**OFFICIAL** Abit BG7 (i845G) Thread

[5150Joker]

I got the BG7 in today and let it do burn-in at default speed most of the day. I did a quick test earlier at 150 fsb and it booted in fine. Ran 3dmark about 10 times in a loop as well as Sandra and Quake 3. This is with the 4:5 memory ratio so I'm fairly impressed so far. I'll have to do some more testing tomorrow to know for sure how stable it is at 150 fsb+.

 

[WideMouth40]

Glad to hear it came in. I;ve been dying to hear another report on this board. Sweet that you got it to 150 FSB out of the chute. Keep us posted!

 

[oldfart]

5150Joker, can you do 150 FSB and 3:4 DDR400? I'd like to see some benches with that setup.

 

[5150Joker]

Originally posted by: oldfart
5150Joker, can you do 150 FSB and 3:4 DDR400? I'd like to see some benches with that setup.

After I burn in my ram some more today I'll try 3:4 for ddr 400. It's so true what they say about needing to burn in samsung memory. Initially I tried 133 fsb and it didn't post at all so I thought, "damn this board really does blow". So I figured I might as well burn in my components at default speed and then gave 133 fsb another try about 8 hrs later and it worked great at default voltages. I hit 150 fsb and ran those tests w/2.7 vdimm and 5% vcore which isn't bad. I think this board reads the core temperatures since it says 41 C idle and I have a volcano 7+ mounted on there. Oh yeah, soon as I received the board the first thing I did was clean up the northbridge and apply some AS3. I'm going to mount a fan on it soon to help keep its temperature down.

 

[oldfart]

I've heard that about Samsung ram. Its odd. I've done (2) 1.6A @ 2.4 systems with Samsung ram. Both went to 150 FSB DDR400 right out of the gate. How do you burn in ram? Have you tried MBM5 to check temps?

 

[5150Joker]

Originally posted by: oldfart
I've heard that about Samsung ram. Its odd. I've done (2) 1.6A @ 2.4 systems with Samsung ram. Both went to 150 FSB DDR400 right out of the gate. How do you burn in ram? Have you tried MBM5 to check temps?

This thread on vr-zone explains exactly what you need to do for ram burn-in. I tried MBM5 and it reports the same temperature as the bios--41C idle and 43C load. Not bad at all if it's a core value. In contrast my athlons 1800xp's core used to run in the 50s and high 40's at idle.

Edit: My cpu is a malay 4/30/02 chip, is that fairly new?

 

[oldfart]

41C idle and 43C load.

Thats not right. 2C between idle and load? Do you mean 41C and 53C?

 

[5150Joker]

Nope I mean a 2 degree difference. Not sure what's up with that, I expected 45 C. I'm going to update the bios later on and see if anything changes w/the temperature readout.


Edit: Did sandra at 150 fsb w/spd settings and it got 2905/2902 memory scores. If i try to set memory timings manually (rather than spd) it gives two long beeps and wont post so i must be doing something wrong (i'm newbie to intel boards). Also, I checked the temperature readout in the bios, it showd 45 C so it looks like MBM was wrong.

 

[5150Joker]

Also, this board "feels" really cheap. Like it was intended for budget consumers to use. It looks really underwhelming when you first pick it up The performance on this thing is stellar so far but I'm really really curious to see how IT7 does. I think I might order an IT7 board on tuesday just to compare them.

 

[Barrei]

Something wrong with those vcore temps should have a much bigger difference between them about 9c and thats with premium heatsink like Alpha , with stock intel heatsink can be as high as 15c between load and idle.

 

[5150Joker]

The volcano 7+ cools better than the alpha for intels. Also I did mention that MBM 5 was incorrect. The true variation was 41 C idle to 45-46C load. It never goes higher than 46C load while i'm at 2.4 Ghz. Anyways, here's a 3dmark I got with this board: http://service.madonion.com/compare?2k1=3598208 Impressed?

 

[Barrei]

After reading reviews on that heatsink your temps do seem in line from idle to load , you must be in a high ambient temperature for your idle to be at 40c, where did you see that this was better then the Alpha ? Can you post it I would like to read it .

 

[5150Joker]

Originally posted by: Barrei
After reading reviews on that heatsink your temps do seem in line from idle to load , you must be in a high ambient temperature for your idle to be at 40c, where did you see that this was better then the Alpha ? Can you post it I would like to read it .

No problem
http://www.overclockers.com/articles373/ scroll down to the air cooling die simulator tests.

 

[5150Joker]

Update: Works perfectly at 155 fsb with vcore at 1.6 and vdimm at 2.8. Here's my newest 3dmark score, i broke 13!!! http://service.madonion.com/compare?2k1=3598785 Also, I broke the 3k sandra mark for memory!! I got 3005/3002.

 

[WideMouth40]

Wooohoooooo, broke right past the 150 mark...more more more

Peace

 

[Barrei]

I saw that test page before and its one of the worst I've ever seen he uses different cfm rated fans on different heatsinks and then gives results , the only real way to test them is to use the same cfm on every heatsink, don't get me wrong I have a stock heatsink so I'm not in favor of one over the other and obviously the volcano 7+ is good , however to judge it on that test is not a very good idea . Buy the way great score on 3d markse .

 

[5150Joker]

Originally posted by: Barrei
I saw that test page before and its one of the worst I've ever seen he uses different cfm rated fans on different heatsinks and then gives results , the only real way to test them is to use the same cfm on every heatsink, don't get me wrong I have a stock heatsink so I'm not in favor of one over the other and obviously the volcano 7+ is good , however to judge it on that test is not a very good idea . Buy the way great score on 3d markse .

Edit: n/m just saw that you said cfm and not same fan You're right they should've used equivalent cfm's.

 

[oldfart]

So you are now running 155 FSB, 4:5 = 388DDR, 2.5 GHz? If you could do 160 FSB, 4:5, DDR400, 2.56 GHz, that would a sweet setup! (not that it isn't already).

 

[5150Joker]

(Howdy all! Joker's friend here, typing this all up for him as he's too busy playing with his BG7 )

OK so far this Abit BG7 using the i845G chipset is performing sweet.

Here's what I got so far using a P4 1.6A, 1x 512Mb Samsung "true" PC2700 DDR module and a Antec PP-412X 400W PSU:

16x140MHz FSB (VCore@Default=1.5v in BIOS) | 3:4 (DDR373) | 2.5:2:2:5 w/vdimm 2.5v
--- Sandra Mem. Score: 2790/2750 | Tests: Sandra Burn-in and many loops of 3DMark2001 | Result: No crashes/unstability so far.

16x150MHz FSB (VCore@5%=1.575v in BIOS) | 3:4 (DDR400) | 2.5:2:2:5 w/vdimm 2.7v
--- Sandra Mem. Score: 2905/2902 | Tests: Sandra Burn-in and many loops of 3DMark2001 | Result: No crashes/unstability so far.

16x155MHz FSB (VCore@10%=1.65v in BIOS) | 3:4 (DDR413) | 2.5:2:2:5 w/vdimm 2.8v (reckon 2.7v might work too)
--- Sandra Mem. Score: 3009/3006 | Tests: Sandra Burn-in and 1x of 3DMark2001 | Result: No crashes/unstability so far but continuing to test for unstability.

Abit BG7, although looks like a weakling little motherboard, performs like a gem so far.

 

[oldfart]

Its odd that 150 FSB 3:4 DDR400 SiSoft mem score is lower than My 4BDA2+ 845D score I would expect just the opposite.

 

[TuffGuy]

Yes it is. On my BG7:

@145Mhz fsb, 2320MHz CPU and 3:4 fsb:memory

Sandra Memory - [/b]2993 / 2990[/b]
Sandra CPU - 4512 / 2852

 

[5150Joker]

I have a bd7II-raid coming in from newegg that I just ordered. It should get here by Tuesday since i overnighted it. I'll post some comparisons between the two when it arrives. Going to keep whichever performs better Right now I have the BG7 at 160 fsb 4:5 and it's running 3dmark.


Edit: Scored 13, 338 in 3dmark. Not too bad.

 

[oldfart]

That BG7/BD7 should be a good comparison. I like the 160 FSB 4:5 DDR400 setup. That is pretty much an ideal set of ratios.

 

[NanoMem]

5150Joker, just curious about your seemingly low sandra mem bandwidth at 150 MHZ FSB, are you using integrated video or an AGP card?

Peace!

 

[5150Joker]

I wouldn't call thoe memory scores low, they were slightly below what old fart listed on his 845E board. I'm using a Leadtek Geforce 4 Ti4600 card with this setup. I'll be receiving my BD7-II Raid tomorrow and plan to test against this BG7 extensively. I'll post my results and see how it turns out. So far BG7 doesn't score any better in memory than 845E boards. I thinkt hey exaggerated it's memory controller's effeciency--it's basically an 845E w/integrated graphics and a 4:5 ratio. Also, it won't let me post higher than 160 fsb so far and I'm fairly certain it's not my ram or cpu (cpu is 4/30/02 malay and ram is samsung pc2700 may batch).