dinos22

hi guys

I ended up getting a nice Week 17 Venice so my Intel Pentium D Dual Core setup is on the backburner now. I bought a new motherboard to go with my Venice. It's an Abit AN8 Ultra with 3.55vdimm for BH5/UTT fans. I would appreciate your input with my setup/testing/tweaking/all else...

Here is some info on the product:
ABIT's Website: http://www.abit-usa.com/products/mb/...es=1&model=278
Review by Solo: http://www.net-forums.net/forums/Abi...ed-t22067.html

I started testing yesterday afternoon and here are some preliminary results. PLS keep in mind that this motherboard JUST got released and I am one of a very small group of people worldwide that got my dirty fingers on it so bios files are in its infancy......There is a known issue with Venice guys and these new Abit AN8 series mobos so bios updates will have to iron out some of those....hopefully.....One of the well know issues is 1T not working or unstable with Venice CPUs....works great on Clawhammer....another issue is CPU temp reading being WAY TOO HIGH.....

TEST SETUP
A64 3000+ Venice 0517EPMW
Abit AN8 Ultra + Freezer64
1GB Corsair TwinX1024-3200C2v1.5 UTT
1GB Corsair TwinX1024-4400C25 TCCD
XpertVision 6600GT 558/1035
Wavemaster + Enermax 535W EPS PSU


I started to to play with 1GB Corsair TwinX1024-3200C2v1.5 UTT first....the board looped memtest test#5 error free at 238MHz @3.2V which is sligthly higher than DFI and also I used really tight timings unlike on DFI setup.......BUT I ran into memtest errors at test#8 which i originally didn't test for in my last post so i dropped it back to 225Mhz on test#8 and still got errors.....damn.......so i decided to stop messing around and......

.......go back to square one......I started with tictac's suggested timings enclosed below....i am on 9x multiplier though

Memtest 1.55.1 Test 5-----quick 6 loops
Memtest 1.55.1 Test 8-----quick 2 loops
(they are quick loops and once i run into errors at desired speed i will drop MHz back for lengthier tests)

I USED TIGHTER TIMINGS:
* Read to Write Delay: 1
* Row Refresh Recycle Time: 12
ALSO EXPERIENCED PROBLEM WITH 1T so CHANGED to 2T half way through the test....all in notes below...INITIAL TESTS ARE 1T.


TEST RESULTS:

TEST A: 2.8VOLTS on vdimm
TEST1: 200MHz Test5&8----PASS
TEST2: 205MHz Test5&8----PASS
TEST3: 210MHz Test5&8----PASS
TEST4: 215MHz Test5&8----PASS

TEST5: 220MHz Test5----FAIL(errors); Test8----PASS(errors);

TEST B: 2.9VOLTS on vdimm
TEST1: 220MHz Test5&8----PASS
TEST2: 225MHz Test5&8----FAIL(errors)

TEST C: 3.0VOLTS on vdimm
TEST1: 225MHz Test5----PASS; Test8----FAIL(errors)

TEST D: 3.1VOLTS on vdimm
TEST1: 225MHz Test5----PASS; Test8----FAIL(loads of errors)

TEST E: 3.2VOLTS on vdimm
TEST1: 225MHz Test5----PASS; Test8----FAIL(loads of errors)
TEST2: 238MHz Test5----PASS; Test8----FAIL(loads of errors)

238MHz at 1T 3.2Vdimm was as far as it goes error free in test#5.....

I tried to see if a much faster fan would make a difference in terms of cooling and it made NO difference

There is a known problem with Venice + this mobo (Bios13) with 1T which i just confirmed so I CHANGED TIMING TO 2T and here are the results:

[EDIT] NOTE REGARDING UNDERVOLTING BUG
Motherboard was undervolting from 0.07-0.1 in bios13. I marked the tests below as bios selected voltage hence in real life it is actually lower by up to 0.1V.....hence 3.55V in BIOS gives me 3.45V in real life......and 3V actually gave me 2.93V....The positive thing about this is i got 258MHz with tightest timings at 3.45V ONLY.....I cannot wait to get the bug fix with newer bios and see what it does with actual 3.55Vdimm........i'll be aiming for 265/DDR530

TEST A: 3.0VOLTS on vdimm [hence 2.93V roughly the real voltage]
TEST1: 225MHz Test5&8----PASS
TEST2: 230MHz Test5&8----PASS
TEST3: 235MHz Test5----FAIL(errors)...moved to 3.1V

TEST B: 3.1VOLTS on vdimm [hence 3.03V roughly the real voltage]
TEST1: 235MHz Test5&8----PASS
TEST2: 240MHz Test5----FAIL(errors); Test8----PASS

TEST C: 3.2VOLTS on vdimm [hence 3.12V roughly the real voltage]
TEST1: 240MHz Test5&8----PASS
TEST2: 245MHz Test5----FAIL(tens of thousands of errors); Test8----PASS

TEST D: 3.3VOLTS on vdimm [hence 3.20V roughly the real voltage]
TEST1: 245MHz Test5----FAIL(40 errors); Test8----PASS

TEST E: 3.4VOLTS on vdimm [hence 3.30V roughly the real voltage]
TEST1: 245MHz Test5&8----PASS
TEST2: 250MHz Test5----FAIL(errors); Test8----PASS

TEST F: 3.55VOLTS on vdimm [hence 3.45V roughly the real voltage]
TEST1: 250MHz Test5&8----PASS
TEST2: 254MHz Test5&8----PASS
[EDIT]...TEST3: 258MHz Test5&8----PASS (AFTER 5 hour Burn-in)
(End of test there at TIGHT timings set out by tictac...I don't see much point loosening the timings to reach 265......but I might burn them in a little as see what happens)
[EDIT]...TEST4: 261MHz Test5&8----PASS (AFTER 24 hour Burn-in)
(End of test there at TIGHT timings set out....Screenshots included with a SuperPI 32M/PCMARK04/05/3DMARK04/05 run at 2875MHz)


BTW.....the mosfet heatsink that sits next to that heatpipe copper heatsink is not even warm......more like luke warm at 3.55Vdimm...i was expecting to be able to fry a couple of eggs.....well DFI is supposed to be pretty hot at these volts anyhow......I only have active cooling on the RAM.....92mm fan spinning at 1500RPM....slow just to blow some air over the RAM

I'm happy :leet:



NOTE ON INITIAL 5 Hour Burn-in
after 5 hour burn-in at 255MHz and 440xpasses of test#5 i had 20 errors all up....last error appearing on around loop 260......

after reboot I tightened the timings just a bit more......last digit on tictac's suggested timing dropped to 1 (from 2) and went to 258MHz error free......I will need to check the voltage with multimeter and burn in more to get a few more MHz......I can see a stable 260MHz easy

also guys I touched the MOSFET heatsink and it was almost COLD.......either I'm checking at the wrong place or this motherboard RULES

COLD BOOT TEST:
I haven't experienced any problems yet. I will check for COLD BOOT when I get home today.....I left it at 250MHz 1:1 at 3.55Vdimm....... ....[EDIT]Checked...no cold boot......I was also suggested to use 300+ HTT which i checked and found no problems with cold boot

Problems so far:
***Speedfan is showing DDRVDD AS 3.45V...bios selected 3.55V (I need to check this with multimeter....how do I do that....would anyone know)
***Temps are ridiculous even at idle on the Venice.....Abit has a know problem here......my test on Asus had load temps at 42C at 2.8GHz at sligthly lower voltage.....
***OC results are 2T....well anything over 220MHz error free....there is an issue with Venice specifically so seems a bios update might help there


PC BENCHMARKS (on Corsair HC200 W.C.)

WATER PROJECT LINK (Lots of photos)make sure you mark me as referred if you join www.i4memory.com


Best Aquamark3 Run:
A64 3000+@2976MHz LINK


SuperPI/mod1.4 RUNS:
1M SuperPI at 2949MHz


32M SuperPI at 2875Mhz




Some 3DMARK/PCMARK RUNS:

3DMARK03 at 2875MHz



3DMARK05 at 2875MHz



PCMARK04 at 2875MHz



PCMARK05 at 2875MHz

1234

i considered this mb but the only review said it wasn't that good when compared to the dfi so i have held back until people like you iron out the issues b/c i don't have the time.

but, good results nonetheless. interesting to see how this mb pans out. do you know if the an8-sli comes with 3.55vdimm?

dinos22

Thanks, yeah there will be some issues initially but none of them are really that bad.....2T will give me 2% less real world performance but i will compensate for that with higher clocks 2T gives, I know that temps are incorrect so no brainer there....

All NB heatpipe motherboards are supposed to have 3.55vdimm as far as i know

babyelf

nice one dinos

but 1.5+v is the minimum voltage for those speeds?

dinos22

thanks dude

1.5V on CPU yeah.....i had it at 1.52V on Asus at 2.8GHz so this one either needs more volts or needs a better bios....or not reporting properly

/\/\r.Abdul

yeah man thats nice whats the difference between this and the fatality? besides ram cooling and so on?

dinos22

it's cheaper.....I don't really know....i know that the other mobos have that wanker's name on it....also you get this wanker front panel with fatal1ty....it's all the same ****

this motherboard has a ****load of LEDs placed all over it and lights up a dark room pretty nice in red.......looks pretty hard core

babyelf

2 more abit engineers left for asus lol

they do make good boards.. but bios.. =/