Jump to content
HWBOT Community Forums

Masterchief79

Members
  • Posts

    242
  • Joined

  • Last visited

Posts posted by Masterchief79

  1. Hi guys,

    searching for rather exotic or rare GFX. I live in Germany, so offers from EU are probably favored :)

    Thinking of something like:

     

    Asus 7800GT Dual

    GTX480 Lightning

    GTX580 Lightning XE (3GB) / EVGA GTX580 Classy / Asus GTX580 Matrix

     

    Dual X1900XT

    Sapphire HD2600XT2 (dual HD2600XT)

    maybe Asus Mars and Ares, depending on the price

    and engineering samples and/or models that never made it to the retail market (HD2900 XTX for example).

     

    Greetings,

    MC79

  2. So I found a similar thread for a GTX780Ti DC2 (will search link if someones interested), where an ASP1212 was modded.

    I compared the datasheets and those 2 controllers (CHL8318 and ASP1212) are absolutely identical. I mean, they even just copy-pasted the text in the datasheet.

    In that thread, K404 just soldered to pin 4 (VCPU) and it worked, so I did the same, and it worked, too ;)

     

    Just running 1,28V without further adjustments (used 50 Ohm potentiometer for vGPU), resistance went down from 8,7 Ohm to 7,5 Ohm. GPU Temperature 53°C @ air, VRM 45°C @ air.

    Would take photos but have no cam here.

  3. That eVc thing looks interesting, you know where I can get one of those Roman?

     

    I'm used to FB Mods so that wouldn't be the problem, it's rather that I'm not familiar with the pinout of this controller and Kepler cards in general. I have no idea if you can just solder happily on the IC and the card boosts still as it should etc.

    Gonna measure the resistance now and edit it any second.

     

    Pin 4 to GND = 7.2 Ohm

  4. Hey guys,

    wanna ask for help with my Gigabyte GTX680 SOC.

     

    726665d1395829584-voltmod-diskussions-und-laberthread-img_9878.jpg

     

    vGPU-Controller: CHiL CHL8318 (http://www.irf.com/product-info/datasheets/data/pb-chl8318.pdf)

    vMem-Controller: Texas Instruments M430F5510 (full name MSP430F5510) (http://www.ti.com/lit/ds/symlink/msp430f5510.pdf)

     

    vGPU is most important. It's not a bad chip on that card (1,4GHz@1,212V at air through 3DM11), so I wanna see what's possible with voltage tweaking.

    Any idea on how to vMod that CHL8318 though?

  5. Indeed, it does. The original AM3.exe always works, unless some drivers aren't properly installed or the hardware just isn't capable of running the bench (not DX6-compatible). So that's fine there. Just the wrapper reports the Non-Default-Settings Error with some older cards.

    If it would help you, I could make a list, I've got quite a few cards here.

    Any chances of getting it to work with newer drivers? And thanks for help so far, didn't really expect an answer so quickly ;)

  6. With older AGP cards I frequently get the "Non-Default-Settings" error which doesn't permit the benchmark to run. Even with freshly installed drivers and default settings there.

    Right at the moment I got the prob with a MX440-8x 64 bit. Some fix yet? Always annoying when you can't complete your benchmark set cause of cranky AM3.

  7. What do you mean with stuffed exactly? If it means what I think - most probably no :D Brocken socket = broken board, I mean you would have like to solder a new socket onto it... I fixed a M4G once with about 20 bent pins once but that's it...

     

    To my rex: Works mostly fine now, some small issues (hangs when I try to save bios settings etc.), but much better than before. I tested for max FSB and without much tweaking or clock skew adjusting I got to about 640FSB CPU-Z.

    Here is the previous E84 32M run: http://hwbot.org/submission/2300403_masterchief79_superpi_32m_core_2_e8400_%283.0ghz%29_9min_21sec_922ms

     

    Got that down to 8m53@5350MHz now, with still 2 cores active, further testing tomorrow (or today - it's 3.30 AM here). And we'll also see if I can break my 5608MHz validation from the RF (2C/2T).

    So, atm I'm happy, apparently got 2 640+ boards :D

  8. Hey, Update!

    As I used my Sandy system (which I normally use 24/7) otherwise, I had to build another system, you know, for gaming etc, from parts that were laying around. I used the REX I was talking about here with a Q6600@3,8GHz and a XFX 5970 BE Limited.

     

    So, a few days ago I broke my M4E (again) and repaired it (again -> this board is awesome, one socket pin broke off, one conductor path burned since one of my pals inserted a ram module the wrong way... and one condensator exploded - repaired still working@5GHz 24/7), so I decided to go back for my Sandy 24/7.

     

    I then used the REX for benching an E7200 office CPU and got to 595MHz FSB, wait whut? Remember, initially I was stuck at 570... Yea kinda funny, maybe the everyday usage did its thing to the resoldered northbridge phase. As the E7200 only has a 266 default FSB and 595 are already pretty high for this chip^^ I rebenched my E8400, 610MHz FSB just working their way through SPi 32M. Can't explain this change, same settings as before.

     

    Concerning ROG boards, I got pretty lucky I guess :D Three killed, three repaired ;)

×
×
  • Create New...