Author Topic: Need An Intel P-4 Expert.... (Read 747 times)

Cabby44 · « **on:** August 25, 2003, 10:32:25 PM »

Put together a P-4 3.0C and an Abit IC7-G Mobo and it sucks. Big time.

System:
Intel P-4 3.0C 800 FSB Processor Hyper-Threading enabled
Abit IC7-G MaxII with ver. 16 BIOS
Paired OCZ EL DDR PC-4000 / 500mhz / Gold Edition / Dual Channel 256MB Sticks populating all 4 slots=1GB
Maxtor 40MB ATA-133 HDD in SATA(using SATA Adapter)
WinXP Pro SP-1
Audigy2 Platinum
Radeon 9700 Pro AGP 8x
DX9.0b
Antec 480W PSU

My Sandra(latest version) scores are around 1100/688 Dhrystone/Whetstone which is utterly RIDICULOUS and HDD/Memory performance is abysmal. The machine will barely run 3DMark2003 and takes so long to load the benchmarks i finally just exited 3DMark when CPU Test Two ran at 1FPS. What a joke.

My AMD 2800+ Barton/Asus Nforce2 400mhz FSB machine gets 3DMark2003 scores in the 5400 range with the SAME Radeon 9700 Pro vid card i have in the Intel machine.

I've tried everything and the CPU just won't perform. WCPUID identifies the CPU correctly as does Sandra but the CPU performs like a Pentium II 450. The BIOS posts with CPU @ 3.01 Mhz "Memory in Dual Channel 200/200". The OCZ Memory is also correctly identified in Sandra. I've tried CPU default settings, NB Strap, manual settings, and the same with the memory. I'm at my wits end. Should i RMA this piece of crap or am i missing something?? I've read the manual from cover to cover and don't see anything obvious i've missed.

Cabby

mipoikel · « **Reply #1 on:** August 26, 2003, 12:36:09 AM »

First thing I would do is to test HD in normal IDE channel. I dont think SATA channel is any better if you dont use SATA HD.

bloom25 · « **Reply #2 on:** August 26, 2003, 01:59:26 AM »

IDE performance shouldn't have a major effect on CPU synthetic benchmarks.

My first thought is that the P4 is possibly clock throttling because it's running way too hot. What is the CPU temperature Cabby?

Cabby44 · « **Reply #3 on:** August 26, 2003, 03:48:49 AM »

^51 deg. C

Cabby

Skuzzy · « **Reply #4 on:** August 26, 2003, 10:14:35 AM »

That is pretty hot for a P4. Is that the idle temperature or when playing games? Either way, it is pretty hot.

Never seen a P4 run that hot before. Odd.

Siaf__csf · « **Reply #5 on:** August 26, 2003, 12:22:17 PM »

Take it from this P4 expert - Buy AMD.

LoneStarBuckeye · « **Reply #6 on:** August 26, 2003, 01:23:59 PM »

I've got exactly the same system, except for the hard drives (two 10k rpm Raptors in RAID 0 configuration) and the video card (9800 Pro). Mine runs at 43 C (reported) with four monster case fans (Vantec Tornados) and a monster processor fan (same fan). (I've got a four-channel rheostat so that I can adjust the fan RPMs. Otherwise, the noise would be deafening.)

There were considerable problems with the IC7-G board reporting the wrong temperature when it was first released, but I thought they remedied them with subsequent BIOS releases. You may want to check the Abit IC7 forums, as there has been much discussion of this topic.

Having said that, you shouldn't have any physical problems running at 51 C (i.e., you shouldn't damage the processor). There are reports on the Abit site of P4s running MUCH warmer than that. Indeed, before I added the monster fans and swapped out the default heatsink, my processor would idle at about 48-49 C and peak in the mid-50s during heavy utilization.

One possible culprit is the hardware monitoring software (It's called Hardware Doctor or something like that) that comes with the motherboard. I know that it has threshold levels and one of the options may be to throttle the processor if a preset temperature level is exceeded. You may want to check it out.

Good luck!

qts · « **Reply #7 on:** August 26, 2003, 02:41:14 PM »

Did you install the motherboard drivers first? They must be installed immediately after installing the OS.

bloom25 · « **Reply #8 on:** August 26, 2003, 05:53:51 PM »

If that's 51C at idle, and the temperature probe is accurate, that's too hot. P4s will begin clock throttling at around 60C. Running a synthetic benchmark like Sandra (100% CPU load) may get temperatures to 60C pretty quickly.

Cabby, try running the system with the side of the case off and see if that makes any difference.

Qts is right about the chipset drivers as well. (You should have them on the CD that came with the board, or you can download them from Intel directly.) There are actually two drivers, Intel Application Accelerator (IDE driver, won't effect CPU benchmarks) and the actual chipset drivers. They should typically be installed right after applying your Windows service packs.

Cabby44 · « **Reply #9 on:** August 30, 2003, 07:46:12 PM »

Thanks for the replies, guys.

I had an interesting problem to be sure. I removed the HSF, cleaned-off the Intel-supplied thermal grease, put on a fresh dab of Arctic Silver, and re-mounted the HSF. Booted the machine up and ran a quick Sandra CPU bench. The results were right where they were supposed to be with CPU temps about 118 deg F. Yayyy!!!

Believing the problem to be solved(i figured i made a botch of the HSF installation but i didn't know how, i've installed HSF's of all types on numerous types of Mobos for years with no problems), i re-assembled the machine and fired it back up. Again a very slow boot-up and i ran another Sandra CPU bench and got dismal results.

Oh hell, WTF is going on?? I again slid the mobo out and removed the HSF. You ain't gonna believe this. Stuck lightly to the bottom of the copper heatsink with thermal grease was my brand-new P-4 3.0C chip!! I couldn't believe it!! Before i could react the CPU fell-off the HSF and hit the case with a bang which mashed several of the pins on the CPU rendering it useless.

I couldn't figure out what happened. The ZIF locking-arm was down and locked. Was the Abit's ZIF socket defective?? I called the reseller and told her what happened and she didn't believe the story. I admitted i had never heard of such a thing happening either.

I RMA'd the motherboard and called Intel(i almost always buy Retail CPU's for the warranty)and told them the story. I reckoned that a pin on the CPU must have been defective or bent and i hadn't seen it or the Abit's ZIF was defective, even though the CPU dropped right into the ZIF with no problem.

Intel said to ship them the CPU and they would look at it and determine if they would replace it depending on the outcome of my reseller testing the Abit's ZIF socket with a new CPU(i told my reseller i refused to put a new CPU in that socket and that they were welcome to try it out themselves. No way i was going to risk another chip on that motherboard). If the thing tests ok they will return the mobo and if it's defective they will replace it. They(ATACOM of Fremont, CA.) are being very co-operative.

I find it hard to believe the machine even booted-up with a bad, defective installation which allowed the CPU to be literally ripped-out of the ZIF socket. No wonder the thing performed so poorly.

BTW, Intel said that in the interest of customer relations they would more than likely replace the damaged CPU. Man, i hope so. In any case, i'm sidelined from PC gaming until i get this thing sorted out.

Until then, i wait.....

C.

Skuzzy · « **Reply #10 on:** August 30, 2003, 09:30:00 PM »

That's just bizarre.

Cabby44 · « **Reply #11 on:** August 30, 2003, 11:08:02 PM »

^What's more bizarre is that since my little incident i have heard of others accidentaly pulling their CPU's out of ZIF sockets because ceramic-based thermal grease had bonded(due to heat) to the HSF...

C.

LoneStarBuckeye · « **Reply #12 on:** August 31, 2003, 09:26:55 AM »

A similar thing happened to me. When I first installed the CPU, I mounted the retail heatsink with artic silver. After seeing my initally high temps, I decided to replaced it with an Alpha heatsink and high-volume fan. When I attempted to remove the heatsink, it ripped the processor out of the ZIF socket, with the locking arm engaged. I thought the heatsink and processor were fused at that point, but I managed to get the heatsink off of the processor with a little work. I didn't bend any pins, but I got some of the Arctic Silver on some of them. I cleaned off the pins and the top of the processor with some alcohol, reapplied Artctic Silver to the top and mounted the new heatsink and fan. I really didn't expect the mobo to boot up, but it has worked just fine since then, with somewhat lower temps.

bloom25 · « **Reply #13 on:** August 31, 2003, 03:55:26 PM »

I heard of this happening on Socket 478 on more than one occasion. If the heatsink I am using comes with a thermal pad (phase change pad), I always remove it. On P4s it tends to bond so securely to the top of the heatspreader that you can tear the CPU right out of the socket if you aren't careful. (I have heard of pins being pulled right off the bottom of the CPU.) You have to be really careful on Athlons as well, because you can break the CPU die itself if you try to pry the heatsink off.

I typically just use standard heatsink compound, but sometimes Artic Silver. I can't say that I've ever had either of them bond well enough to pull a P4 out of it's socket, so that's new to me!

Gixer · « **Reply #14 on:** August 31, 2003, 07:16:15 PM »

Why buy a P4 anyway? From my understand percentage increase over an AMD on average is like 10% and you'd only notice that during CAD,Graphics rendering work. Seems you pay alot more for a P4 system over an AMD for some high end performance you'd never see anyway.

At the moment I'm building a AMD 2.5 Barton (overclock to 3+) system with Epox board and eventually water cooling. Once I get it up and running I'll post some bench marks.

As for your P4, what kind of cooling are you using?

...-Gixer
-Hells Angels-