http://www.neoseeker.com/resourcelink.html?rlid=131317
4 phase ;)
So far I got 410mhz/ddr820 with 1T. 400mhz 100% 32m and 3d stable.
http://www.neoseeker.com/resourcelink.html?rlid=131317
4 phase ;)
So far I got 410mhz/ddr820 with 1T. 400mhz 100% 32m and 3d stable.
Now this is interesting. I tested 3-3-3-8 1t 363mhz vs 4-4-4-12 2t 483mhz at 2900mhz CPU speed in 8m and 3dmark and the results are quite surprising:
http://afireinside.evilspork.net/AFI/AM2/3338-8m.JPG
http://afireinside.evilspork.net/AFI/AM2/44412-8m.JPG
Leads me to believe there is around a 100mhz difference in pi between the two sets of timings... Sounds reasonable. On to 3dmark!
http://afireinside.evilspork.net/AFI/AM2/3338-01.JPG
http://afireinside.evilspork.net/AFI/AM2/44412-01.JPG
At first I thought "damn nice car low or something" then I saw improvements across the board. Very strange...
Stock X800pro if anyone is wondering. Win XP and cat 6.5s.
Ok so.
Sandra prefers Cas 3 1T over Cas 4 2T.
But 3dmark prefers Cas 4 2T over Cas 3 1T.
Too bad I care more about Sandra :p: Nice score for stock x800 pro :) Those scores used to be for stock X800XTs on stock FX-57s :)
Perkam
wow at 3dmark scores, its interesting that 3dmark likes mhz more than timings
1.625v running flawlessly :) Almost had 2.95 through 8m but it crashed as the final loop started "not convergent in main" Doing 2932mhz now, probably my max on this heatsink.
Very nice comparisons AFI :up: Same conclusion I'm coming up with after playing a bit with my AM2 setup - sheer memory speed is what counts - DDR2-1080 8-4-4-4.0-2T wipes the floor with DDR2-900 8-3-3-3.0-2T :fact: Just like in your comparos @ 2900MHz :)
Good stuff with testing :toast:
The BIOSTAR I used to build my dads AM2 system had a 4-Pin Fan connector as well...I am not sure what they plan to be using...Quote:
Originally Posted by afireinside
Did you notice how tiny the boxes are for AMD now...I ordered a 3500+ and it was so tiny...
Thanks :)Quote:
Originally Posted by bachus_anonym
Tight timings with 2T seems kind of useless. They can hold their own but only at 1t in pi. 3d is a slaughter. I have not been able to pass DDR820 @ 1T for the life of me. These sticks can't do over DDR1020 at 4-4-4-12 either and tras = 8 at ddr1000 gives me an insta lock.
http://afireinside.evilspork.net/AFI...7-32m-band.JPG
Vcore = 1.65v
Vdimm = 2.3
That's all for tonight. I'll try to get some better cooling tomorrow. Kind of odd that 2947 passes 32m and 3004 failed 1m first loop :(
4pin fan-connector is used on most btx boards. the fourth pin is to pwm-controll the fan-speed.
Go for 8k :D
Perkam
So what`s the max HT for your board afi ?
I'll test that today
While the forums were down...
http://afireinside.evilspork.net/AFI/AM2/571-32m.JPG
http://afireinside.evilspork.net/AFI/AM2/571-3d01.JPG
Vcore = 1.65
Vdimm = 2.5
Need a vdimm mod :(
Max HTT was around 350-360. This board is no HTT champ.
Still can't get over 2.95ghz... Tried 1.675 and 1.7v but it round errors half way through pi. Not sure what's up with that considering 2.95 is 100% 3d and 32m pi stable.
Same ram clocks down to 4-4-3-8 :D Working on tweaking 32m right now. CPU @ 2.85ghz 1.65v is loading 60C during the last few loops of 32m... No wonder I can't hit 3ghz.
Not a champ, but certainly not a looser.
350-360 is enough to OC any CPU to the limit with air/water.
thanks for testing that :)
afi , try it with dice
I will soon, need a hold down.Quote:
Originally Posted by Revage
Nice work AFI. Looking forward to that DI action...3 inch mousepot? :toast:
I have run sandra at 306fsb with the foxconn, but not higher. I'll be doing some head to head comparisons with the Asus am2 board in the next week or two. If the Asus can run some mean timings at high fsb with no CB issues, it should beat these results easily.
Best sandra results on foxconn/fx62/corsair C5 2.35vdimm cpu@3350mhz:
http://pic20.picturetrail.com/VOL256.../158975447.jpg
7k unbuffered
http://pic20.picturetrail.com/VOL256.../158975444.jpg
Nice stuff :D I wonder if I can get my tras down to 5... I'll have to try that.
Yeah 3" mousepot. Jason said he has some delrin AM2 hold downs made but I haven't caught him on AIM recently so I'm just going to figure something out over the weekend.
I look forward to what you do with the ASUS board. If Shamino's ram clocks mean anything it should be a monster!
Machinist problems I think is what he said.Quote:
Originally Posted by afireinside
For me, when I was benching 3d and trying to put am2 at the top of the orb, the Foxconn proved to be better. I had way more issues with the mem controller on the chip with that asus mobo when trying to run really tight timings(3-3-3-5-1T-12trc) + lowest die temps on the chip.
It's one thing to run SP sandra etc all day on air or chilled water with low cpu speed(33-3400mhz)..but as soon as you push the chip hard over 3600mhz for 3d or heavy load, especially one with a sh*t cold bug, the rules change. High fsb doesn't matter anymore at all. Only thing that matters is how tight you can get the ram/chipset timings at 210htt or lower:D
Chipset timings? What are they? I've only been playing with the ram and drive strengths.
What speed do you run the SPP<->MCP at? Does it make a difference?
http://afireinside.evilspork.net/AFI/AM2/IMG_0711.JPG
http://afireinside.evilspork.net/AFI/AM2/IMG_0715.JPG
Orleans is under water :lol: Doin it AFI style. I'll rig up the water setup after I get back from bestbuy for some ice water benching @ 3g+
Oh yeah forgot to save the pic to my thumb drive but I have a shot of 32m @ 571mhz ram 4-4-3-5 TRC=11 :D
haha nice mounting, cant wait to see those results
cant wait to see the action