still have some chips left..
  12. Nope.. I checked out your items but you only send within the US..
  13. This will be interesting, I'm currently bidding on an Opteron 165: CCBWE 0607UPBW. I sure hope it'll reach ~2.7Ghz...
  14. Time to RMA the motherboard methinks.
  15. I can't see how enabling SLI-jumpers would make any difference since we're talking about two completely different ATI-cards from different generations.
  16. If you get into bios, change the startup voltage to 1.500 volts. If it's a severe case of cold-boot bugginess that may help. Kick up the voltage for the RAM one notch as well. And drop the LDT ratio to 3x.
  17. Put the RAM modules in the ORANGE slots, not the yellow ones.. I couldn't boot after putting my machine together and I spent hours on trying to find out what was wrong. Formatted and installed 2K Pro on my old P2 300 server (ahem) and managed to get online and search the threads here @ dfi-street. And yes, I put the modules in the orange slots instead, removed the battery for 15 minutes, put it back and the machine booted.. Good luck
  18. Unless you're in dire need of that x600 SE, throw it out.. If windows is using that for gaming the performance is pretty much expected. It would be a real waste to use the x1600 Pro for an extra monitor and the x600 SE for gaming. Do you know which is master and which is "slave"? And which one of them are used for gaming?
  19. Interesting. After 2-3 hours of messing with the settings and running the built-in memtest86+ I finally figured out what caused the error at the exact same adress at the exact point of the test. The BIOS I had was OCZ Tony's 702bt2. I set everything in bios for the RAM to auto, removing the custom values and voilá. No errors. Another thing in the 702bt2 BIOS is that there was no 150 divider. Sometimes it showed up but it worked as 140 or 166, can't remember which. Flashed to 623-3, set everything to my o/c settings and no errors whatsoever. I'm only running RAM @ stock with 1T getting 5.5Gb/s, but it's better than knowing that there's an error while 300mb/s faster.
  20. The difference between 1T & 2T with the exact same settings otherwise is 1.1Gb/s. This is in Sandra, it may not be noticeable otherwise. Still, it's annoying as hell. When I've got some money to spare I'll buy some good, overclockable RAM. I'll have to RMA these though, can't sell RAM that may/will crash a computer at random times just because it can't run at it's default T1 setting.
  21. I just upgraded from 1Gb to 2Gb (sold my 2*512mb TwinMOS and bough 2*1024 TwinMOS), not particularly O/C friendly but it's okay, CPU-speed makes a lot more difference than RAM going from 200->230Mhz. However, the new RAM is probably going back on RMA. SPD for it is CL3, 3-3-6 1T. However, when I use 1T and run the built-in Memtest86+ there's a single error at the same address, same bit. If I use 2T there's not a single error even if I put CL to 2.5. But booting into Windows gives me 4.7Gb/s speed while 1T gives me 5.8Gb/s. That's a huge difference just for 1T. Doesn't matter which voltage I use (tried 2.7v->3.0v), I always get the same error at the exact same address after a while. Running windows doesn't show any problems of course, the chance of crucial code to hit the adress @ 1786.6Mb is a longshot. But I'm going to do some more testing and see if I can hit a combination to get rid of that error. But it does seem like one of the RAM-sticks are bad since the RAM is supposed to run @ 1T. The RAM is running @ 200Mhz right now, CL3-3-3-6 and I've tried different LDT's and LDT multipliers. Sorry, just wanted to rant about being annoyed, one single bit/byte out of 2Gb that is giving errors.
  22. If CPU is DOA the 4 leds shouldn't all be lit.
  23. Did you try your RAM in ONLY the orange ram slots, not the yellow ones. I couldn't (can't) boot my board with ram in the yellow sockets.
  24. Actually, it's a good thing I have to point my sink "up" since the CPU-fan is sucking air through the copperfins instead of blowing down on them. That way it will suck air past the RAM that get somewhat cooled.
×
×
  • Create New...