Jump to content
HWBOT Community Forums

TerraRaptor

Members
  • Posts

    1661
  • Joined

  • Last visited

  • Days Won

    23

Everything posted by TerraRaptor

  1. Seems like Gorod was right saying that PCIE may influence the clocks on nv43 (http://www.xtremesystems.org/forums/showthread.php?248552-6600GT-834MHz-gt-03-05-06-WR&p=4321051&viewfull=1#post4321051) - I did a quick test myself dropping bclk with 2500K from 104 to 100 and card lost stability. Probably nv43 needs a combo of temperature, voltage and pcie clock to show good gains. For me, going cold gained at least 100MHz.
  2. Data in your table displays just a normal variance imho. I think after 100 runs both datasets will average to the same values.
  3. I estimate power consumption as 100W here, vdrop is not big (i actually did a pretest with single pair of wires, then soldered another pair - no magic difference even with single pair of wires). Capmod didn't help either. These cards seem to prefer cold a lot. Tested a total of 3 cards (2 asus with infineon 2ns, 1 giga with samsung 1.6ns). Important note on 6600GT (and other cards with ISL6534) - you better epower both vgpu and vmem as ISL6534 controls them both and likes to trigger protection if only single section is epowered. ISL6534 also controls PCIE 1.2v voltage but epowering it or changing it didn't help at all - so can be left at stock.
  4. Should have MMX only option for oldest platforms (sdram/ddr). Run time of around 50-60 seconds maximum no matter what speed is (so it is not "how much time it took to finish job" but "how much of the job was done"). Probably a separate workload for different DRAM technologies - to be able to address specific platform issues more easily? SDRAM+DDR (MMX only available)/DDR2+RDRAM (SSE2)/DDR3 (SSE3)/DDR4 (AVX+)/DDR5/DDR6. Anyway, the most valuable thing should be consistency, repeatability and coverage of most hardware.
  5. right, not the best one - wasn't able to quickly boot 540x7. Batch L734A770, Malay
  6. it exists in x48 ddr3 too. 2:3 strap 400 is displayed as 3:5 too. It is common CPUZ bug I believe for all P45/X48 boards (and I assume X38/P35).
  7. Contact UL support. Why asking here without error screenshot etc? PS. Leeghoofd proves to be great person again - I would ban you to save time and nerves.
  8. Cheap dc/dc converters rule nowadays (i believe these were not easily obtainable those days - so ocz booster or one from hipro5 were things to consider), so I don't even have a tiny will to own ocz booster:) it will make mount of ss more difficult too.
  9. Good score. Another reminder AMD was doing great at prescott days.
  10. Can't use early catalyst for newer cards anyway - looks like 3850 first appearance was in cat 8.6 and 4670 in cat 9.6:)
  11. I was expecting the last two registers (like 16 18 or 26 18) to be drive strengths for receiver/transmitter aka cpu side and NB side. I have killed two cpus when experimenting with that part of SIP, though should re-validate it by killing another cpu setting it FF FF.
  12. I tried several ones from your server - quite same bus limit of 250-255 for cpu confirmed to run 275 bus 1m aircooled. And the higher the bus the more likely usb will drop (os still works). I tried agp 4x, disabling FW and SBA - no go. AGP clock doesn't matter too. Seems like excessive bus load by 3850, probably need to try agp drive strength in powerstrip as Antinomy suggests.
×
×
  • Create New...