Asus G73 Benchmark Thread

The error code is a "Kernal-Power 41 (63)" which is just a complete loss of power, as if I held the power button for 7 seconds or removed a desktops power plug, kind of shutdown.

As for the 920xm... getting the 180w brick would add 30w more to the machine, which if the processor, via throttlestop, I can increase the wattage and amperage intake of the processor. Given my error code, it prolly tried sucking too much juice and the machine couldn't run. But the brick would alleviate that issue, am I correct?

 

hummm, well when you beat my 4.320ghz over clock....i may give into that..but for now...nope.
when benching just the cpu...you have plenty of power. but the code you will get is error 101 which means..you need more cpu voltage.

 

Where I got that code was in the windows logs, R-click Computer, Manage and logs. That error pops up whenever the computer is shut down improperly, irregardless of how it happened. It could be a failure to recover from sleepstates, or long-pressing the power button, to loss of power.
That's why I think the 180w brick will help. What do you think there, given that information, as I am still running off my standard 150 right now.

And I will be looking for the infamous Error-101 when I do get my new brick.

And im glad for the positive feedback on a potentially destructive hobby, altho the 920xm seems to handle itself pretty well as idiot-proof.

 

you can click my banner to get an idea of what i might know....

 

All I can say is very impressive numbers.
On a side note, how much power is being drawn by that 920xm under test loads, and your TDP/TDC settings for it? I just wanna know exactly what you did to achieve that speed. Using setFSB, I set my procs frequency from 133 to 150, so my DDR3 memory 600mhz a stick vs 533, and setting the thermal limits of TDP/ TDC, if those are higher, my comp just turns off under load, versus lower TDP/TDC settings and getting a Halt bluescreen if my multiplier is higher. I have yet to hit the 101 error (voltage) yet.

I wanna see if my unit and proc are capable of similar numbers if power input (brick) is an issue right now. (I can hear the plastic go crick once in a while when it warms up)

 

i can run tdp/tdc at 99/65 with multipliers at 26/27 while benching wprime 1.55. this should be enough to give you the bsod and the error code 101 or 124 or 24, but you have to wait for the first blue screen to disappear first! you can also use the vantage cpu test as well..3dmark06 cpu is not as hard on cpu over clock.

raising fsb incurs more heat than good really.

yu have yet to post any screen shots of where your at..so start with the easy one...cpuz

side note:
and i can do the same with my 150 brick as well as with my 180 brick.

 

I shall be grabbing those real soon, I was on my Droid2 for the last umpteen posts that I made.

I have all the software downloaded to bench/monitor. And I know raising the FSB can have more drastic changes to heat, as the memory speed increases as well with that trick.

Aside CPUZ, any others you want posted? And when I try those W/A amounts, my computer wont even bluescreen, it just shuts off, and heat is within normal ranges too.

http://3dmark.com/3dm06/15270482 CPU was clocked @ 25/25/27/28
http://3dmark.com/3dm11/391206 CPU clocked as above.
3dMark Vantage requires a key above trial, as the resolution it wants my 1600x900 Screen cannot support.

 

a few things.
you can't bench 06 because you dont have the right screen for it.
i asked for cpuz and wprime 1.55 because this is something you can do.
you need all cores at either 25 26 27 28 not mixed.

i take it you dont know how turbo works?

johnksss's 6sec 546ms wPrime 32m run with Core i7 920XM @ 3709MHz

here, we can run it together..ill even use my 150W psu
my 3dmark06 score at the right resolution
17,269
Result

 

I know how the turbo works, if its running off all 4 cores, it will default to the @4 Cores speed, but if you get single or dualcore activity only, it will bump the multiplier to the @2 or @1 core speeds. If temperatures get in the way, or too much power is drawn, it will lower its clockrate down to Base clock speeds of 2ghz.

I am noticing one thing, as I go to run the tests... CPU-Z (uploading the file now, had to take care of a few things) reports my CPU as a 920xm, yet wPrime 1.55 reports it as the following:

Intel(R) Core(TM) i7 CPU X 920 @ 2.00GHz (Engineering Sample)
Core:
Clock: 3458 MHz
Vcore: 0.713 V
FSB: 531.7 MHz
L2 Cache: 256 KB

Would that have anything to do with it? When I put the chip in, it had the SLBLW markings on it.
I am registering so I can check the posts I make there more easier, but the tests I am running 32m on 8 threads, is taking 9-10seconds @ stock speeds or 26x multi I will have those posted in a few.

 

yeah, that is wrong...it is using cpuz 1.41 which is of course...well well out of date...lol
dont worry, your cool. you have a retail..

 

I am just configuring and cleaning up my desktop, getting some saved screenshots and all that jazz.

One thing I want to know, why would I have a .5sec difference between 2.26@4 Cores vs 3.33@4 Cores (Max Turbo it can supply, not guaranteed though)

And after I run a test with 3.46@4 Cores, it will shut off after a minute, no BSoD or anything.

 

depends on your tdp/tdc formula
you send to many amps in and it will cut off..the same for watts.
if your temps hit 100C on to many cores it will do it as well.
once you post some pictures..i can try to recreate and see what happens.

 

I posted all the relevant information that was needed, I think. I have tried this OC twice, with FSB@133 and higher multipliers, and FSB@150 with lower multipliers, same result.

TDP=80 (Wattage, I know this)
TDC=50 (Amperage)

I can make the TDP go higher without a crash, but the OC wont go any further, and the TDC wont rise any higher without instabilities, regardless of my FSB amount. Raising both allows for a higher base "Turbo" Speed, but instabilities occur just past 50TDC. Also, I can hear my power pack make a slight crack sound sometimes, like the plastic is moving a lil. Think its a dud power supply waitin to pop?

The only thing I am down to, is potentially reloading Windows and reinstalling everything fresh.

 

1: set it to 24x multiplier
2: turn off that turbo mode button.
3: set the tdp to 99
4: set the tdc to 65
then run the test.
set the temp to C not F

 

Will do, and as for the FSB, leave it at 133 Right?
I have a feeling that will cause my comp to just shut off, gonna try it tho.

 

133, right.
nah, it should stay on..nice and happy.
provided your chip is really oem or retail...

 

I certainly hope it is, the later version of CPU-Z shows it is, all the steppings and everything match. Or I may have just gotten a dud (not really, just in OCing). Cause I can reach 2 cores @ 3.6ghz no problem, 1 core @ 3.7ghz pretty easily without issue.

**Edit** I ran at that multiplier, and tried one higher, sofar so good. But one of my cores did max at 100*C.
Score now became 8.416sec for the same test. Processor peaked at 3.33 across all 4 cores then held 3.05ghz for a majority of the test.

 

did you run all 4 cores @ 24x?

 

Yes, all 4 cores @ 24x, the processor would peak at the 3.2 Max, then drop to 3.05-3.07 due to heat. It was within 5-10*C of maxxing out.

 

ok, then that's your problem now...heat...

 

Well, I applied Tunix-TX2 to it, I applied roughly what the Pea Sized blob was, think theres not enough on it?

 

Ya blow out those fans yet? Perhaps you might need to apply more pressure while repasting. I did and it helped a bit.

 

Oh yeah, I blew them fans out good. Had a thin layer of hair/dust on them.

I will be on this thread via my Droid2, may be opening this up soon.

 

and make sure you dont bend the heat sink brackets when putting it back on else one side will be far hotter than the other.

 

Yeah, I have heard about them retention mods and how they can potentially cool the system MUCH more than normal, but, my processor is almost identical on each core when its idle, and under load, just varies based on which core is stressed for work.

 

then your good. you would know once you ran wprime. half is 95C and the other half is like 75C.
and the retention mod isn't for us...it's on the alienware machine.

 

most likely for the alienware only, but whats to stop a fanatic from making it work for us? LOL

Last Question before i really embark on something time consuming. I know that the older HeatSink Compounds, if you had a bit too much on, it would actually insulate the processor, causing overheating to happen much faster. Now, stuff like Tunix and IC Diamond 7, if we applied a lil much, since its non-conductive it wont fry our boards, but would it still Thermally Insulate the CPU?

 

when you put the compound on and press the heat sink on. wait about 15 seconds..then take it off and look at what you see. 90 percent of the time..most of it will be off to the side on the pcb portion. i dont even do the pea thing. i spread it then put the heat sink on and done.

 

So to me then, it sounds like I applied too little in that case.

 

nope. probably just the right amount...

 

any bennchmarks with the 11.1a?

 

i mean in comparation with older drivers (stock clocks)
10.10 , 10.11 or 10.12

 

Did somebody tried to replace his HD 5870M from his G73 by a Graphic Card MXM 3 ?

 

While the G73 uses a physical MXM 3.0 connector, electrically it is different. It will not work with any other cards.

 

But did somebody tried ?

 

"Also tried MSI's 5870 card on his unit and the system does post without any problems, put his card on another G73JH and the result is the same, won't boot from anything."

This is quote from same forum here and comming from gentech retailer.
scook9 how you know that no other mxm 3.0b card wont work in g73? Did you try it?

 

Link? 10char.

 

I think that was a quote from me when my vBIOS update crashed and Ken put a MSI 5870m in my G73. It fit and did post, the G73 5870m PCB is of a proprietary size and does not fit in any other laptop (at least at the time, I wouldn't know now).

Read Update #2

 

I remember reading that, and I thought the conclusion was the asus PCB is just too big to go elsewhere...

 

So i can put an other graphic card (HD 5870M from MxM-upgrade) in my G73JW ?

 

Yeah but here we talking about if other mxm 3.0b vga board would work on g73jh.

 

http://forum.notebookreview.com/asu...elp-i-bricked-my-g73-vbios-9.html#post6395285

 

Its not a G73JH but a G73JW !!!
What are the differences betwen The JW bios and the JH bios ??? Would a JW bios support a ATI card ??

 

As long as bios properly recognise new hardware it should not be problem. I`m still not completely sure if our mxm slot is 100% compatible with MXM 3.0b standard though.

 

In the past Asus has always altered the MXM slot electronically to make other vendors cards non compatible

 

Yeah thats true. But as you say in the past. Hopefully they change that rule

 

I think I will buy a HD6950M to try !!

 

I wish you luck man. We need someone willing to spend money to find out truth. We are all just talking here and nobody knows for sure becouse nobody tried .

I was in same dilema becouse my gpu is dead for now and I have found 6970m on internet but it costs 500 euros plus taxes. It is still to expensive for my pocket anyway.

 

Be carreful, HD6970M is a 75-100w card so if you take it, you will have to buy a powel supply of Min 180w !!! Thats why i want to buy a HD6950M, in notebookcheck, they said that she need 75w !!!

 

Thats a smallest problem man if I was sure it is compatible and is little bit cheaper, then ac adapter is least I would worry about