JCviggen
08-05-2002, 11:24 AM
Okidoki, I got one in today.
I've been following the huge thread at HardOCP about this board, and I wasnt expecting it to be an easy ride.
I dont need rock stable high clocks, if they get me through 3Dmark at good clocks and its stable at more "normal" clocks for everyday operation then thats fine. But this board really does have issues. Issues with a capital I
The only spot of good news is that it will boot into Windows at 160x4 on the memory. 3Dmark runs a bit before i'm welcomed by a blue screen. But the memory isnt my main concern at the moment. This board acts very weird. i've only tested it for one hour or so, but here's already some first thoughts.
The "auto" setting for Vcore tells me its at 1.50V, and it measures 1.58 in windows. If I put it on user define and select 1.50V, it will give me around 1.68 measured in windows. However, I took some screenshots of the voltage under load.
Screenshot 1 shows the voltage at the user define setting, 1.50V, idle in windows at 150x16. All is normal, voltages look OK. Everything is within limits of what I call "stable" voltage. Every board has small voltage swings, the ones I see on the ASUS are normal, the 12V rail is the same as it was on my TH7-II as well.
http://users.pandora.be/JCviggen/Vcore1.JPG
you can follow the graph at the bottom a bit.
Then, I stressed the CPU by turning on prime95. The Voltage dropped to 1.62, a normal drop by stressing the CPU.
http://users.pandora.be/JCviggen/Vcore2.JPG
As you can see, the CPU load goes to 100% (blue line) and the temp rises slowly.
third pic, the voltage takes a sudden dive (and I mean sudden) to 1.52-1.55V and it stays there. It takes a while for the voltage to suddenly jump up again to above 1.60. the scale is kind of tiny but you can watch how the voltage goes down, and then up again a bit later. But it stayed low for a while after the CPU load had gone to practically 0. I would have expected it to come back as soon as prime stopped.
At exactly the point the voltage dropped, prime errored out. But here's the thing, my 1.60A runs about 2500-2550 MHz at 1.50V, so a drop to 1.52V shouldnt crash it at only 2400 MHz.
http://users.pandora.be/JCviggen/Vcore3.JPG
Here's another thing to think about : follow the CPU temperature line. At the exact point that it touches 50°C, it crashes. I have repeated the test several times at all kinds of speeds (seriously low ones also) and for some reason, when the CPU temp touches 50C, it errors out on prime95. Its so exactly spot-on the 50C mark at different settings that it cant be a coincidence.
I then decided to put the thing in my prometeia. Looked pretty good, BIOS reported CPU temp of -10°C. Booted into windows and ... CPU Cool is no longer reporting the CPU temp sensor. its gone. MBM reckoned my CPU was 116°C, but since the back of my mobo was gathering some ice, I am prepared to challenge that reading. If I stressed the CPU a bit though, I was welcomed by a freeze or blue screen. Even at speed that should work perfectly. In fact things are a lot more "workable" with just the air cooling now, but i'll try again tomorrow with the phase change.
I have now knocked it to 133x16 and auto Vcore, and i'll see what happens.
From the "feel" of things, this board is a worse CPU overclocker than my TH7-II was, and most ppl know that a TH7-II pretty much sucks. An ASUS P4T-E or any abit DDR board will give you about 100 MHz more at less voltage. But this has yet to be confirmed.
Right now I cant recommend this board. I wanted to check out the problems for myself, and they are pretty big. I dont know if its both the board and memory that need work, but I am sure about the board Ã:banana:nd its BIOS. Not a good job by ASUS
JC
PS : no clue why the images show up as links, but we'll have to live with it
I've been following the huge thread at HardOCP about this board, and I wasnt expecting it to be an easy ride.
I dont need rock stable high clocks, if they get me through 3Dmark at good clocks and its stable at more "normal" clocks for everyday operation then thats fine. But this board really does have issues. Issues with a capital I
The only spot of good news is that it will boot into Windows at 160x4 on the memory. 3Dmark runs a bit before i'm welcomed by a blue screen. But the memory isnt my main concern at the moment. This board acts very weird. i've only tested it for one hour or so, but here's already some first thoughts.
The "auto" setting for Vcore tells me its at 1.50V, and it measures 1.58 in windows. If I put it on user define and select 1.50V, it will give me around 1.68 measured in windows. However, I took some screenshots of the voltage under load.
Screenshot 1 shows the voltage at the user define setting, 1.50V, idle in windows at 150x16. All is normal, voltages look OK. Everything is within limits of what I call "stable" voltage. Every board has small voltage swings, the ones I see on the ASUS are normal, the 12V rail is the same as it was on my TH7-II as well.
http://users.pandora.be/JCviggen/Vcore1.JPG
you can follow the graph at the bottom a bit.
Then, I stressed the CPU by turning on prime95. The Voltage dropped to 1.62, a normal drop by stressing the CPU.
http://users.pandora.be/JCviggen/Vcore2.JPG
As you can see, the CPU load goes to 100% (blue line) and the temp rises slowly.
third pic, the voltage takes a sudden dive (and I mean sudden) to 1.52-1.55V and it stays there. It takes a while for the voltage to suddenly jump up again to above 1.60. the scale is kind of tiny but you can watch how the voltage goes down, and then up again a bit later. But it stayed low for a while after the CPU load had gone to practically 0. I would have expected it to come back as soon as prime stopped.
At exactly the point the voltage dropped, prime errored out. But here's the thing, my 1.60A runs about 2500-2550 MHz at 1.50V, so a drop to 1.52V shouldnt crash it at only 2400 MHz.
http://users.pandora.be/JCviggen/Vcore3.JPG
Here's another thing to think about : follow the CPU temperature line. At the exact point that it touches 50°C, it crashes. I have repeated the test several times at all kinds of speeds (seriously low ones also) and for some reason, when the CPU temp touches 50C, it errors out on prime95. Its so exactly spot-on the 50C mark at different settings that it cant be a coincidence.
I then decided to put the thing in my prometeia. Looked pretty good, BIOS reported CPU temp of -10°C. Booted into windows and ... CPU Cool is no longer reporting the CPU temp sensor. its gone. MBM reckoned my CPU was 116°C, but since the back of my mobo was gathering some ice, I am prepared to challenge that reading. If I stressed the CPU a bit though, I was welcomed by a freeze or blue screen. Even at speed that should work perfectly. In fact things are a lot more "workable" with just the air cooling now, but i'll try again tomorrow with the phase change.
I have now knocked it to 133x16 and auto Vcore, and i'll see what happens.
From the "feel" of things, this board is a worse CPU overclocker than my TH7-II was, and most ppl know that a TH7-II pretty much sucks. An ASUS P4T-E or any abit DDR board will give you about 100 MHz more at less voltage. But this has yet to be confirmed.
Right now I cant recommend this board. I wanted to check out the problems for myself, and they are pretty big. I dont know if its both the board and memory that need work, but I am sure about the board Ã:banana:nd its BIOS. Not a good job by ASUS
JC
PS : no clue why the images show up as links, but we'll have to live with it