TerraRaptor

Hi all we have 3 categories of 8600GT here @hwbot: http://hwbot.org/hardware/videocard/geforce_8600_gt_128mb_gddr3/ http://hwbot.org/hardware/videocard/geforce_8600_gt_256mb_gddr3/ http://hwbot.org/hardware/videocard/geforce_8600_gt_512mb_gddr3/ As far as I understand the only difference is memory size. Can I submit results with 128mb card to 256mb category? I beleive these cards may be merged as they are quite the same and i don't want loose hwboints just because someone thinks that 256M cards have an advantage over 128mb cards in memory intensive tests.

Ohuenno, bro! Congrats to reunited ZOT and personally to you, GraduS!

I broke the pins on processor (and soldered a bridge to ak4 spot on processor) - so it is not an issue. Probably it is socket that saw too many cpus - sometimes i need to resit processor in socket several times to make even coppermine to work.

Thanks for the answer but I actually did the same mod and tried even more (G35-G37 bridge, isolation of Detect etc) - no go. With Evil Inside bios too. All voltages are in place (vcore, vtt, vnb, vagp, vdimm) - but still black screen. Probably should do a more complex connection of AK4 either to VTT or PWGD with resistor and capacitor.

Did you mod cpu or mobo to support PIII-S? I failed to make PIII-S with mods run on CUSL2, although it still works in TUSL2. Modded Celly 1.1 1.5v works fine. But modded Celly 1.1 1.475 doesn't. Dunno what to do - don't want to mod motherboard.

Da BEST!

Money in the game? This will bring out too much shit and will destroy a spirit of old-school everclocking "not for boints, just for fun".

+ D3D.

well done bro)

No problem, Caos Wolf) we all know sdr binning is a problem. We just need to find golden chips - and probably sit them on good pcb.

everyone here gave a try to hundreds of sticks I think) M.tecs suck so we need exactly 0103 to give it a try.

>pc100@219 2 2 2 7 9...bh-5?lol Why not? My pc133@202 2-2-2-7-9 or Gigio's pc133@210 2-2-2-7-9 would be also "hard to beleive" 1 year ago.

I hope not to hear cool story next time that you had 250 2-2-2-5-7 memory, unlocked cpu with x12 multi being capable of 3ghz and sub 50s result) Anyway this result is the best here (as well as pifast dd yesterday).

I'm dead sure me and GraduS will break 1m in next 3-4 months using 1:1. Gigio already has memory capable <1m.

easier voltage regulation (stock vrm allows up to 2.05v with pinmod and up to 2.5-2.7v 2 with complicated VR mod) and probably some boost with long benches (32m, 1024m) - signle phase vrm overheats quickly with >2min load. I didn't notice any gain even with cpuz so far if compare modded stock vrm vs epower.

Photo 3 - Asus ROG OC Panel - acting like thermometer (left for cpu, right - for memory) Photo 4 - top view of the system with EVGA Epower soldered instead of stock vrm.

nice John!

Well done bro. Can't beleive you didn't bench 06 for globals)

powerful sub!

stupid challenge

Good cpu bro!

well done bro!

nope. With proper waza this will brake 1.02 for sure and probably 1.01. My estimations are that I need 1.4S @200fsb to break 1m. This setup is really close to my estimation in terms of performance. Proper waza + well tweaked os + known tweaks to superpi + tweaks to i815 are way to go below 1m with this setup. I had 1.03.5xx @194.5mhz with 1.4S. Almost 10% difference in ram should be more than 500ms and 70 mhz in cpu.

nice result, but why no tweaks?

okay, they then need be warped in dragon butts skin. Probably this may be of help)