New Posts  All Forums:Forum Nav:

12 pipe mobility 9800 - Page 20

post #381 of 567
Thread Starter 
Anticitizen 2 could probably get into the 5k club by setting all driver settings to max performance. If that doesn't work then put the fans on max and stick it in the freezer and see how high it can be overclocked
post #382 of 567
Anticitizen 2 did you only connect the right 2 dots or all of them and please tell us what bios you are running currently thanks!
post #383 of 567
Quote:
Originally Posted by raelcun
Anticitizen 2 could probably get into the 5k club by setting all driver settings to max performance. If that doesn't work then put the fans on max and stick it in the freezer and see how high it can be overclocked
rolf...in the freezer
post #384 of 567
Quote:
Originally Posted by Anticitizen 2
just took benchmarks at 435mhz/418mhz max safe setting that ati tool detected. 4757 in 3dmark05 and 90.68fps in CS:S with all settings maxed at 1280x800 res. can post pictoral proof if you tell me a good site i can host the pics from...
I wonder why your scores are so much better then mine, ATI Tools is showing that I have 16 pipelines too but my scores are almost the same as the people that have 12
post #385 of 567
Quote:
Originally Posted by duckhunter-deux
rolf...in the freezer
HA! I benched my laptop in the freezer once. I didn't notice anything much!
post #386 of 567
Hi, guyz. I just tested something and I am wondering if you guyz that got the 12 pipes working can help me. I used some equipment my dad has, (I think it's called polymeter) to test if there is conductivity between two points. There are another two points like the ones you connected, at about the middle at the right hand side of the card. Can you guyz with the succesful mods test if there is conductivity between these two points? Cause there is none in mine and it's one of the new ones that the mod didn't work. Maybe we have to connect these two points as well. But before we go destroying cards it will be good if you guyz could give us some feedback.
Thanks
post #387 of 567
Quote:
Originally Posted by KAROLOYDI
Hi, guyz. I just tested something and I am wondering if you guyz that got the 12 pipes working can help me. I used some equipment my dad has, (I think it's called polymeter) to test if there is conductivity between two points. There are another two points like the ones you connected, at about the middle at the right hand side of the card. Can you guyz with the succesful mods test if there is conductivity between these two points? Cause there is none in mine and it's one of the new ones that the mod didn't work. Maybe we have to connect these two points as well. But before we go destroying cards it will be good if you guyz could give us some feedback.
Thanks
Since I have a GC 34th week series there is no reason why I shouldn't at least be geting conductivity in my jumps even if the pipes are bad. The curcuit writer that I seem to of wasted $13us (and $6us for acetone) is very unadhesive for running a wire microscopicly (damn I feel like a robot every time I use that eyeloop or an alien ). Also, the wire wire would have to coming out from under the heatsink and slip under the keybord with the computer half apart just to see if power is available.
post #388 of 567
Has anyone modded with xps yet, I have just noticed 9100's with successful mods. Can you successful one's please post your lappy model and lappy's bios version?
post #389 of 567
The laptop is not a factor in modding the card. It is strictly a video card mod. Since the 9800 is designed to go into either the iXPS or the i9100 it shouldnt matter what computer it is in. I think there are just more 9100 users (it's a cheeper rig for basically the same performance) than XPS users on these forums now a days.
post #390 of 567
raeclun has an XPS and he was the first to successfully complete the mod.
post #391 of 567
well what motherboard bios are you successful guys using? I am using A05
post #392 of 567
Quote:
Originally Posted by raelcun
Given that some people already have 16 pipes without a flash I am skeptical that a bios will change anything.
The passage below is from the guide for flashing a 12 pipe x800pro vivo to a 16 pipe x800XT.
It is still running with only 12 Pipelines, even after flashing an XT Bios to it. The reason why the XT Bios does not unlock the additional pipelines can be found in the Flashrom flashing tool. It "preserves" certain ROM Bios settings when flashing another BIOS:

Several people have used a Gigabyte Bios and flashing tool to get around this, but for some reason you can not flash back to 12 pipelines with this method- if you have defective pipelines you keep them.
Rumor has it that Gigabyte X800 Pro VIVOs come with activated 16 pipelines out of the box. http://www.techpowerup.com/articles/...ng/vidcard/100

So my previous point still stands. Do we need a particular flash utility to unlock the 4 additional pipes (i.e., a utility that does not preserve the ROM settings)?

With respect to the fact that some people are getting 16 pipes immediately after the mod, maybe these cards were enabled from the beginning (similar to gigabyte x800 pro vivos, which as rumor has it, also had 16 pipes enabled out-of-the-box).
post #393 of 567
Thread Starter 
The key thing to point out is:

Quote:
Originally Posted by NigelTufnel
The reason why the XT Bios does not unlock the additional pipelines can be found in the Flashrom flashing tool. It "preserves" certain ROM Bios settings when flashing another BIOS.
In the case of those desktop cards they are flashing another bios which is known to enable the 16 pipelines. All of the mr9800 cards (16p, 12p, and 8p) use the same bios so there is no bios to flash to that will give you 16 pipes. Even if you use the modified flashrom, which I have done, you will be flashing the same bios that the card had on it in the first place. If the flashrom does not preserve anything you are flashing the same bios anyways so it will not matter.

Here is another theory for you:
Whether or not you get 8p, 12p, or 16p is purely hardware related in the case of the mr9800. There are two levels of hardware protection:
1) The laser cut we all know about. Maybe the laser cut can go to different depths and therefore make it so that connecting the surface dots does not actually make the connection?
2) Something else which is not visible?

There is one other thing that has been a bit confusing so far. I would have expected some people to connect up the two dots and enable the extra pipelines (4 or 8) but then find that they have bad pipelines and have to revert back. That has not happened yet. All that has happened is some people have been unable to activate any at all. Is there something that is detecting the bad pipelines and simply not allowing them to be activated because they are bad?

So, we need to attack this problem from different angles and not get stuck on one idea which leads nowhere. That is what happened in the initial bios hacking attempts. Everyone was focused on the bios and now we have some people with 16p and no bios hacking or dead cards. We also need to consider the possibility that some cards have up to 8 bad pipelines which is why they are not being enabled when connecting the dots. Maybe ATI does something else to disable bad pipelines so they cannot be enabled but allows the good ones to be enabled or disabled via the laser cut. Any thoughts?
post #394 of 567
EDIT: Double Post Read Below
post #395 of 567
Someone mentioned another laser cut lower on the card. That might have something to do with it.

A theory I've had for sometime now is that the amount of paste you put on might have something to do with it. The first time I put it on, I put a very very thin layer. Didn't work. So I looked at pictures of successful mods, saw everyone else had put a glob of it on, and put some more on. Maybe for some reason the amount of silver paste applied affects how well the bridge is. The more paste, the better the bridge, and thus more pipes. This could explain the failure of some older cards (aside from bad pipes). If someone with 16 pipes could show us how much paste was put on, maybe we could confrim or deny the theory.
post #396 of 567
the first time i tried the mod i put a thin line and it didnt work. i looked at the trace i made with a magnifier and i noticed that the trace was not covering the dots completely. i took paint thinner and removed the old trace and then put a larger glob that completely covered the 2 dots on the right but no other dots. i would post a pic, but i dont have a good enough camera.

as for trying to get into the xps2 5k club i would try but i am a bit afraid of killing my card now that it has more pipes and the new cards dont seem to work. is it even safe over the long run to have these cards at x800 speeds? i have been thinking of different ways of beefing up the cooling system and they range from taking a large fan and increasing airflow to making a hood so cool air from the AC can be taken directly into the cooling fins to taking precise measurements of the plates that go over the card and having new ones machined out of pure copper. or, of course, maybe a little AS5 and a copper shim would probably lower temps a bit too...

on a side note concerning bios, i had the original bios until the new ones were released around nov. i just flashed over the old ones with the dell flashing utility. i have A06 motherboard bios and A01 video bios that were flashed over the A00 bios
post #397 of 567
I had the same concern about the longevity of this mod. I keep my card at stock speeds unless I'm gaming or benchmarking, and I've propped my lappy up on some CDs to help airflow. I think I'm gonna get the Vantec LapCool2 fan thingermabobber, since I read a review that showed it dropping temps up to 6*C. I might try the copper mod, and if I get the Vantec I'll get some AS5 too. My card got up to 66*C when I had it at 435/415 and I don't know the max temps these cards can hit safely. On a positive note, I got ~110 fps with everything on high and 1680x1050 res in UT2004 at those speeds.

You can probably push your card until you see artifacts, and then immediatley drop back. I'm guessing you could get the core up to 450 and the mem up to maybe 430.

I'll try updating both my CPU and video BIOS when I get home.
post #398 of 567
66*C? during the loading breaks in 3dmark05 there was enough time to see the temps that I8kfanGUI posted in the tray... i have cpu and gpu temps displayed in the tray and during the overclock in the test it only got up to 57*C. i heard that in I8kfanGUI the ram and gpu temps are switched around...is this true? if so, then my gpu does get up to around 66*C. were do you have your laptop sitting? mine is in my basement with air conditioning upstairs and it is propped up with 2 decks of cards.
post #399 of 567
I'm using I9fanGUI. Mine didn't get that high during 3dmark, it did after playing UT2004 for about an hour. I have my lappy in my 75-80*F room with the air conditioning on. If the gpu and dimm temps are switched then my GPU is more like 60.
post #400 of 567
Here are my card readings in case anyone is interested:

GPU:

216pecga12f
gc5900.1
0435AA

Ram:

Samsung 425
k4d55323qf-gc33
New Posts  All Forums:Forum Nav:
  Return Home