Particle: If you have v1.0 of the boards, it might be because the vdroop on vcore? Before i did the mod the vcore could drop from 1.54 in idle to 1.41 in load. :O I had big issues getting over 5GHz.
Printable View
Particle: If you have v1.0 of the boards, it might be because the vdroop on vcore? Before i did the mod the vcore could drop from 1.54 in idle to 1.41 in load. :O I had big issues getting over 5GHz.
I've got a v1.0 board, but I get overshoot instead when under load. If I set to 1.4375V, I get about 1.51V under load.
Ok, first Cinebench run, just to get overall feeling for the chip.
Nothing spectacular, but wanted to see 5GHz and if it can run the GTX2's at their rated speed and timings. Yes, it can!
Temperature in the room is 22*C, real voltage under full load was 1.443V.
I have no drop in frequency, but can't install AOD and PSCheck. I should change my bloated OS.
Cooling is PA140.3 with 3x 140mm YateLoon 2300rpm. I'm a little worried about the board though. Not going to push more this night.
I'm happy with the magical 5GHz for now :).
What is the most appropriate way to detect default VID? CoreTemp says 1.275, here's cpu-z report as well. Is it really 1.275V and if yes - what are the chances of getting 8GHz+ validation?
# of P-States 7
P-State FID 0x1A - VID 0x0B - IDD 18 (21.00x - 1.412 V)
P-State FID 0x17 - VID 0x0B - IDD 18 (19.50x - 1.412 V)
P-State FID 0x19 - VID 0x16 - IDD 12 (20.50x - 1.275 V)
P-State FID 0x11 - VID 0x1B - IDD 10 (16.50x - 1.212 V)
P-State FID 0xB - VID 0x25 - IDD 7 (13.50x - 1.087 V)
P-State FID 0x5 - VID 0x30 - IDD 5 (10.50x - 0.950 V)
P-State FID 0x10C - VID 0x38 - IDD 3 (7.00x - 0.850 V)
https://dl.dropbox.com/u/30576042/20...2022.07.29.png
Nice one Ivan! Scary how hot these chips run isn't it? The Crosshair 4 F is a very solid board but it feels a bit flakey with bios 3027 imo. It seems to clock your GTX 2's very well though. Good luck pushing it.
Very encouraging results with the Hypers. I hope we can do 2133 CL7 prime stable...
They were running 2160 7-7-6-20 1T 90ns 1.675V, but didn't wait for the Pi32M to finish, I was eager to try Cinebench, lol.
Memory clocking seems on par with my SB. Not sure what would be the Prime stable frequencies and if I can stabilize them at all.
Not sure if I'm going to do excessive Prime testing any soon. Really worried about the board and will stick to shorter tests for now.
Feels like the chip is torturing it. I don't want to end up with melted power connector :).
I don't think the powerconnector is a common problem for these boards. I pointed the IR thermometer at the connector during heavy load and it didn't get any hotter than the rest of the board.
You just need to make sure the pins on the connector are gripping firmly.
I have it for 3 hours now, so not going to push any further this evening, but expect some competition in CB in the future :D
At least I'm going to try very hard. My score is low as well, but will get higher with a fresh install.
BTW, Zeus, do you have manual voltage adjustment in CH4F bios or it's only offset mode that is available?
Batch is 1142GPM.
I'd probably need upwards of 1.71-1.72v to match that, but extremely encouraging.Quote:
Originally Posted by I.nfraR.ed
Looks like 2200 7-7-6 in Cinebench should be a peice of cake ;)
I think it had manual voltage as well but it defaults to offset voltage which i hate. I'm not 100% sure cause i might mix things up with my current CVF. It should be in the extreme tweaker menu.
The score is so low...it must be the OS or I'm mising something :(.
Going to rerun it tomorrow with a new os.
Chip didn't scale very much. More voltage was making it worse. Lower clocked memory and Northbridge didn't help either.
http://dl.dropbox.com/u/30576042/cb-1.png
Link
I can't say for sure whats happening on your rig, but I did notice if I pushed the IMC too hard it actually lowered my scores (throttling :shrug:)...Quote:
Originally Posted by I.nfraR.ed
Try lowering the IMC multi by 1x and try that run again, you might be surprise! ;)
mines scoring all over the place LOL 5.06 was a 3.1ghz
then 5.17 at the same clocks.
then just see if i could match the FX 8150 I set 3.6ghz and only scored 5.27 so then I realized something was up.
I had core temp open with the fastest clocking reading setting 100ms.
down clocked to 2.8ghz instead of running 3.6ghz. I tried the AOD trick that work but then it stopped working.
so then I say screw it let me set this 3.6ghz in bios and just for fun I set 2.6ghz NB and left the memory at 1866mhz it score 6.44. :/ i have no been able to repeat it at all.
Hey Ivan, if that is at normal ambient temps it is a nice clocking chip. Max cinebench 11.5 stable i can get at 20ºC is ~5150. The score is a bit odd though but i'm sure you'll work that out. It's not the efficiency of the board, i even think efficiency of the CIVF is a tad better than the CVF.
Good luck pushing bud.
yes, these are great scores, specially,if is it with normal ambients. Il have very hard with aircooling get 5 GHZ in R11.5 think.
The luck you have with chips, I'm sure you'll do fine. Unless you get a chip like mine....:p:
Mine seems to want some "Xtreme" cold.
No, max on 22*C was smth like 5160MHz. This is with 15C ambient. It just doesn't like more voltage.
Dave have awesome chip, 100MHz better than mine. Low score might be a result of using RC Tweakit to monitor temps/voltages.
Will try without it.
This 1,3375 VID FX-8150 overclocks little better than my 1,2625v VID. This is done with about 23c ambient temps and watercooling:
http://i10.aijaa.com/b/00333/9088685.jpg
Nice one Asmola. :up: Looks like 5140-5160MHz is about average for good watercooling and normal ambient temps for these FX chips in this bench.
Ambient temp ~23 celsius degrees and air-cooling:
http://loppmarknader.se/blossas_albu...rmal_CB826.png
Maybe I can do 5080-5100MHz with this ambient temp, but it sure is starting to get hot now. :D Max SP1M is just over 5400MHz.
Well, I found the reason. It was not the ROG Connect, Northbridge or the OS. It was some combination of cpu multi and HTRef.
Tried with lower NB, but the results were still bad. Then dropped the multi and increased HT. The score went up, just where it should be, but as a side effect I was not able to complete CB at 5250.
It kept freezing on the last few or single square on the top left corner.
FX-8150 @ 5225MHz | Crosshair IV Formula | Water Cooling
8.57pts
http://i.imgur.com/KCXzLs.jpg
Image is clickable
That's odd. I did not experience any of that with my C4F. Maybe it has something to do with different straps? (internal latency in NB) Freezing at the very last row of squares or even the very last square sounds familiar. How annoying!
It always seem the fastest runs that freeze at the very end. :D
Updated, FX's only...
http://img257.imageshack.us/img257/4...bench115fx.png
Code:Name /Post# Score Cooling CPU MHz ScreenShot
01 Dumo /461 8.85 1-Stage X8 FX-8150 5418 01
02 Daveburt714 /582 8.84 H20 X8 FX-8120 5389 02
03 equinox1 /445 8.72 H20 X8 FX-8120 5331 03
04 krjalone /556 8.65 H20 X8 FX-8150 5310 04
05 Zeus /550 8.60 H20 X8 FX-8120 5250 05
06 I.nfraR.ed /621 8.57 H20 X8 FX-8150 5226 06
07 Asmola /618 8.50 H20 X8 FX-8150 5168 07
08 wez /530 8.45 H20 X8 FX-8120 5140 08
09 El Gappo /411 8.34 AIR X8 FX-8150 5090 09
10 blossa /620 8.26 AIR X8 FX-8150 5051 10
11 Particle /511 8.24 DICE X8 FX-8120 5235 11
12 PolRoger /515 8.18 AIR X8 FX-8150 5017 12
13 charged3800z24 /512 8.16 H20 X8 FX-8120 5017 13
14 Zeus /602 8.14 H20 X8 FX-8150 5019 14
15 FlanK3r /408 8.08 AIR X8 FX-8120 4941 15
16 sandlewood /500 7.55 - X8 FX-8150 4823 16
17 Baam /486 6.99 - X8 FX-8120 4275 17
18 gOtVoltage /521 6.14 H20 X6 FX-6100 5001 18
19 The Stilt /510 4.29 AIR X4 FX-4100 5201 19
Nice run man, glad you got it sorted out... :clap:Quote:
Originally Posted by I.nfraR.ed
That Mem and your IMC are REALLY impressive BTW!
My cores might clock a little better than yours, but my IMC craps out ~2600 and v's over 1.25 don't help (didn't push it farther, scaling stops ~ 1.2125).
Would you mind posting an AIDA mem bench with those mem/IMC clocks?
I'm still getting really wierd #'s on that (especially L2 read), varies between 17000 and 42000, strange.... :shrug:
Yes, same for me about L2. One run is 18k, other run - 48k or whatever it was.
I don't have aida run on those particular settings, but have a screen from the previous Cinebench session.
CPU-NB voltage was 1.29V, vMEM = 1.7V. Didn't try max aida frequencies.
IMC seems above average. Mine needs 1.14V for 2450MHz Prime95 Blend.
http://i.imgur.com/nyb1v.png
Now, when I think about the low scores again...don't remember if I tried lowering mem divider. Needs more investigation, but I don't have time right now.
On a side note, it seems Hypers are running fine on the CHIV. 2000+ 6-6-6-18 should be possible with GTX2's.
@Asmola: Very nice! Just like Eugene said, it seems 5150 - 5160 is where most of the chips top on normal watercooling and ambient temps.