2008 XS Cheapazchips 3D challenge

[T_M]

Dino, when your PCB is running at subzero temperatures, covering the RAM aint a problem.
Anyway, seems my samsung are letting me down, but i'll happily take the crown for record GPU (1053MHz)
First sessions today on LN2, planning another go at it tomorrow but the RAM is just a let down. It really will not budge past 513MHz, even with an extra 0.1V (which shows its not a temp problem Dino )
GPU pot temp was kept at almost exactly -70C (but -65C was fine too) for the max clocks today.
GPU = 1.93V idle, MEM = 2.15V

T_M 26191 ANYTHING GOES (1053/2106/513) http://service.futuremark.com/compare?2k1=9339088

T_M 9577 ANYTHING GOES (1053/2106/513) http://service.futuremark.com/compare?2k3=5742854







http://service.futuremark.com/compare?2k3=5742854


http://service.futuremark.com/compare?2k1=9339088

[dinos22]

nice man

have you tried higher shader

[T_M]

Nah today was just getting a feel for things, tomorrow ill push core/shaders as well as tweak CPU/RAM a bit. So far i didnt have any problems at all running 1.93V idle (~1.96V loaded).
Testing some high MHz Patriot as we speak.

If anyone has some '01 order and LOD tips they are willing to share, i'd love to hear it

[zakelwe]

Nah today was just getting a feel for things, tomorrow ill push core/shaders as well as tweak CPU/RAM a bit. So far i didnt have any problems at all running 1.93V idle (~1.96V loaded).
Testing some high MHz Patriot as we speak.

If anyone has some '01 order and LOD tips they are willing to share, i'd love to hear it

Hi T_M,

Excellent core clocks there : toast: nice to se LN2 make an appearance.

I'll put your scores up later.

For 2001 I use 4.3 for nature and 2,8 for all other tests, seems to work well for my combination.

My card, after last nights struggles, will currently only do 2.5D rather than 3D Might be a "throw in the bin job".

Regards

Andy

[SpikeP]

back again to claim the top 3dmark03 spot again!

need to get a new cooler for the core now. Little lump of aluminium isn't cutting it anymore.

Anyway, quick run after work. memory finally made 600mhz stable after changing the voltage up and down to find a sweet spot of 2.15v. What kind of voltage is everyone else pumping through them?

core was 770. shaders were 2000. still need to get that shader one higher. New cooler and see what i can do.

Link to ORB: http://service.futuremark.com/compare?2k3=5743439

[STEvil]

I think the vreg circuits are holding us back for the memory.

Wish I had a 2v-3v variable power supply to test that theory with..

[STEvil]

craaaaaap

Sparkle 8400GS is 64-bit lol.. stupid people advertising the 128-bit HDR!

3ns Elixer ram on it.

[LyP0]

T_M : ramp up the shader clock. the card is hugely criplled right now. My card did need a shader clock of 2700+ @ 1100mhz gpu speed. The diffirence between 2200mhz and 2600mhz was almost 10% so that means you could easy get 11k to crack.

@ people in holland. I've got an 8400gs with a power management that can handle 2.78V ( but the card is defenately dead, the core was cracked of the pcb when some condensation between the core and pcb froze :@ ) I'd be happy to send you the card so you can use the powermanagement.

[STEvil]

This 3ns Elixir looks to be cherry, go figure. http://sg.vr-zone.com/?i=3116

MVKTech claims 1000mhz near the bottom of the page... ! http://www.mvktech.net/content/view/3800/39/1/2/ and then prove it http://www.mvktech.net/content/view/3800/39/1/7/


Damn, 500mhz on 3ns chips at stock volts.. watch out people lol!

[T_M]

Great thanks for that LyPo, so it sounds like my GPU clock wasnt the highest ever done. Whats the WR clock so far?
Thanks for sharing your LODs zakelwe, truly in the spirit of this competition of sharing
Yesterday i ran LOD 15 for 3D'03, and LOD 5 for all 3D'01

[G H Z]

Yeah crank on the shaders T M, not as big an impact as ram clocks but definitely worth it. Try LOD 3.6 for '03

[T_M]

Will spedn the afternoon with her and see if she can crack those top spots

[zakelwe]

Good luck T_M! Your's and SpikeP's scores added this morning.

SpikeP, have I missed your 2001 score, or are you concentrating on 03? Let me know if I have slipped up.

It would be nice to have T_M's core speed and Spikep's ram. Getting towards 29k and 12k I think.

Regards

Andy

[G H Z]

Getting towards 29k and 12k I think.

Soooooo...you really don't have a 30K backup

There may be hope yet

[zakelwe]

Soooooo...you really don't have a 30K backup

There may be hope yet

No I was trying to bluff you into either giving up or blowing up

Regards

Andy

[SpikeP]

Good luck T_M! Your's and SpikeP's scores added this morning.

SpikeP, have I missed your 2001 score, or are you concentrating on 03? Let me know if I have slipped up.

i am just concentrating on 2003. Need to make some time to get some more results done. Looks like wednesday maybe be the last day i do some benchmarks on it. Need to get onto XP as well!

- Paul

[STEvil]

Great thanks for that LyPo, so it sounds like my GPU clock wasnt the highest ever done. Whats the WR clock so far?
Thanks for sharing your LODs zakelwe, truly in the spirit of this competition of sharing
Yesterday i ran LOD 15 for 3D'03, and LOD 5 for all 3D'01

Aw man, cranking LOD makes it no fun!

Suppose I could try increasing mine some so my clocks perform for what they are lol (been running 3.6 only on all benches).

If I get the overcurrent mod figured out for sure on my asus card I might break 1000/2600 on water

[Rol-Co]

Great thanks for that LyPo, so it sounds like my GPU clock wasnt the highest ever done. Whats the WR clock so far?
Thanks for sharing your LODs zakelwe, truly in the spirit of this competition of sharing
Yesterday i ran LOD 15 for 3D'03, and LOD 5 for all 3D'01

still some clocks to go T_M

[T_M]

Pushed it further today and broke the 10k barrier. I tell ya it was really hard work with this RAM holding it all back. Temperatures and pouring really played a big part.
vGPU = 2.12V idle, GPU pot temp = 50C, vMEM = irrelevant
Next session ill push the vGPU higher, for now i was trying to play it "safe".
Focus has purely been on '03, and '01 was just an afterthought.
LOD = 3.8 worked well, definately 15 is and got me 100pts less.
I have run higher GPU MHz (1080) but the score was a couple points less.

10015
http://service.futuremark.com/compare?2k3=5744732


T_M 10015 ANYTHING GOES(1066.5/2646/513) http://service.futuremark.com/compare?2k3=5744732
T_M 26889 ANYTHING GOES(1053/2646/513) http://service.futuremark.com/compare?2k1=9339836

[dinos22]

nice work T_M

bummer Feb all out and i haven't done any benching yet lol

[G H Z]

Nice clocks T_M & Rol-Co

No I was trying to bluff you into either giving up or blowing up

Did some benching outside Sat night, with cold air and more voltage I managed a decent '01.

28,462 (972/2484/551) @ 1.87V

[zakelwe]

Nice work G H Z on top spot and T_M nice jumps up as well

Corrected Pauls 03 score which was right postion but his old score still.

I still have a feeling that we may see some late entries ( dino ?? ) and a dark horse come up on the rails, so keep benching ( as Brucey would say).

Regards

Andy

[Johnny Bravo]

I still have a feeling that we may see some late entries ( dino ?? ) and a dark horse come up on the rails, so keep benching ( as Brucey would say).

You Know nothing of my work Andy

[Rol-Co]

Those clocks are not mine GHZ, was somebody playing dry ice .....the max clock i've seen on the core.

[T_M]

Suprisingly the cold bug on these "beasts" gets warmer as you put more volts/clocks thru.
By the end of my session i was upto -50C pot temp coldbug, and also rate of temperature change would affect bugging too.
Also, my mosfet sinks are still not geting warm.