gdp's Comments

Dr. Alison Marr : A Rainy Day in Victoria

Mon, 9 Aug 2010 13:14:35 +0000

More great photos. Looks like you two had a wonderful time. Hope to hear from Jeff when he gets back to the brutal heat here in Austin. Warning! At least three out of the next five days (Monday August 9 through Friday August 13, look out!) will feature triple digit temps. Ouch! --Ed--

Dr. Alison Marr : Our Little Cabin in the Woods

Sat, 7 Aug 2010 13:49:59 +0000

Keep up the good work, guys. Sounds like you're having at least as much fun (and the same kind of weather) as we did on our recent trip to Maine. Word of warning: you won't want to come home to the "no relief in sight from triple digit temps" that currently passes as our so-called weather forecast. Ouch!____See ya soon!__--Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 16:24:11 +0000

You're welcome! Comments about my sanity, troubleshooting skills, and general competence notwithstanding, I really enjoyed--and continue to enjoy--this entire adventure! --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 16:20:22 +0000

Had that system running for 2-plus years before making the 32 to 64 bit switch. HW Monitor reported temperatures of 27-37 Celsius at the cores, and never over 40 Celsius at the socket. Just re-ran HW Monitor on that machine and temps are as follows (wish I could just stick a screenshot in here): mobo: 24-27 CPU socket: 25-32 CPU core 1: 35-38 CPU core 2: 33-35 CPU core 3: 38-41 CPU core 4: 38-41 (all preceding temps Celsius) I *seriously* doubt it was thermal. I always watch my temps and that build ran cool for so long in advance of the switchover that I'd be surprised if temps caused my problem. Besides, why would the same cooler/CPU work (and install) Win7 x86 flawlessly but bork on x64. Surely, both would overheat at more or less the same time? Interesting hypothesis, though. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 16:05:34 +0000

I did try the original configuration with only the CPU different. Same problem. Thanks for asking, --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 16:04:47 +0000

Thank you. Thank you. Thank you. Nice to hear my own longstanding experience vindicated by somebody else. This is the first "partial CPU failure" I've ever experienced. Numerous dead CPUs over the years (occupational hazard when you're assigned to overclock systems) but none that worked mostly, except for this one little thing... To me, at least, that's what made this story interesting and why I pitched it to Esther in the first place. Thanks again, --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 16:02:33 +0000

The original CPU runs 32-bit Win7 fine, but it won't permit me to install Win7 x64 as documented. That's what I know for sure anyway! --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:54:28 +0000

I don't know why the part dislays its particular and unique (in my experience, and in my research of available information, forum postings, and so forth). I had hoped to get Intel to dig in and see what was up but they were emphatically not interested. If you have some ideas on how I might go about testing further, I'm certainly game to try. Thanks for the feedback. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:52:47 +0000

Gee that's pretty much the order I followed myself, except for putting the mobo and PSU ahead of the CPU. As I have now stated in replies to numerous other posts, I'd actually forgotten the chip was an engineering sample, because it had been working without any problems for over two years (in 32-bit mode, however). I thought it was an interesting exercise in troubleshooting nevertheless, and wrote it up as such. Thanks for adding some cogent and relevant advice to the overall content of the story. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:50:16 +0000

Hey! I used that system for 2-plus years on 32-bit OSes without any problems at all. I started working in IT in 1981/2, possibly before you were born. If the system hadn't worked fine for so long I wouldnt' have been so cavalier about switching from 32- to 64-bit. As I said in a reply to an earlier posting, it had been so long since Intel sent it to me that I'd actually forgotten it was an engineering sample until I popped it out and check the package information. I just thought it was an interesting troubleshooting exercise and wrote it up as such. It's your privilege (and apparent pleasure) to think otherwise. Thanks for the feedback, --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:47:17 +0000

I'm running BIOS version 0205 on the P5Q3, and had updated that some time last year. I assure you this was not the cause of the problem. I'll make you the same offer I made Intel: I'll sent you the chip (and if you like) the motherboard as well, and you can try it out to see if you can replicate my problem. I always visit the mobo website to grab the latest drivers and firmware as part of my initial setup routine. I should have documented the BIOS version in my story to forestall your analysis and (as it turns out) incorrect conclusion. Thanks for pointing this out, though: it's a valid point and newbie system builders will benefit from taking this into account in the future. I am not a newbie. --Ed-- PS: I'm looking at the box for the replacement motherboard, and I assure both board and box identify the product in question as Asus P5E3 Pro> here's the page on the Asus site for that very product:http://www.asus.com/product.aspx?P_ID=dsHHbxSbteK... PPS: If you want to take the time to try out the parts, you can find my complete contact information on my website atwww.edtittel.com.

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:39:30 +0000

Dear Lucien: I blame only myself for this. I had been using that chip for over 2 years on 32-bit Vista and Win7 installations without any problems at all. It had been so long since Intel had sent me the chip that I had absolutely forgotten it was an engineering sample until I plucked it out and looked up the on-package info. Indeed, I am aware that engineering samples are not warranted, not supported, and generally not suitable for production use. I wanted to make the point that all components should be suspect when troubleshooting systems, especially when "truly weird" things start happening--as they did in this particular case. Thanks for your feedback. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:36:34 +0000

I've got a Sea Sonic Power Angel that measures actual power consumption at the socket. This build never exceeds 280 Watts of power draw. I was pretty sure the PSU wasn't at fault, but in my desire to avoid suspecting the CPU (if you like), or my desire to check everything (which I like), I decided it would be worth swapping out in the interests of thorough coverage. Where did you get the idea that I had the BIOS settings wrong from the get-go? I left the defaults intact (I assume you're talking about the 2nd mobo, correct) on the faulty assumption that the vendor would go with basic defaults (as has been my experience on every other Asus motherboard I've ever used, for a total of somewhere north of 20 such boards). I may be guilty of overlooking a cursory check, but I didn't change the BIOS in error: I changed it to turn off what proved to be an overly aggressive default. Glad to bring joy, sunshine, and chuckle into your life nevertheless. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:30:58 +0000

Yes, the original configuration with new proc works, but with old proc does not. I did verify that at some point, but decided that detail wasn't relevant to my story. I'm really not after Intel here, I just want to make the point that "positive paranoia" (it's *GOOD* to suspect EVERYTHING, that is) is a valuable approach when serious troubleshooting is required. Thanks for your feedback. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:29:28 +0000

Gee, your scenario sounds even more "interesting" than mine! My sympathies. I've dealt with flaky memory many times and appreciate your approach. Glad you were able to find a workable compromise. Thanks for your feedback, --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:27:58 +0000

Dear Penang: 1. I did move the QX9650 from the old P5EQ3 to the new P5E3 Pro. 2. The original build ran Vista Ultimate for 1.5 years and Win7 Ultimate for nearly a year (both x86/32-bit) without any apparent problems 3. It pretty much has to be the CPU, because I can make things work by holding all other parts constant and using a different LGA 775 CPU, then make them fail by holding all other parts constant and using this particular CPU. I agree with your point about MS including a facility to slipstream drivers into an install more easily. This can be done using various tools like Sysprep, but it's time consuming and not terribly fun. Thanks for your feedback. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:24:25 +0000

Dear NTH: I guess I should count my lucky stars that I use DriverAgent to bring drivers up to snuff, and it steered me to the ATK0110/ACPI driver at the Asus site instead of the MS download. So far, no problems on that front, though I've seen it happen on other builds myself. Thanks for the feedback. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:21:53 +0000

I absolutely don't blame Intel for this: in fact, they are very careful to eschew any and all support and warranty on engineering sample parts. It was my unwillingness to consider the CPU as a possible cause that forced me to solve this problem "the long way around." I'm much more of an Intel fanboy than an MS fanboy in any case, but didn't see how the OS could have caused this particular problem. Thanks for your feedback. --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:20:09 +0000

Heh! Heh! I couldn't agree with you more: I was *stunned* to realize that Asus turns overclocking/AI tweaking on in its latest motherboards by default. I generally tend to start up with really basic set-ups: that is, exact clocking of the CPU at rated speed with FSB settings to match, or slower. As soon as I put 4 DIMMs into that mobo, I knew I'd have to turn the FSB clock down because ICH9 memory controllers don't do anywhere near as well with all slots filled as they do with a single bank filled instead (I would normally go with 2 x 4GB to get to 8, but I don't have any spare 4GB DDR3 DIMMs laying around, and they still cost about $200 a pop so I don't want to run out and buy any right now, either). Thanks for your feedback, --Ed--

IT Expert Voice : Lessons in Troubleshooting: Focus and Systematic Coverage Are Key

Tue, 13 Apr 2010 15:17:21 +0000

Good suggestion about other 64-bit OSes. I'll try that next month when I have some time off, and report back here. As for faith in Windows OSes, I've been working with them too long to exactly consider my attitude toward them one of faith--it's more like "what is it *THIS* time?" Gave me a good chuckle, though. I actually have another sample from the same production batch (it's on my primary production machine right now, though, so I don't want to mess with it). I will be building myself a new i7 PC next month (again, during time off) and will then try installing 64-bit on the other "same processor" just to see what happens. Thanks for your feedback, --Ed--