View Full Version : Solutions to this baffling Tualatin Celeron equation? Please read...
br0adband
12-28-01, 07:29 AM
Ok, I'm a total noob to this Forum, but I've been reading the website itself for a long time. Haven't toyed around with overclocking much in the past, just my old GF2 Pro (with great success).
Just bought the following yesterday:
Intel Celeron 1.2GHz Tualatin (we need a name for these, maybe Celeron T's or CT's or TC's or just Celeron 3)
Asus TUSL2-C (so far, I've read that this isn't the best board for this processor, read on)
Vantec FCE-62540D Heatsink and Fan on the CPU with Arctic Silver II (applied with instructions from www.arcticsilver.com)
256MB PC133 SDRAM with Infineon chips
MSI Geforce 3 Ti 200 (default speeds)
IBM Deskstar 60GXP 40GB 7200rpm drive
SBLive! Value 5.1
TDK 24/10/40 CDRW
350 watt Power Supply
No other hardware, and no I don't have a floppy installed (no need for that antiquated hardware anymore)
That's it. This whole machine is just put together, it's running Windows XP Pro at the stock settings across the board (CPU vcore and FSB; memory at SPD timings which are 3-3-3 but it runs *stable* at 2-2-2 also). Nothing else out of the ordinary.
Here are the current case and CPU temps as I'm typing this:
CPU 28° C
Case 31° C
Ok, after running Quake 3 Arena for 1 hour at all these settings I get the following (temps measured as soon as I get back to the Desktop, I'm using Motherboard Monitor 5.1.0.2):
CPU 31° C
Case 31° C
Before I get too in-depth here, let me say up front that I live in Death Valley, CA, *normally* the hottest place on Earth. But my ambient temps in my room stay pretty much the same all the time, and I'm running all this with the case ON and no case fans (yes, I know they'll help, I just haven't had a chance to pick any up yet - that's the one thing I forgot). And since Las Vegas is a two hour drive each way, I'm not running back into town just yet.
Now the question is, are these good temps? The CPU temp rarely goes over the ambient temp no matter what I'm doing. I've run Prime 95 for 3 hours (not going to run it for 24 hours straight, seems a bit pointless at the moment, but I might) and it's still 31° C for the CPU temp. I've run 3DMark 2001 in 30 loops, run Q3 demos till my eyes hurt, Prime95 for the above mentioned 3 hours, CPUBurn and several other torturing programs for this machine.
Why am I telling you people this? Simple. Because I can't get this thing to get to a Desktop at anything over the default 1200MHz.
Seems really screwy don't it? Yeah, I'm wondering the same thing.
This is a Phillipines made Celeron running at stock 1.475 vcore. The BIOS on my TUSL2-C is the most current (1011 beta 007), everything is default across the board (I think I already said that, just making sure it's understood).
Now I'm not a total noob in all respects about overclocking. I've owned several Asus boards, an Abit board or two, one Shuttle board and two Gigabyte boards and never even messed with overclocking before with any of them. Now, in my first real attempt at doing this, I can't get a damned thing to come up.
What happens is this: I made the assumption that this thing *should* be able to hit *at least* 1320 (12 x 110 FSB) since the mobo itself and the chipset (i815EP) are designed to do the 133MHz dance out of the box. I'd love to get to 1440 (12 x 120 FSB) or better yet the fabled 1600 (12 x 133 FSB) but I'm dead in the proverbial water at anything over 1200.
Just so you know, I'm started running Prime95 once again (version 2.1) about 35 minutes ago as began typing this, case temp is still 31° C and now the CPU temp is hanging in there at 30° C (it actually went down a bit).
I'm suspecting that my choice of the Vantec HSF could be limiting me in some way, but damn... not going more than 1° over ambient? How can that be a bad thing? I had a Tbird 750 running on an Asus A7V133 that idled at 40°C and 52° C at full load, and this Tualatin DOES run cooler, obviously. But... hell, I can't figure this out.
Before I make a total jackass out of myself (this could go on for a while), I'm going to cut this short in the hopes that someone might be able to offer some assistance. Not meaning to sound pushy, but if you're going to suggest I get another mobo, forget it. I should be able to make this hardware work AT LEAST 10% faster (the 12 x 110 FSB target) than what it's currently doing.
Any suggestions are welcome, and keep up the great work here in the Forum. Excellent stuff here to work with, just hope someone can point me in the direction of ... well, I'd just like to see what this hardware can do, as do the rest of you I'm sure.
:)
Many thanks in advance,
br0adband
EDIT --- I neglected to mention that with this new BIOS 1011 beta 007 I do have the vcore adjustments of 1.475 up to 1.675 in .05 increments. I have tried each of these settings with many different FSB/RAM speeds to no avail. Maybe I just put together a bad combination of hardware, who knows. That's what I'd like to find out. Thanks!
wolfsid
12-28-01, 08:41 AM
welcome to the forum seems like you will be a good asset.
Kingslayer
12-28-01, 10:55 AM
I dont think the problem is your heatsink. Your staying around your ambient temps. That's pretty good for a heatsink. Your ambient temps are a little high, but an average of 31 degrees aint bad. It's not a temp that would stop you from a decent overclock. Your not even breaking 90 degrees F. That's not bad.
But I honestly don't know what is holding you up. I agree with your speed goals. They arent overzealous.
What kind of TIM do you have on that heatsink. It's the only thing that I can see holding you up. The fact that your CPU temps were 1 degree below ambient kinda bothered me, but that sensor could be in a air path and giving an untrue reading.
Also, if you have it, heat up your CPU. Shut down, and immediately check you temps in your bios. See if they match what your monitoring software is showing. Maybe your getting hotter than you think.
And case fans would help. :)
muddocktor
12-28-01, 11:07 AM
Another welcome to the forums from me, also.
It seems that you have done your homework on the components in your system; no obvious weak spots. The only thing that jumps right out at me is the cpu temps being lower than your case temps as that is impossible without active cooling of some sort(bong cooler or TEC). I have noticed that the Tualatin procs run really cool; I have a P3 1.26S server proc and I've got it presently running at 1500 mhz with the stock Intel hsf and it doesn't even get as high as 30° C, even when Folding @Home. What temps are you showing in bios if you go reboot the system after running Prime95 for a while?
What are your symptons when trying to boot up overclocked? Do you get BSOD's or does it just sit there and not do anything? I had a BH6 board that had an option in bios called "Speed Error Hold" and if it was enabled the machine wouldn't boot up. I hope you get the problems sorted out, that is a fine setup that you have.
ol' man
12-28-01, 11:10 AM
Well I can't think of what it would be off hand! That is truly strange. Your PSU should be enough @ 350w. I guarentee that chip should do at least 1500 but you have the better version being a philli chip so 1600MHz should be ideally easy to hit.
Personally I am wondering why if you have heard the many problems people have had with the tusl2 why you decided to get it? I have owned one as others here. It has some faults! The first place I would maybe look is there. That is not a good board. It doesn't like high FSB's.
You should have got a ST6 or Gigabyte GA60XET-C. Theoretically you should hit 134fsb (yeah 134fsb) with that chip so the tusl2 should be fine but that is besides the point. The TUSL2 does not allow the PCI to be 33MHz until 134fsb and not 133 which is not usually a problem with the philli chip but you are seeing some problems.
It is really strange and I would maybe put the blame on the mobo. Not sure what else it would be. 1MHz increase should be done without a whimper. I have seen some get your chip to 133fsb on default vcore.
Once again if you can get a different mobo the ST6, GA60XET and GA60XET-C allow 1/4 dividers under 133fsb. You should try one of those boards and bring back the ASUS if you can. I had to RMA my TUSL2 mobo.
As for the temps???? They are okay but once you hit 1600MHz you may need some better cooling if your temps start hitting above 45 deg. C and the extreme limit is 50 deg. C. Anything higher and no matter the speed of the chip it will fail.
Wait I see you are running winxp?????? Maybe that is the problem??????? I don't know, I dislike winxp about as much as the TUSL2:D
pappypete
12-28-01, 11:37 AM
Welcome.....Looks like a grest system. My thoughts? start
over....install only what you need. See if you can get o/c, and
add components one at a time.
ol' man
12-28-01, 11:45 AM
Originally posted by pappypete
Welcome.....Looks like a grest system. My thoughts? start
over....install only what you need. See if you can get o/c, and
add components one at a time.
That soujnds good but he is running XP?????????? Maybe this will count it as a different configuration every time! Pretty soon after 7 times of this BS it will lock up then. Yippppeeeeeeeeeeeee. winXP sucks!
MilkPowder-2
12-28-01, 11:49 AM
Originally posted by ol' man
That soujnds good but he is running XP?????????? Maybe this will count it as a different configuration every time! Pretty soon after 7 times of this BS it will lock up then. Yippppeeeeeeeeeeeee. winXP sucks!
I second that ol'man :)
br0adband
12-29-01, 04:34 AM
The mystery continues...
Well, I came home from work tonight all fired up about getting this thing to go higher. I can get it to boot at the simpler settings under 1320, but not much else.
I pulled everything apart tonight and redid the thermal compound (again, Arctic Silver II) and was miffed to notice that in the 24 hours or so that I had this thing up and running, the heatsink on the CPU (the forementioned Vantec FCE-62540D) had marks on it, more like nicks in the metal. I then came to the conclusion that I wasn't getting a good contact between the IHS on the Tualatin Celeron and the HSF itself. I don't have the proper tools to lap these two items, so I redid the compound a *bit* thicker this time on the CPU. I also sanded down the Blue Orb I bought and finally put it on my GF3 Ti 200, same problem, the Blue Orb wasn't flat enough from the factory to ensure adequate contact with the GPU core. Same thing, except after sanding it down to a nice flat surface (albeit a rough one, couldn't find any 1000 or 2000 grit sand paper, could only find 80, 100, and 200 for the final pass). Slapped it all back together and fired it back up.
Now I can't measure the effect of the Blue Orb on the GF3, but I'm sure that now I've got a far better contact surface, it's bound to be running cooler. I'm not going to seriously overclock this video card till I get my Koolance case (www.koolance.com) in a few weeks. But what IS interesting is that since I redid the compound on the Tualatin CPU my BIOS temps are down 8° C !!! I'm running the exact same setup I have in the first post in this thread, the only thing I changed is the *amount* of Arctic Silver II on the CPU core itself. I put a thicker layer on it to account for the mismatched surfaces of the core and the HSF bottom. Hmmm... maybe I'm on to something here. Read on...
I left it running at the BIOS hardware monitoring screen for about 15 minutes and it just hovered at 38° C which is down from what it was previously idling at, 44-46° C. Wow... and I'm *still* not getting decent contact from the HSF. I'm going to ask a friend I work with here (he has some experience with lapping HSF's and CPU cores, as well as watercooling) if he'd like to assist me in this project of mine to lap the HSF and CPU core.
Here's a question for any of you who might be running this Tualatin Celeron 1.2GHz and an Asus TUSL2-C motherboard:
What are your BIOS hardware monitoring temps showing? I realize that if you go into the BIOS and then into the hardware monitoring screen and just let it sit there for 10-30 minutes that you'll get an accurate reading of the max temps for the CPU. The reason I'm pretty sure this is true is because that when just running the CPU during a BIOS screen the chip is running full bore, never idle. When running a modern OS like Win2K or WinXP or even Linux the OS will issue HLT instructions to keep the CPU idle for most of the time, thereby keeping it cooler. But in the BIOS screens this doesn't happen, so after about 10-30 minutes you should be looking at an accurate measure of your CPU's operating temps at full load. If anyone can elaborate on this further I'd appreciate it. If I'm wrong, tell me. If I'm right on the money, tell me that as well. :)
Ok, here are my current BIOS temps after 20 minutes of just sitting there watching the readings:
MB temp: 28° C
CPU Temp: 37-38° C
Under WinXP I get the following after 20 minutes of doing not much of anything, no background apps running other than the tray icons, I even uninstalled Motherboard Monitor just in case it was affecting the temps:
MB temp: 30° C
CPU temp: 25° C
What am I missing here? Do I need to configure Motherboard Monitor to recognize the discrepancy between the BIOS temps and what it's reading while running WinXP, or am I really getting a 13° C drop in the CPU temp while running WinXP (because of the HLT instructions)?????
What's getting on my nerves here is that the motherboard temps are almost always the same, within 1-3° C of each other, while the CPU temps are 10-15° different.
Am I making sense here, or just plain crazy? :P
Hope someone else can see something I'm missing here. Trying not to make these posts *too* long but I never want to be accused of not providing enough information for someone to base a solution on.
Thanks in advance for any help, system specs are posted in the first message in this thread so I won't post them again. Right now at this moment I'm still running defaults on everything across the board, just like I showed in my rigs specs, I have only increased the amount of Arctic Silver II on the CPU core to account for the non-efficient mating surfaces (I'll be lapping these puppies together soon).
br0adband
EDIT
Here's a kicker for you. I decided to take a chance on installing Asus Probe (since I'm using an Asus mobo -- Go figure). Believe it or not, it's reporting my current CPU temp at this moment as:
44° C (running idle while typing this post)
Then I played Q3 for about 15 minutes on a server with over 15 people on it. The reading:
45° C
Ok, now I'm REALLY confused. Supposedly Motherboard Monitor has been wrong all the time. Asus Probe is apparently reading the correct temps, but even after playing Q3 for 15-20 minutes full load I'm still only going up 1-3° C??? WTF is going on here...
Is my HSF so good that it's keeping my full load temps only 1-3° C above my idle temps? If it is, I'd say I'm doing rather well, but the extremely high idle temp of 43-44° C seems a *bit* high, doesn't it?
This just gets more baffling all the time... lemme go run Prime 95 for an hour or so and I'll post those results also.
Bah :(
I've found that the Tualatin Cellys cpu temps don't range very far when comparing idle to load temps. I idle around 38-39c and load temps are 42c after an hour of Sandra. I'm basically in the same spot as you; I'm case fanless, but overclocked.
I find that the BIOS temp readings are the most accurate since there is essentially, no configuration to deal with so you can't mess it up. You idling in the 40's is just too high, especially if you have adequate cooling.
From what I gather, doesn't look like you've got the sucker overclocked, so something is definitely screwy if you're peddling through in the 40's. Even if those temps are correct, your cpu temps are rather high for a non OCed chip. I agree that you should get your comp under load, watch the temps in mbm, and then confirm in the bios.
br0adband
12-29-01, 06:42 AM
Oh boy, now you folks are gonna REALLY love this one...
I decided to start fresh, completely from scratch even less than 48 hours after building this here monstrosity.
By starting fresh, I mean another OS. I dumped WinXP in favor of Win2K for testing purposes. Here are the results of what I've just happened upon.
1) Set the computer for 1.5 GHz at 1.575 vcore.
2) Booted off the Win2K Pro SP2 CD and installed Win2K Pro SP2 clean with no errors.
(read that again, I'll wait)
<insert dramatic sounds of shock and looks of amazement here>
Yes, you read correctly. I set the CPU to 1500 MHz in the BIOS, upped the vcore to 1.575 volts and went from there. I have no problems POSTing at anything other than 1600 ( it won't do that no matter what I do, I get that silly Asus *beep beep beep beep* from the PC speaker, most likely the temps again). It won't go at the stock 1.475 or 1.525 vcore, but it works fine at 1.575. Never got around to testing it because I wanted to install my video drivers (for my GF3 Ti 200, again read the first post for my rig's specs). And guess what? It stopped cold on the first reboot with more current drivers. With the default VGA driver I can boot into Win2K Pro SP2 every single time with no errors or lockups. Ran several Sandra 2001 CPU and memory benchmarks, ran Prime95 for a short time, but if I boot to any drivers other than the default VGA I'm dead in the water. I can boot into Safe Mode or Safe Mode with VGA anytime. Weird, isn't it?
Yep, we're narrowing it down one part at a time.
I'm thinking it's now the video card that's holding me back for some reason, maybe it just can't do anything over the AGP spec for bus speeds. If this is true, I'm gonna be peeved, but I'll live with it. Hopefully someone reading this will be able to offer some assistance. I'm thinking about upping the VIO voltage on this mobo to 3.60 (which is the max, set by a jumper on the TUSL2-C), but I'll wait a bit until you people have a chance to read this.
Just to summarize: I installed Win2K Pro SP2 a short time ago, with the CPU running at 1.5 GHz with vcore at 1.575 with no errors. However, this is with the default VGA driver in Win2K. As soon as I installed the most current Detonator XP drivers (23.11) or even trying the drivers that came with my GF3 Ti 200 (21.83) at default speeds I get nowhere. I can POST at 1.5 GHz just fine, even get into loading Win2K but then it just craps out dead. BUT, again, when I reboot then choose VGA Mode I get into Win2K just fine, no errors, able to run just about anything with no problems.
I haven't reinstalled Asus Probe yet, I'm about to do that right now and reboot into VGA Mode to watch the temps when I run Prime95 for a bit.
I'll keep you folks posted.
br0adband
br0adband
12-29-01, 07:07 AM
Ah baloney...
How lame can you get. Asus Probe won't even run in VGA mode. Great, just great.
Well, right now I'm running at 1.5 GHz with 1.575 vcore as stated above. Can't test the temps with Asus Probe because it won't run, decided not to install Motherboard Monitor since it's so obviously off-base as far as temperature readings are concerned.
I'm online, running Prime95 and CPUBurn at the same time, Explorer, IE and mIRC all running comfortably, no errors, no lockups, nothing.
Anyone able to offer any advice whatsoever? Talk about confusing... it's fine at this speed unless I boot to the normal video drivers.
I am completely stumped on this one. Soundcard works, CD burner works, hard drive is up and running, Sandra benchmarks are right where they should be... I'm just stumped.
br0adband
Godfodda
12-29-01, 10:39 AM
I don't know **** about vid cards, but this point popped into my head. Maybe it's a 4xAGP issue with the card? I don't know how they work, but maybe the VGA drivers don't enable 4x mode but the normal drivers do, and the combo doesn't like that. I don't know, but I am reading with great interest. Bumpity-bump.
:EDIT: Maybe you could disable 4x in the BIOS to check, if you think it might be part of the problem. Forgive my ignorance if I'm way off on this.
pappypete
12-29-01, 10:58 AM
Glad to see your gaining.......I to have a Vantec HS/F, 6035D, at the moment it's on my desk. I believe the markings on the HS
are the result of vibration, caused by high RPM fan. I'm in the process of installing rubber "O" rings between HS and fan, also,
under screw heads. This might reduce vibration = better contact
with CPU.
ol' man
12-29-01, 01:26 PM
Originally posted by br0adband
Ah baloney...
How lame can you get. Asus Probe won't even run in VGA mode. Great, just great.
Well, right now I'm running at 1.5 GHz with 1.575 vcore as stated above. Can't test the temps with Asus Probe because it won't run, decided not to install Motherboard Monitor since it's so obviously off-base as far as temperature readings are concerned.
I'm online, running Prime95 and CPUBurn at the same time, Explorer, IE and mIRC all running comfortably, no errors, no lockups, nothing.
Anyone able to offer any advice whatsoever? Talk about confusing... it's fine at this speed unless I boot to the normal video drivers.
I am completely stumped on this one. Soundcard works, CD burner works, hard drive is up and running, Sandra benchmarks are right where they should be... I'm just stumped.
br0adband
Well man I told you that 1500MHz is the limit you are going to hit about with that chip if you can't hit 134fsb on the TUSL2. You don't have the neccesary dividers under134fsb. No sense in beating your self up about it. That is why I said that the tusl2 sucks!!!!!!!!!! You need to hit 134fsb now or you will not get higher. The tusl2 hits its max PCI at around 38~40MHz. Do the math at 125fsb you are sitting at 41MHz PCI which is very good for the tusl2 but you do not have 1/4 dividers under 134fsb with the crappy TUSL2. EIther you hit 134fsb or you call it a day(year whatever) cause you will not hit higher with out reaching that goal. The tusl2 sucks, I can't stand that piece crap board. I know you got it for cheap but that is what the board is, CHEAP!
Do you understand what your problem is here?
Hit higher hurtz on the fsb or quit dinking with it. Maybe up the vcore to 1.7v or better.
If you get around 45~50 deg. C you will experience instablity problems too especially OC'd to 1500+MHz.
Also don't jam more thermal paste into the HSF thinking it is going to solve any overheating problems. You need to lap that equipment if it is not touching 100%.
The tusl2 mobo problem should have a sticky.
I have warned many people over and over about the tusl2 and its crappyness but they do not get it. If you have a chip that you know will do 134fsb then it will be fine but if not ..... well we get what we have here!
MilkPowder-2
12-29-01, 03:52 PM
Originally posted by ol' man
Well man I told you that 1500MHz is the limit you are going to hit about with that chip if you can't hit 134fsb on the TUSL2. You don't have the neccesary dividers under134fsb. No sense in beating your self up about it. That is why I said that the tusl2 sucks!!!!!!!!!! You need to hit 134fsb now or you will not get higher. The tusl2 hits its max PCI at around 38~40MHz. Do the math at 125fsb you are sitting at 41MHz PCI which is very good for the tusl2 but you do not have 1/4 dividers under 134fsb with the crappy TUSL2. EIther you hit 134fsb or you call it a day(year whatever) cause you will not hit higher with out reaching that goal. The tusl2 sucks, I can't stand that piece crap board. I know you got it for cheap but that is what the board is, CHEAP!
Do you understand what your problem is here?
Hit higher hurtz on the fsb or quit dinking with it. Maybe up the vcore to 1.7v or better.
If you get around 45~50 deg. C you will experience instablity problems too especially OC'd to 1500+MHz.
Also don't jam more thermal paste into the HSF thinking it is going to solve any overheating problems. You need to lap that equipment if it is not touching 100%.
The tusl2 mobo problem should have a sticky.
I have warned many people over and over about the tusl2 and its crappyness but they do not get it. If you have a chip that you know will do 134fsb then it will be fine but if not ..... well we get what we have here!
I'm glad you mentioned this again, because you are correct.
Olfart(not Ol'man) was aware of this and avoiding the problem with auto PCI frequency divider with some boards including ASUS TUSL2C. He was worrying about what if he can't OC high enough where he might have to run it at upper 120's FSB upto 133 fsb and the board would get stuck in its PCI divider@ 1/3, that would be problem. This is what olman is talking about also. ST-6 has manual override on the PCI n AGP divider. TUSL2 don't. My Abit BE6-2 BX mobo has manual override on the those settings too. So what Ol'man is trying to say about pushing it to 134fsb or Up is correct because your TUSL2 will auto kick-in to 1/4 PCI starting at 134fsb. Either you go higher or forget it. I have seen similar matter a few years ago!!! back then didn't know what the hell was causing the problem. Also, some mobos don't handle well at high PCI frequency setting something like Zip drive not working properly, or on-board audio get wackkkked.
br0adband
12-29-01, 07:39 PM
Oh well, I'm still completely baffled by this. I left it running in VGA Mode for over 10 hours with CPUBurn, Prime95 and 3DMark 2001 running continuously. Not one problem during that whole time.
I'm inclined to believe it's the video card right now. A friend is bringing his TNT2 over tonight for me to run testing on that card to see if it's better able to handle the 41 MHz PCI speeds. If that's what it turns out to be and my GF3 Ti 200 can't do it, I'll return the card as defective and get another one. MSI makes decent products, although not the best of the best. Most people I've read about have had great luck with the Gainward Golden Sample GF3 Ti 200 card, so I might get that one instead. The PNY GF3 Ti 200 is actually my card, it's manufactured by MSI for them.
Other than this, I can't think of any other reasons this won't come up at 1.5 GHz with real video drivers. Again, if anyone has suggestions, let me know.
br0adband
wolfsid
12-29-01, 07:54 PM
Originally posted by br0adband
Oh well, I'm still completely baffled by this. I left it running in VGA Mode for over 10 hours with CPUBurn, Prime95 and 3DMark 2001 running continuously. Not one problem during that whole time.
I'm inclined to believe it's the video card right now. A friend is bringing his TNT2 over tonight for me to run testing on that card to see if it's better able to handle the 41 MHz PCI speeds. If that's what it turns out to be and my GF3 Ti 200 can't do it, I'll return the card as defective and get another one. MSI makes decent products, although not the best of the best. Most people I've read about have had great luck with the Gainward Golden Sample GF3 Ti 200 card, so I might get that one instead. The PNY GF3 Ti 200 is actually my card, it's manufactured by MSI for them.
Other than this, I can't think of any other reasons this won't come up at 1.5 GHz with real video drivers. Again, if anyone has suggestions, let me know.
Br0adband
I have had that same problem once before some vid card don't like the higher mhz, the way i fixed the problem is bought a asus card they handle the a higher mhz when overclocking and haven't had a problem with it yet.
I don't think TNT2's like out of spec FSB's either....Try a Geforce 2 MX or higher card....
ol' man
12-29-01, 11:39 PM
Originally posted by br0adband
Oh well, I'm still completely baffled by this. I left it running in VGA Mode for over 10 hours with CPUBurn, Prime95 and 3DMark 2001 running continuously. Not one problem during that whole time.
I'm inclined to believe it's the video card right now. A friend is bringing his TNT2 over tonight for me to run testing on that card to see if it's better able to handle the 41 MHz PCI speeds. If that's what it turns out to be and my GF3 Ti 200 can't do it, I'll return the card as defective and get another one. MSI makes decent products, although not the best of the best. Most people I've read about have had great luck with the Gainward Golden Sample GF3 Ti 200 card, so I might get that one instead. The PNY GF3 Ti 200 is actually my card, it's manufactured by MSI for them.
Other than this, I can't think of any other reasons this won't come up at 1.5 GHz with real video drivers. Again, if anyone has suggestions, let me know.
br0adband
Well I gues I can lead ya to water but I can't make ya drink! The tusl2 does not like PCI speeds above 38~40MHz. That is not hard science. I am done this thread is hopeless!
Godfodda
12-30-01, 12:57 AM
Originally posted by ol' man
Well I gues I can lead ya to water but I can't make ya drink! The tusl2 does not like PCI speeds above 38~40MHz. That is not hard science. I am done this thread is hopeless!
Geez. Settle down, ol'man. Have a bit of "prune juice". :beer: Not everyone will drop everything just because one person says all is lost. Don't get me wrong, I'm NOT belittling your contribution to the forums. I really enjoy reading your diatribes. But we're overclockers, and that means we think there's gotta be SOME way to beat the damned machine (hopeless as it may be). :D
You havin a bad day ol'man?
ol' man
12-30-01, 01:25 AM
The only way to beat it is too up the vcore and get the fsb to 134MHz. Plain and simple!
MilkPowder-2
12-30-01, 01:34 AM
Hey check out my siggy. It's done. hehehe.. :D :burn:
br0adband
12-30-01, 01:39 AM
Well, maybe ol' man has something there.
I wonder myself because of this interesting thing I've discovered.
I don't have 134 MHz as an option for an FSB setting. Weird, isn't it? This BIOS on the TUSL2-C I have, version 1011 beta 007, doesn't list 134 MHz as a selectable speed when I'm trying to manually set the FSB, it goes from 100 up to 133 then SKIPS 134 and goes to 135 MHz and higher.
Like I've said before, it just gets more weird and more confusing every time.
So now I'm wondering if Asus knows something about this board that most people don't (not counting ol' man there, he seems to be the resident *expert* on the TUSL2-C since no one else has piped up about these boards). I'm still lost but continuing on in my quest for 1500 with this board.
br0adband
MilkPowder-2
12-30-01, 01:59 AM
Have you tried 135mhz FSB yet? If you haven't, try it. =)
br0adband
12-30-01, 03:59 AM
Ok, for those people in the audience, drum roll please...
<sound of drums a'rolling>
Guess what? I'm running in Win2K with current Detonator XP drivers at 1.5 GHz, which has been my goal all along.
How did I do this? Keep reading...
I started doing simple surfing around for various information about this TUSL2-C board and came across info about *tweaked* CUSL2 BIOS modifications. I spent some more time looking around and found out that there was one specifically for the TUSL2-C (my mobo) and decided to give it a shot.
I am happy to announce that I am currently running stable (at least for the past 30 mins now with Prime95 running up a storm) at 1.5 GHz. I flashed a BIOS made by someone called "The MAD" over at http://www.asusboards.com in the BIOS forum and was stunned that it's working. I'm not sure what this guy is doing, but I hope he keeps doing it. Here is the specific link to the page with the TUSL2-C tweaked 1011 beta 007 BIOS I'm currently running WITH GREAT SUCCESS:
http://www.x86-secret.com/articles/tweak/i815twken.htm
I had to up the vcore to 1.575 to get it running at this speed, as I put it at 1.525 and it locked up whenever I ran Q3 or Prime95 for more than just a few minutes. So far at 1.575 vcore I'm still up and running, going on 50 minutes so far. I know this isn't a SERIOUS test of stability, but hell, just an hour ago I couldn't do 1.5 GHz at all. :D :D :D
Thanks for all the help from you people here, I'll still be aprising all of you of my progress as I continue this overclocking journey of mine.
I can't get the system to even post at 1.6 GHz (12 x 133), I keep gettting that "beeboobeeboobeeboo" sound from the PC speaker. I'm assuming that's some kind of alarm letting me know the settings I've chosen just ain't working, but does anyone know if it's specifically a temperature alarm or something else? The manual is pretty vague in some respects.
Again, any assistance is greatly appreciated.
br0adband
ps I'm about to wipe this puppy and install WinXP to make sure XP can handle the speed. I'll keep you folks posted as usual.
Wish me luck! :eek: :eek:
Congrats dude....Good to see you pushed on and kept trying to solve the problem instead of simply giving up! And by the looks of things you won!
Just wondering....Are your 1/4 pci dividers working under 133 mhz FSB?
br0adband
12-30-01, 04:29 AM
Well, I suppose my first question would be:
How the hell would I know? :)
Right now, at 1.5 GHz the FSB/RAM/PCI settings are 125/125/41, so I know it's working as far as the 1/3 divider is concerned. Not sure how to discover the 1/4 divider since I don't see any options for that sort of thing in the BIOS.
Clue me in on what to look for and I'll let you know.
br0adband
br0adband
12-30-01, 04:35 AM
DAMN!!!
I just looked at your system specs Fluid...
33° C with the stock cooler? Geezus... what the hell am I doing wrong?
It's running Prime95 right now, has been for over 2 hours and it's sitting at 48° C according to Asus Probe. Christ I need to calibrate either this program or Motherboard Monitor (which I would prefer to use anyway).
br0adband
Originally posted by br0adband
DAMN!!!
I just looked at your system specs Fluid...
33° C with the stock cooler? Geezus... what the hell am I doing wrong?
Personally, I feel that Fluid's got a buttkicking chip. If he can run that sucker @ 1.5 with only 1.575v, I'm hardly surprised he can get those kind of temps. Although, it wouldn't hurt to know what kind of case cooling he's got in that demon of a tower.
To tell you the truth, I'm still fazed at the temps he gets, especially with what he's cooling it with, and where he resides.
Fluid, I'm on to you ;)
:beer:
I'd like to know more, too. Mine surely won't do that. I've got a nice hefty Copperfinned cooler mated to a 32cfm 60mm fan and today is absolutely the coolest I've ever seen my chip run. I'm sitting in a 68 degree F room and it's idling at 35C. Motherboard temp is 25C.
During warmer times when the temp in this room approaches 80F or so, it's common for me to watch this chip, clocked at 1404 at default voltage, to consistently idle at or around 100F.
This bothers me, because the Coppermines I've run in the past on this same setup would idle at or around 87-89F. I've got excellent case cooling and airflow, and have even lapped the IHS on my chip, but I can't seem to get much improvement from anything. To be perfectly honest, I could run it with the stock cooler, because there's only 1 degree C difference in temps. Also, if I clock the chip back down to default settings all the way, this only results in a 1 degree C or less drop in temps.
The only bright point about all of this is that the chip doesn't seem to mind the heat. I can stress it all day at 1.4ghz in this room with an ambient temp of 82F+ and it never falters. It maxes out at or around 45-46 C under heavy load(CPU Burn and Prime 95 torture test running)and seems extremely comfortable.
I just wonder what the effects of this higher heat will have on the life of the chip. I know I won't own it long enough to care, but someone else probably will and that could wind up a concern of mine.
If anyone has any secrets, short of removing the IHS, I sure wish they would share them with the rest of us.:)
Rick
ol' man
12-30-01, 12:39 PM
Originally posted by br0adband
Well, maybe ol' man has something there.
I wonder myself because of this interesting thing I've discovered.
I don't have 134 MHz as an option for an FSB setting. Weird, isn't it? This BIOS on the TUSL2-C I have, version 1011 beta 007, doesn't list 134 MHz as a selectable speed when I'm trying to manually set the FSB, it goes from 100 up to 133 then SKIPS 134 and goes to 135 MHz and higher.
Like I've said before, it just gets more weird and more confusing every time.
So now I'm wondering if Asus knows something about this board that most people don't (not counting ol' man there, he seems to be the resident *expert* on the TUSL2-C since no one else has piped up about these boards). I'm still lost but continuing on in my quest for 1500 with this board.
br0adband
Hey dude when you set it too 133fsb it automatically goes to 134fsb in reality! That is why I have said you need to hit 134fsb. You are getting better at listening anyway! There is no 134fsb setting in the bios so indeed 135fsb is really 136fsb in reality. You have to remember this. The reason it was so importent to me was because I had a chip that would do 133fsb stable but not 134fsb. In that case I was screwed and limited to 125fsb or lower.
THE TUSL2 DOES NOT ALLOW 1/4 DIVIDERS UNDER 133FSB! HOW MANY TIMES DO I HAVE TO SAY THIS?
You need a very good HS for that chip. I feel like someone needs to do more homework, hint hint!
The trick to running that chip at 1600MHz to have a very good HS(COPPER BASED PREFERABLY), 300+w(350w ideally) PSU and very good contact between the IHS and HS. In the latter case a very simple and fine lapping of the IHS may suffice but done very carefully! Most cannot hit 1600MHz with a 250w PSU!
Originally posted by br0adband
Here is the specific link to the page with the TUSL2-C tweaked 1011 beta 007 BIOS I'm currently running WITH GREAT SUCCESS:
http://www.x86-secret.com/articles/tweak/i815twken.htm
any probs using this bios??? do you flash just like normal
and what is your voltage adjustment range?
To be honest temps usually range between 33 and 35 degrees under load depending on the ambient temp....The other day was a 35 degree C day so it was pretty warm in my computer room and temps went to 37 degrees C. But most of the time 33 is the norm with 35 at times....
This chip idles 'alot' hotter than my P3 1000EB (29 - 30 degrees) cumine but the temps barely seem to fluctuate under load, whereas at times the cumine would start approaching 40 degrees C.
I've checked MBM 5.1 with the temps in my bios and they're identical so there is no inaccuracies there....
I'm not too sure why my temps are so good as i've only got one intake and ine exhaust fan in this monster of a case....The only thing I can think of is this mobo, when I changed from an XE to an XET with my cumine my temps dropped 3 degrees C! And yes, thermal paste application and heatsink fitment was right on in both cases....
Also....I reckon once I get my PSU sorted out I should be able to drop my vcore a little....
I guess all I can say is....
....My chip runs cool, my chip runs cool! Nah, Nah!:D
br0adband
12-30-01, 06:04 PM
Well, it's still up and running at 1.5 with 1.575 vcore, no issues noted since my last report.
Fluid: Can you do me a favor if you get a chance? Fire up your machine and go into the BIOS. Go into the Hardware Monitoring and just let it sit there for 10-20 minutes and record what temps are showing. Then fire up your PC into Windows for another 10-20 minutes (idling, not doing much of anything) and tell me what you're getting.
I'm still wondering about the temps, personally. Right now in the BIOS I'm showing 46-48° C after 10-20 minutes of just watching the Hardware Monitoring. In WinXP MBM (Motherboard Monitor) will show 32-34° C, so I KNOW somewhere there is a pretty serious discrepancy, but I haven't been able to nail it down.
I'll be lapping the IHS and the HSF later this week, but for now I'm happy. :eek:
br0adband
Done....
I get 30 in the bios and 30 in windows sitting for 10-15 mins....
ol' man
12-30-01, 07:06 PM
Room temp 24 deg. C with one intake and one exhaust are 35.5 deg. C at idle and 36.5 deg. C at load. cel 1.2 @ 1.5GHz 1.7v. Bios is the same around 34 deg. C.
br0adband
12-30-01, 07:46 PM
Well, I certainly hope the lapping job makes a difference. I just can't believe that this Vantec HSF is that bad, it received *good* reviews everywhere I could find one. Not *great*, but *good*.
All this work is just a stopgap measure anyway till I get a Koolance case, but it's nice to actually work on this stuff these days instead of just reading about it all the time. :)
br0adband
i have very similar setup.
Running at 1534 Mhz at 1.82 volts.
1600 posts but hangs.
Im getting arctic silver 2 and do some lapping today.
Have dragon orb 3.
Will let u know if i hit 1600.
Have u done the wire trick to get higher voltage.
br0adband
01-05-02, 04:09 AM
Nah, didn't think I'd need to get that drastic since I can do 1500 with 1.575 vcore.
I'm still not 100% sure exactly what is causing me to *not* hit 1600 since I can't even POST to it. I get that *beeboobeeboobeeboo* alarm from the mobo. I asked this once already, and since this post is getting buried by more and more new ones, I'll ask again:
Does anyone that owns an Asus TUSL2-C mobo know what that *beeboobeeboobeeboobeeboo* alarm coming from the PC speaker means when you try to overclock? Does it specifically mean a temperature issue, or maybe a voltage issue? I'm not having any luck finding an answer to this one, the manual is pretty vague about boot errors.
Any help would be appreciated.
UPDATE: Still running comfortably at 1500 right now, lapped the CPU ISF down to bare copper and worked on the HSF bottom also. Still kind of rough since I don't have sandpaper above 400 grit, but I'll be getting some much finer grit stuff in the next few days (600, 1000 and 1500 grit) so I'll be redoing it yet one more time. :)
br0adband
6502kid
01-05-02, 04:39 AM
I am starting to be way skeptical about any temp any
motherboard tells me about.
I can fire up my Abit VT6X4 and Soyo TISU in the same room,
and the Abit tells me case temp is 19c. Soyo tells me case
temp is 28c. Same enlight cases. Both with 3 or 4 extra
fans.
It the soyo system, it would always show the CPU temp at 40-42c
I dropped the same cpu and heat sink into this Gigabyte board
and it tells me CPU temps are 25c.
Who knows. As long as it dont melt I guess.....:mad:
ol' man
01-05-02, 12:33 PM
Originally posted by br0adband
Nah, didn't think I'd need to get that drastic since I can do 1500 with 1.575 vcore.
I'm still not 100% sure exactly what is causing me to *not* hit 1600 since I can't even POST to it. I get that *beeboobeeboobeeboo* alarm from the mobo. I asked this once already, and since this post is getting buried by more and more new ones, I'll ask again:
Does anyone that owns an Asus TUSL2-C mobo know what that *beeboobeeboobeeboobeeboo* alarm coming from the PC speaker means when you try to overclock? Does it specifically mean a temperature issue, or maybe a voltage issue? I'm not having any luck finding an answer to this one, the manual is pretty vague about boot errors.
Any help would be appreciated.
UPDATE: Still running comfortably at 1500 right now, lapped the CPU ISF down to bare copper and worked on the HSF bottom also. Still kind of rough since I don't have sandpaper above 400 grit, but I'll be getting some much finer grit stuff in the next few days (600, 1000 and 1500 grit) so I'll be redoing it yet one more time. :)
br0adband
YOU WHAT! You didn't need anything more corse than 1000grit to get the IHS down where you wanted it. Wow! Take it easy on that chip man! A nice mirror smooth IHS is the best way to do it. Not rough! I would even try it when it is rough!
yeah i don't know how he got it that low mine hovers around 29 idle and 43 or 44 under full load.... awesome chip or something.....
6502kid....
Don't use the Gigabyte hardware health utility....
Use MBM 5.1 (latest)....The Gigabyte utility always reads 25 degrees on Tualatin based chips, don't know why!
If you want to know the settings just ask me....
Celemine1Gig
01-06-02, 10:04 AM
Settings for MBM 5.1 wouild be great!! Could you post them or mail them to me??
My e-mail adress! (ingmarklein@aol.com)
Thy in advance
Ingo
I used the Gigabyte Utility and Speedfan and they both always read out 25°C and I have a coppermine chip!
I am having discrepancies with the hardware monitors also.
Asus Probe shows 39c idle , MBM5.1 shows 39c idle
MBM4.18 shows 26c and Sandra shows 24c idle.....who do you trust? I find it hard to believe that with 26c mb temp that the cpu is at 39c (at idle) cause I am using Alpha Pal 6035 HS with 40cfm fan that runs at 8300 rpm....
that is Tusl2-c and cel 1.2-t
Is there a way to accurately find out which proggie is accurate?
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.