you are running out of memory.
Before exporting your mesh, did you check its quality? (in Gambit: Examine Mesh)

as the screen shot, the quality is ok right? for the distribution

set lower value at 0.9 and update (examine mesh panel)
it will return how many cells are above skewness 0.9.
if you get 0 cell or very few (but with skewness less than 0.95), than I can imagine your mesh is ok (regarding quality)
you will do another checkmesh later in fluent
By the way, are you running your mesh (6.8million cells) on serial solver?
If yes than I assume you need at least 8GB Ram.

serial solver? meaning? just 1 PC?
by the way, the mesh quality i can guaranteed is fine, because all the way during the meshing process, i only got 1 warning regarding poor mesh geometry showing that 3 >0.97 equisized skew element on face ... . Only 1 particular face have that, and the number is just 3!

yes, serial means you run your model on one cpu.
If you get element with skewness > 0.97, then you have probbilities that the check mesh in fluent will fail. And if fluent enables you tostart the iterations, you have lots of chance, that it will affect your connvergence.

then now presume that my mesh were fine, then it is my computer problem now?

if you cannot load your mesh and you run out of memory, then you are on limits from your hadware.
Reduce your mesh or, load it on a more powerfull pc

hmm, reduce my mesh? meaning decrease the number of element right? anyways i not sure how to do that already. as i u can see, some of the spacing i already use 2 instead of 1 (mm) ..

is your mesh full tetra or tetra-hexcore?
If it is full tetra, then switch to tetra-hexcore.
Your mesh size will decrease

mostly tetra.. ok will try it again.. will let u know the outcome :)

Hey max, i glad to tell that, i use hexcore (native) function and i able to reduce the element to 1/4 of the original, which is around 1.8 million only, if i not mistaken.. However, i still feel very laggy when i press display grid function to show the "default interior" which is my fluid, is it normal?

display your mesh on your surface wall.
If you want to see your mesh inside your domain, you may use sweep surface.
Or you can use the plane in Examine Mesh (Gambit)

ok max, i have research in these 2 days, and i have a rally good news for me i guess, not sure if really a good news for u. Take a look at my screen shot
http://i.imgur.com/j5ScK.jpg
http://i.imgur.com/tFTn8.jpg
Look at the number of element, i can reduce to about 300k, i guess it is really good! and now, i guess i am done with gambit? Any other thing i need to refine before i go to FLUENT?

you can fix most cells with size function:
All your cells are attached to thin surfaces, which give cells som stretched aspect.
If you apply a size function with those surfaces as source, then I am pretty sure you can decrease the number of elements (with skewness >0.9)
But you can try starting your calculation from this mesh.
If it failed, you know what you should do

hmm, well.. i know size function is useful, but i really no idea how to use it :(

ok bro, i have asked my supervisor about tis and he say is ok to neglect the minor poor mesh. So i proceed FLUENT. Now at FLUENT, this is the following step right?
1) Grid Check
2) Grid Scale
3) Define --> Solver
(What slover? Pressure or density?, i let pressure for now; Steady or unsteady? i guess it is an steady flow right?)
4) Define --> Energy Equation
5) Define --> Material
6) Define --> Operating Condition
(Pressure is 101325; reference temperature, XYZ i guess i can put all 0 right?; Gravity neglected)
7) Define --> Boundary Condition
Here the problem start.
a)http://i.imgur.com/o2OBE.png
This picture is the boundary condition where i define in GAMBIT. Once i export mesh, gambit auto define a boundary condition named "wall" for few other identities. As u know, this "wall" named boundary condition can just be view in FLUENT, so i view it in FLUENT and picture below tells.
http://i.imgur.com/aYxjT.png
################################################## #######
Besides, in defining boundary conditions, For a wall boundary condition,
http://i.imgur.com/D65iW.jpg
Question are shown as the figure. Besides 1 additional question here, in my case there is no need to use other tab (eg momentum, UDS, etc) right?
Next is, are the exhaust fan and the pressure outlet boundary condition the same type?
The question regarding exhaust fan is as attached in the picture.
http://i.imgur.com/Kjz7D.jpg

From now you should open a new thread.
Other people could help you more than I can, because I am not so experienced with thermal stuffs.
Anyway the 2 first tutorials in Fluent (6.3) are heat transfer:
http://my.fit.edu/itresources/manual...g/main_pre.htm

ok, anyways i really thanks to u for helping me throughout the process.. i will try to do the size function on my own btw.. i will still reply in this thread if i need any help bro :) thanks

Max, I have looked many times through this forum and your many replies in various threads, have helped me a lot. I admire your patience and your willingness to help.

thanks